title
stringlengths 13
150
| url
stringlengths 97
97
| authors
stringlengths 8
467
| detail_url
stringlengths 97
97
| tags
stringclasses 1
value | AuthorFeedback
stringlengths 102
102
⌀ | Bibtex
stringlengths 53
54
| MetaReview
stringlengths 99
99
| Paper
stringlengths 93
93
| Review
stringlengths 95
95
| Supplemental
stringlengths 100
100
⌀ | abstract
stringlengths 53
2k
|
|---|---|---|---|---|---|---|---|---|---|---|---|
Adversarially Robust Few-Shot Learning: A Meta-Learning Approach | https://papers.nips.cc/paper_files/paper/2020/hash/cfee398643cbc3dc5eefc89334cacdc1-Abstract.html | Micah Goldblum, Liam Fowl, Tom Goldstein | https://papers.nips.cc/paper_files/paper/2020/hash/cfee398643cbc3dc5eefc89334cacdc1-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/cfee398643cbc3dc5eefc89334cacdc1-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11225-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/cfee398643cbc3dc5eefc89334cacdc1-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/cfee398643cbc3dc5eefc89334cacdc1-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/cfee398643cbc3dc5eefc89334cacdc1-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/cfee398643cbc3dc5eefc89334cacdc1-Supplemental.pdf | Previous work on adversarially robust neural networks for image classification requires large training sets and computationally expensive training procedures. On the other hand, few-shot learning methods are highly vulnerable to adversarial examples. The goal of our work is to produce networks which both perform well at few-shot classification tasks and are simultaneously robust to adversarial examples. We develop an algorithm, called Adversarial Querying (AQ), for producing adversarially robust meta-learners, and we thoroughly investigate the causes for adversarial vulnerability. Moreover, our method achieves far superior robust performance on few-shot image classification tasks, such as Mini-ImageNet and CIFAR-FS, than robust transfer learning. |
Neural Anisotropy Directions | https://papers.nips.cc/paper_files/paper/2020/hash/cff02a74da64d145a4aed3a577a106ab-Abstract.html | Guillermo Ortiz-Jimenez, Apostolos Modas, Seyed-Mohsen Moosavi, Pascal Frossard | https://papers.nips.cc/paper_files/paper/2020/hash/cff02a74da64d145a4aed3a577a106ab-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/cff02a74da64d145a4aed3a577a106ab-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11226-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/cff02a74da64d145a4aed3a577a106ab-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/cff02a74da64d145a4aed3a577a106ab-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/cff02a74da64d145a4aed3a577a106ab-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/cff02a74da64d145a4aed3a577a106ab-Supplemental.pdf | In this work, we analyze the role of the network architecture in shaping the inductive bias of deep classifiers. To that end, we start by focusing on a very simple problem, i.e., classifying a class of linearly separable distributions, and show that, depending on the direction of the discriminative feature of the distribution, many state-of-the-art deep convolutional neural networks (CNNs) have a surprisingly hard time solving this simple task. We then define as neural anisotropy directions (NADs) the vectors that encapsulate the directional inductive bias of an architecture. These vectors, which are specific for each architecture and hence act as a signature, encode the preference of a network to separate the input data based on some particular features. We provide an efficient method to identify NADs for several CNN architectures and thus reveal their directional inductive biases. Furthermore, we show that, for the CIFAR-10 dataset, NADs characterize the features used by CNNs to discriminate between different classes. |
Digraph Inception Convolutional Networks | https://papers.nips.cc/paper_files/paper/2020/hash/cffb6e2288a630c2a787a64ccc67097c-Abstract.html | Zekun Tong, Yuxuan Liang, Changsheng Sun, Xinke Li, David Rosenblum, Andrew Lim | https://papers.nips.cc/paper_files/paper/2020/hash/cffb6e2288a630c2a787a64ccc67097c-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/cffb6e2288a630c2a787a64ccc67097c-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11227-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/cffb6e2288a630c2a787a64ccc67097c-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/cffb6e2288a630c2a787a64ccc67097c-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/cffb6e2288a630c2a787a64ccc67097c-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/cffb6e2288a630c2a787a64ccc67097c-Supplemental.pdf | Graph Convolutional Networks (GCNs) have shown promising results in modeling graph-structured data. However, they have difficulty with processing digraphs because of two reasons: 1) transforming directed to undirected graph to guarantee the symmetry of graph Laplacian is not reasonable since it not only misleads message passing scheme to aggregate incorrect weights but also deprives the unique characteristics of digraph structure; 2) due to the fixed receptive field in each layer, GCNs fail to obtain multi-scale features that can boost their performance. In this paper, we theoretically extend spectral-based graph convolution to digraphs and derive a simplified form using personalized PageRank. Specifically, we present the Digraph Inception Convolutional Networks (DiGCN) which utilizes digraph convolution and kth-order proximity to achieve larger receptive fields and learn multi-scale features in digraphs. We empirically show that DiGCN can encode more structural information from digraphs than GCNs and help achieve better performance when generalized to other models. Moreover, experiments on various benchmarks demonstrate its superiority against the state-of-the-art methods. |
PAC-Bayesian Bound for the Conditional Value at Risk | https://papers.nips.cc/paper_files/paper/2020/hash/d02e9bdc27a894e882fa0c9055c99722-Abstract.html | Zakaria Mhammedi, Benjamin Guedj, Robert C. Williamson | https://papers.nips.cc/paper_files/paper/2020/hash/d02e9bdc27a894e882fa0c9055c99722-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d02e9bdc27a894e882fa0c9055c99722-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11228-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d02e9bdc27a894e882fa0c9055c99722-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d02e9bdc27a894e882fa0c9055c99722-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d02e9bdc27a894e882fa0c9055c99722-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d02e9bdc27a894e882fa0c9055c99722-Supplemental.pdf | Conditional Value at Risk ($\textsc{CVaR}$) is a ``coherent risk measure'' which generalizes expectation (reduced to a boundary parameter setting).
Widely used in mathematical finance, it is garnering increasing interest in machine learning as an alternate approach to regularization, and as a means for ensuring fairness.
This paper presents a generalization bound for learning algorithms that minimize the $\textsc{CVaR}$ of the empirical loss.
The bound is of PAC-Bayesian type and is guaranteed to be small when the empirical $\textsc{CVaR}$ is small.
We achieve this by reducing the problem of estimating $\textsc{CVaR}$ to that of merely estimating an expectation. This then enables us, as a by-product, to obtain concentration inequalities for $\textsc{CVaR}$ even when the random variable in question is unbounded. |
Stochastic Stein Discrepancies | https://papers.nips.cc/paper_files/paper/2020/hash/d03a857a23b5285736c4d55e0bb067c8-Abstract.html | Jackson Gorham, Anant Raj, Lester Mackey | https://papers.nips.cc/paper_files/paper/2020/hash/d03a857a23b5285736c4d55e0bb067c8-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d03a857a23b5285736c4d55e0bb067c8-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11229-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d03a857a23b5285736c4d55e0bb067c8-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d03a857a23b5285736c4d55e0bb067c8-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d03a857a23b5285736c4d55e0bb067c8-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d03a857a23b5285736c4d55e0bb067c8-Supplemental.pdf | Stein discrepancies (SDs) monitor convergence and non-convergence in approximate inference when exact integration and sampling are intractable. However, the computation of a Stein discrepancy can be prohibitive if the Stein operator -- often a sum over likelihood terms or potentials -- is expensive to evaluate. To address this deficiency, we show that stochastic Stein discrepancies (SSDs) based on subsampled approximations of the Stein operator inherit the convergence control properties of standard SDs with probability 1. Along the way, we establish the convergence of Stein variational gradient descent (SVGD) on unbounded domains, resolving an open question of Liu (2017). In our experiments with biased Markov chain Monte Carlo (MCMC) hyperparameter tuning, approximate MCMC sampler selection, and stochastic SVGD, SSDs deliver comparable inferences to standard SDs with orders of magnitude fewer likelihood evaluations. |
On the Role of Sparsity and DAG Constraints for Learning Linear DAGs | https://papers.nips.cc/paper_files/paper/2020/hash/d04d42cdf14579cd294e5079e0745411-Abstract.html | Ignavier Ng, AmirEmad Ghassami, Kun Zhang | https://papers.nips.cc/paper_files/paper/2020/hash/d04d42cdf14579cd294e5079e0745411-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d04d42cdf14579cd294e5079e0745411-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11230-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d04d42cdf14579cd294e5079e0745411-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d04d42cdf14579cd294e5079e0745411-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d04d42cdf14579cd294e5079e0745411-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d04d42cdf14579cd294e5079e0745411-Supplemental.pdf | Learning graphical structure based on Directed Acyclic Graphs (DAGs) is a challenging problem, partly owing to the large search space of possible graphs. A recent line of work formulates the structure learning problem as a continuous constrained optimization task using the least squares objective and an algebraic characterization of DAGs. However, the formulation requires a hard DAG constraint and may lead to optimization difficulties. In this paper, we study the asymptotic role of the sparsity and DAG constraints for learning DAG models in the linear Gaussian and non-Gaussian cases, and investigate their usefulness in the finite sample regime. Based on the theoretical results, we formulate a likelihood-based score function, and show that one only has to apply soft sparsity and DAG constraints to learn a DAG equivalent to the ground truth DAG. This leads to an unconstrained optimization problem that is much easier to solve. Using gradient-based optimization and GPU acceleration, our procedure can easily handle thousands of nodes while retaining a high accuracy. Extensive experiments validate the effectiveness of our proposed method and show that the DAG-penalized likelihood objective is indeed favorable over the least squares one with the hard DAG constraint. |
Cream of the Crop: Distilling Prioritized Paths For One-Shot Neural Architecture Search | https://papers.nips.cc/paper_files/paper/2020/hash/d072677d210ac4c03ba046120f0802ec-Abstract.html | Houwen Peng, Hao Du, Hongyuan Yu, QI LI, Jing Liao, Jianlong Fu | https://papers.nips.cc/paper_files/paper/2020/hash/d072677d210ac4c03ba046120f0802ec-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d072677d210ac4c03ba046120f0802ec-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11231-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d072677d210ac4c03ba046120f0802ec-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d072677d210ac4c03ba046120f0802ec-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d072677d210ac4c03ba046120f0802ec-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d072677d210ac4c03ba046120f0802ec-Supplemental.pdf | One-shot weight sharing methods have recently drawn great attention in neural architecture search due to high efficiency and competitive performance. However, weight sharing across models has an inherent deficiency, i.e., insufficient training of subnetworks in the hypernetwork. To alleviate this problem, we present a simple yet effective architecture distillation method. The central idea is that subnetworks can learn collaboratively and teach each other throughout the training process, aiming to boost the convergence of individual models. We introduce the concept of prioritized path, which refers to the architecture candidates exhibiting superior performance during training. Distilling knowledge from the prioritized paths is able to boost the training of subnetworks. Since the prioritized paths are changed on the fly depending on their performance and complexity, the final obtained paths are the cream of the crop. We directly select the most promising one from the prioritized paths as the final architecture, without using other complex search methods, such as reinforcement learning or evolution algorithms. The experiments on ImageNet verify such path distillation method can improve the convergence ratio and performance of the hypernetwork, as well as boosting the training of subnetworks. The discovered architectures achieve superior performance compared to the recent MobileNetV3 and EfficientNet families under aligned settings. Moreover, the experiments on object detection and more challenging search space show the generality and robustness of the proposed method. Code and models are available at \url{https://github.com/neurips-20/cream.git}. |
Fair Multiple Decision Making Through Soft Interventions | https://papers.nips.cc/paper_files/paper/2020/hash/d0921d442ee91b896ad95059d13df618-Abstract.html | Yaowei Hu, Yongkai Wu, Lu Zhang, Xintao Wu | https://papers.nips.cc/paper_files/paper/2020/hash/d0921d442ee91b896ad95059d13df618-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d0921d442ee91b896ad95059d13df618-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11232-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d0921d442ee91b896ad95059d13df618-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d0921d442ee91b896ad95059d13df618-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d0921d442ee91b896ad95059d13df618-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d0921d442ee91b896ad95059d13df618-Supplemental.pdf | Previous research in fair classification mostly focuses on a single decision model. In reality, there usually exist multiple decision models within a system and all of which may contain a certain amount of discrimination. Such realistic scenarios introduce new challenges to fair classification: since discrimination may be transmitted from upstream models to downstream models, building decision models separately without taking upstream models into consideration cannot guarantee to achieve fairness. In this paper, we propose an approach that learns multiple classifiers and achieves fairness for all of them simultaneously, by treating each decision model as a soft intervention and inferring the post-intervention distributions to formulate the loss function as well as the fairness constraints. We adopt surrogate functions to smooth the loss function and constraints, and theoretically show that the excess risk of the proposed loss function can be bounded in a form that is the same as that for traditional surrogated loss functions. Experiments using both synthetic and real-world datasets show the effectiveness of our approach. |
Representation Learning for Integrating Multi-domain Outcomes to Optimize Individualized Treatment | https://papers.nips.cc/paper_files/paper/2020/hash/d0bb8259d8fe3c7df4554dab9d7da3c9-Abstract.html | Yuan Chen, Donglin Zeng, Tianchen Xu, Yuanjia Wang | https://papers.nips.cc/paper_files/paper/2020/hash/d0bb8259d8fe3c7df4554dab9d7da3c9-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d0bb8259d8fe3c7df4554dab9d7da3c9-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11233-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d0bb8259d8fe3c7df4554dab9d7da3c9-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d0bb8259d8fe3c7df4554dab9d7da3c9-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d0bb8259d8fe3c7df4554dab9d7da3c9-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d0bb8259d8fe3c7df4554dab9d7da3c9-Supplemental.pdf | For mental disorders, patients' underlying mental states are non-observed latent constructs which have to be inferred from observed multi-domain measurements such as diagnostic symptoms and patient functioning scores. Additionally, substantial heterogeneity in the disease diagnosis between patients needs to be addressed for optimizing individualized treatment policy in order to achieve precision medicine. To address these challenges, we propose an integrated learning framework that can simultaneously learn patients' underlying mental states and recommend optimal treatments for each individual. This learning framework is based on the measurement theory in psychiatry for modeling multiple disease diagnostic measures as arising from the underlying causes (true mental states). It allows incorporation of the multivariate pre- and post-treatment outcomes as well as biological measures while preserving the invariant structure for representing patients' latent mental states. A multi-layer neural network is used to allow complex treatment effect heterogeneity. Optimal treatment policy can be inferred for future patients by comparing their potential mental states under different treatments given the observed multi-domain pre-treatment measurements. Experiments on simulated data and a real-world clinical trial data show that the learned treatment polices compare favorably to alternative methods on heterogeneous treatment effects, and have broad utilities which lead to better patient outcomes on multiple domains. |
Learning to Play No-Press Diplomacy with Best Response Policy Iteration | https://papers.nips.cc/paper_files/paper/2020/hash/d1419302db9c022ab1d48681b13d5f8b-Abstract.html | Thomas Anthony, Tom Eccles, Andrea Tacchetti, János Kramár, Ian Gemp, Thomas Hudson, Nicolas Porcel, Marc Lanctot, Julien Perolat, Richard Everett, Satinder Singh, Thore Graepel, Yoram Bachrach | https://papers.nips.cc/paper_files/paper/2020/hash/d1419302db9c022ab1d48681b13d5f8b-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d1419302db9c022ab1d48681b13d5f8b-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11234-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d1419302db9c022ab1d48681b13d5f8b-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d1419302db9c022ab1d48681b13d5f8b-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d1419302db9c022ab1d48681b13d5f8b-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d1419302db9c022ab1d48681b13d5f8b-Supplemental.pdf | Recent advances in deep reinforcement learning (RL) have led to considerable progress in many 2-player zero-sum games, such as Go, Poker and Starcraft. The purely adversarial nature of such games allows for conceptually simple and principled application of RL methods. However real-world settings are many-agent, and agent interactions are complex mixtures of common-interest and competitive aspects. We consider Diplomacy, a 7-player board game designed to accentuate dilemmas resulting from many-agent interactions. It also features a large combinatorial action space and simultaneous moves, which are challenging for RL algorithms. We propose a simple yet effective approximate best response operator, designed to handle large combinatorial action spaces and simultaneous moves. We also introduce a family of policy iteration methods that approximate fictitious play. With these methods, we successfully apply RL to Diplomacy: we show that our agents convincingly outperform the previous state-of-the-art, and game theoretic equilibrium analysis shows that the new process yields consistent improvements. |
Inverse Learning of Symmetries | https://papers.nips.cc/paper_files/paper/2020/hash/d15426b9c324676610fbb01360473ed8-Abstract.html | Mario Wieser, Sonali Parbhoo, Aleksander Wieczorek, Volker Roth | https://papers.nips.cc/paper_files/paper/2020/hash/d15426b9c324676610fbb01360473ed8-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d15426b9c324676610fbb01360473ed8-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11235-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d15426b9c324676610fbb01360473ed8-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d15426b9c324676610fbb01360473ed8-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d15426b9c324676610fbb01360473ed8-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d15426b9c324676610fbb01360473ed8-Supplemental.pdf | Symmetry transformations induce invariances and are a crucial building block of modern machine learning algorithms. In many complex domains, such as the chemical space, invariances can be observed, yet the corresponding symmetry transformation cannot be formulated analytically. We propose to learn the symmetry transformation with a model consisting of two latent subspaces, where the first subspace captures the target and the second subspace the remaining invariant information. Our approach is based on the deep information bottleneck in combination with a continuous mutual information regulariser. Unlike previous methods, we focus on the challenging task of minimising mutual information in continuous domains. To this end, we base the calculation of mutual information on correlation matrices in combination with a bijective variable transformation. Extensive experiments demonstrate that our model outperforms state-of-the-art methods on artificial and molecular datasets. |
DiffGCN: Graph Convolutional Networks via Differential Operators and Algebraic Multigrid Pooling | https://papers.nips.cc/paper_files/paper/2020/hash/d16a974d4d6d0d71b29bfbfe045f1da7-Abstract.html | Moshe Eliasof, Eran Treister | https://papers.nips.cc/paper_files/paper/2020/hash/d16a974d4d6d0d71b29bfbfe045f1da7-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d16a974d4d6d0d71b29bfbfe045f1da7-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11236-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d16a974d4d6d0d71b29bfbfe045f1da7-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d16a974d4d6d0d71b29bfbfe045f1da7-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d16a974d4d6d0d71b29bfbfe045f1da7-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d16a974d4d6d0d71b29bfbfe045f1da7-Supplemental.zip | Graph Convolutional Networks (GCNs) have shown to be effective in handling
unordered data like point clouds and meshes. In this work we propose novel
approaches for graph convolution, pooling and unpooling, inspired from finite
differences and algebraic multigrid frameworks. We form a parameterized convolution
kernel based on discretized differential operators, leveraging the graph mass,
gradient and Laplacian. This way, the parameterization does not depend on the
graph structure, only on the meaning of the network convolutions as differential
operators. To allow hierarchical representations of the input, we propose pooling
and unpooling operations that are based on algebraic multigrid methods, which
are mainly used to solve partial differential equations on unstructured grids. To
motivate and explain our method, we compare it to standard convolutional neural
networks, and show their similarities and relations in the case of a regular grid. Our
proposed method is demonstrated in various experiments like classification and
part-segmentation, achieving on par or better than state of the art results. We also
analyze the computational cost of our method compared to other GCNs. |
Distributed Newton Can Communicate Less and Resist Byzantine Workers | https://papers.nips.cc/paper_files/paper/2020/hash/d17e6bcbcef8de3f7a00195cfa5706f1-Abstract.html | Avishek Ghosh, Raj Kumar Maity, Arya Mazumdar | https://papers.nips.cc/paper_files/paper/2020/hash/d17e6bcbcef8de3f7a00195cfa5706f1-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d17e6bcbcef8de3f7a00195cfa5706f1-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11237-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d17e6bcbcef8de3f7a00195cfa5706f1-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d17e6bcbcef8de3f7a00195cfa5706f1-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d17e6bcbcef8de3f7a00195cfa5706f1-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d17e6bcbcef8de3f7a00195cfa5706f1-Supplemental.zip | We develop a distributed second order optimization algorithm that is communication-efficient as well as robust against Byzantine failures of the worker machines. We propose an iterative approximate Newton-type algorithm, where the worker machines communicate \emph{only once} per iteration with the central machine. This is in sharp contrast with the state-of-the-art distributed second order algorithms like GIANT \cite{giant}, DINGO\cite{dingo}, where the worker machines send (functions of) local gradient and Hessian sequentially; thus ending up communicating twice with the central machine per iteration. Furthermore, we employ a simple norm based thresholding rule to filter-out the Byzantine worker machines. We establish the linear-quadratic rate of convergence of our proposed algorithm and establish that the communication savings and Byzantine resilience attributes only correspond to a small statistical error rate for arbitrary convex loss functions. To the best of our knowledge, this is the first work that addresses the issue of Byzantine resilience in second order distributed optimization. Furthermore, we validate our theoretical results with extensive experiments on synthetically generated and benchmark LIBSVM \cite{libsvm} data-set and demonstrate convergence guarantees. |
Efficient Nonmyopic Bayesian Optimization via One-Shot Multi-Step Trees | https://papers.nips.cc/paper_files/paper/2020/hash/d1d5923fc822531bbfd9d87d4760914b-Abstract.html | Shali Jiang, Daniel Jiang, Maximilian Balandat, Brian Karrer, Jacob Gardner, Roman Garnett | https://papers.nips.cc/paper_files/paper/2020/hash/d1d5923fc822531bbfd9d87d4760914b-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d1d5923fc822531bbfd9d87d4760914b-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11238-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d1d5923fc822531bbfd9d87d4760914b-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d1d5923fc822531bbfd9d87d4760914b-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d1d5923fc822531bbfd9d87d4760914b-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d1d5923fc822531bbfd9d87d4760914b-Supplemental.pdf | Bayesian optimization is a sequential decision making framework for optimizing expensive-to-evaluate black-box functions. Computing a full lookahead policy amounts to solving a highly intractable stochastic dynamic program. Myopic approaches, such as expected improvement, are often adopted in practice, but they ignore the long-term impact of the immediate decision. Existing nonmyopic approaches are mostly heuristic and/or computationally expensive. In this paper, we provide the first efficient implementation of general multi-step lookahead Bayesian optimization, formulated as a sequence of nested optimization problems within a multi-step scenario tree. Instead of solving these problems in a nested way, we equivalently optimize all decision variables in the full tree jointly, in a "one-shot" fashion. Combining this with an efficient method for implementing multi-step Gaussian process "fantasization," we demonstrate that multi-step expected improvement is computationally tractable and exhibits performance superior to existing methods on a wide range of benchmarks. |
Effective Diversity in Population Based Reinforcement Learning | https://papers.nips.cc/paper_files/paper/2020/hash/d1dc3a8270a6f9394f88847d7f0050cf-Abstract.html | Jack Parker-Holder, Aldo Pacchiano, Krzysztof M. Choromanski, Stephen J. Roberts | https://papers.nips.cc/paper_files/paper/2020/hash/d1dc3a8270a6f9394f88847d7f0050cf-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d1dc3a8270a6f9394f88847d7f0050cf-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11239-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d1dc3a8270a6f9394f88847d7f0050cf-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d1dc3a8270a6f9394f88847d7f0050cf-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d1dc3a8270a6f9394f88847d7f0050cf-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d1dc3a8270a6f9394f88847d7f0050cf-Supplemental.pdf | Exploration is a key problem in reinforcement learning, since agents can only learn from data they acquire in the environment. With that in mind, maintaining a population of agents is an attractive method, as it allows data be collected with a diverse set of behaviors. This behavioral diversity is often boosted via multi-objective loss functions. However, those approaches typically leverage mean field updates based on pairwise distances, which makes them susceptible to cycling behaviors and increased redundancy. In addition, explicitly boosting diversity often has a detrimental impact on optimizing already fruitful behaviors for rewards. As such, the reward-diversity trade off typically relies on heuristics. Finally, such methods require behavioral representations, often handcrafted and domain specific. In this paper, we introduce an approach to optimize all members of a population simultaneously. Rather than using pairwise distance, we measure the volume of the entire population in a behavioral manifold, defined by task-agnostic behavioral embeddings. In addition, our algorithm Diversity via Determinants (DvD), adapts the degree of diversity during training using online learning techniques. We introduce both evolutionary and gradient-based instantiations of DvD and show they effectively improve exploration without reducing performance when better exploration is not required. |
Elastic-InfoGAN: Unsupervised Disentangled Representation Learning in Class-Imbalanced Data | https://papers.nips.cc/paper_files/paper/2020/hash/d1e39c9bda5c80ac3d8ea9d658163967-Abstract.html | Utkarsh Ojha, Krishna Kumar Singh, Cho-Jui Hsieh, Yong Jae Lee | https://papers.nips.cc/paper_files/paper/2020/hash/d1e39c9bda5c80ac3d8ea9d658163967-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d1e39c9bda5c80ac3d8ea9d658163967-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11240-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d1e39c9bda5c80ac3d8ea9d658163967-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d1e39c9bda5c80ac3d8ea9d658163967-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d1e39c9bda5c80ac3d8ea9d658163967-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d1e39c9bda5c80ac3d8ea9d658163967-Supplemental.pdf | We propose a novel unsupervised generative model that learns to disentangle object identity from other low-level aspects in class-imbalanced data. We first investigate the issues surrounding the assumptions about uniformity made by InfoGAN, and demonstrate its ineffectiveness to properly disentangle object identity in imbalanced data. Our key idea is to make the discovery of the discrete latent factor of variation invariant to identity-preserving transformations in real images, and use that as a signal to learn the appropriate latent distribution representing object identity. Experiments on both artificial (MNIST, 3D cars, 3D chairs, ShapeNet) and real-world (YouTube-Faces) imbalanced datasets demonstrate the effectiveness of our method in disentangling object identity as a latent factor of variation. |
Direct Policy Gradients: Direct Optimization of Policies in Discrete Action Spaces | https://papers.nips.cc/paper_files/paper/2020/hash/d1e7b08bdb7783ed4fb10abe92c22ffd-Abstract.html | Guy Lorberbom, Chris J. Maddison, Nicolas Heess, Tamir Hazan, Daniel Tarlow | https://papers.nips.cc/paper_files/paper/2020/hash/d1e7b08bdb7783ed4fb10abe92c22ffd-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d1e7b08bdb7783ed4fb10abe92c22ffd-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11241-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d1e7b08bdb7783ed4fb10abe92c22ffd-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d1e7b08bdb7783ed4fb10abe92c22ffd-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d1e7b08bdb7783ed4fb10abe92c22ffd-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d1e7b08bdb7783ed4fb10abe92c22ffd-Supplemental.zip | Direct optimization (McAllester et al., 2010; Song et al., 2016) is an appealing framework that replaces integration with optimization of a random objective for approximating gradients in models with discrete random variables (Lorberbom et al., 2018). A* sampling (Maddison et al., 2014) is a framework for optimizing such random objectives over large spaces. We show how to combine these techniques to yield a reinforcement learning algorithm that approximates a policy gradient by finding trajectories that optimize a random objective. We call the resulting algorithms \emph{direct policy gradient} (DirPG) algorithms. A main benefit of DirPG algorithms is that they allow the insertion of domain knowledge in the form of upper bounds on return-to-go at training time, like is used in heuristic search, while still directly computing a policy gradient. We further analyze their properties, showing there are cases where DirPG has an exponentially larger probability of sampling informative gradients compared to REINFORCE. We also show that there is a built-in variance reduction technique and that a parameter that was previously viewed as a numerical approximation can be interpreted as controlling risk sensitivity. Empirically, we evaluate the effect of key degrees of freedom and show that the algorithm performs well in illustrative domains compared to baselines. |
Hybrid Models for Learning to Branch | https://papers.nips.cc/paper_files/paper/2020/hash/d1e946f4e67db4b362ad23818a6fb78a-Abstract.html | Prateek Gupta, Maxime Gasse, Elias Khalil, Pawan Mudigonda, Andrea Lodi, Yoshua Bengio | https://papers.nips.cc/paper_files/paper/2020/hash/d1e946f4e67db4b362ad23818a6fb78a-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d1e946f4e67db4b362ad23818a6fb78a-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11242-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d1e946f4e67db4b362ad23818a6fb78a-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d1e946f4e67db4b362ad23818a6fb78a-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d1e946f4e67db4b362ad23818a6fb78a-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d1e946f4e67db4b362ad23818a6fb78a-Supplemental.pdf | A recent Graph Neural Network (GNN) approach for learning to branch has been shown to successfully reduce the running time of branch-and-bound algorithms for Mixed Integer Linear Programming (MILP). While the GNN relies on a GPU for inference, MILP solvers are purely CPU-based. This severely limits its application as many practitioners may not have access to high-end GPUs. In this work, we ask two key questions. First, in a more realistic setting where only a CPU is available, is the GNN model still competitive? Second, can we devise an alternate computationally inexpensive model that retains the predictive power of the GNN architecture? We answer the first question in the negative, and address the second question by proposing a new hybrid architecture for efficient branching on CPU machines. The proposed architecture combines the expressive power of GNNs with computationally inexpensive multi-layer perceptrons (MLP) for branching. We evaluate our methods on four classes of MILP problems, and show that they lead to up to 26% reduction in solver running time compared to state-of-the-art methods without a GPU, while extrapolating to harder problems than it was trained on. The code for this project is publicly available at https://github.com/pg2455/Hybrid-learn2branch. |
WoodFisher: Efficient Second-Order Approximation for Neural Network Compression | https://papers.nips.cc/paper_files/paper/2020/hash/d1ff1ec86b62cd5f3903ff19c3a326b2-Abstract.html | Sidak Pal Singh, Dan Alistarh | https://papers.nips.cc/paper_files/paper/2020/hash/d1ff1ec86b62cd5f3903ff19c3a326b2-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d1ff1ec86b62cd5f3903ff19c3a326b2-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11243-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d1ff1ec86b62cd5f3903ff19c3a326b2-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d1ff1ec86b62cd5f3903ff19c3a326b2-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d1ff1ec86b62cd5f3903ff19c3a326b2-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d1ff1ec86b62cd5f3903ff19c3a326b2-Supplemental.pdf | Our main application is to neural network compression, where we build on the classic Optimal Brain Damage/Surgeon framework. We demonstrate that WoodFisher significantly outperforms popular state-of-the-art methods for one-shot pruning. Further, even when iterative, gradual pruning is allowed, our method results in a gain in test accuracy over the state-of-the-art approaches for popular image classification datasets such as ImageNet ILSVRC. Further, we show how our method can be extended to take into account first-order information, and illustrate its ability to automatically set layer-wise pruning thresholds, or perform compression in the limited-data regime. |
Bi-level Score Matching for Learning Energy-based Latent Variable Models | https://papers.nips.cc/paper_files/paper/2020/hash/d25a34b9c2a87db380ecd7f7115882ec-Abstract.html | Fan Bao, Chongxuan LI, Kun Xu, Hang Su, Jun Zhu, Bo Zhang | https://papers.nips.cc/paper_files/paper/2020/hash/d25a34b9c2a87db380ecd7f7115882ec-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d25a34b9c2a87db380ecd7f7115882ec-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11244-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d25a34b9c2a87db380ecd7f7115882ec-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d25a34b9c2a87db380ecd7f7115882ec-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d25a34b9c2a87db380ecd7f7115882ec-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d25a34b9c2a87db380ecd7f7115882ec-Supplemental.pdf | Score matching (SM) provides a compelling approach to learn energy-based models (EBMs) by avoiding the calculation of partition function. However, it remains largely open to learn energy-based latent variable models (EBLVMs), except some special cases. This paper presents a bi-level score matching (BiSM) method to learn EBLVMs with general structures by reformulating SM as a bi-level optimization problem. The higher level introduces a variational posterior of the latent variables and optimizes a modified SM objective, and the lower level optimizes the variational posterior to fit the true posterior. To solve BiSM efficiently, we develop a stochastic optimization algorithm with gradient unrolling. Theoretically, we analyze the consistency of BiSM and the convergence of the stochastic algorithm. Empirically, we show the promise of BiSM in Gaussian restricted Boltzmann machines and highly nonstructural EBLVMs parameterized by deep convolutional neural networks. BiSM is comparable to the widely adopted contrastive divergence and SM methods when they are applicable; and can learn complex EBLVMs with intractable posteriors to generate natural images. |
Counterfactual Contrastive Learning for Weakly-Supervised Vision-Language Grounding | https://papers.nips.cc/paper_files/paper/2020/hash/d27b95cac4c27feb850aaa4070cc4675-Abstract.html | Zhu Zhang, Zhou Zhao, Zhijie Lin, jieming zhu, Xiuqiang He | https://papers.nips.cc/paper_files/paper/2020/hash/d27b95cac4c27feb850aaa4070cc4675-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d27b95cac4c27feb850aaa4070cc4675-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11245-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d27b95cac4c27feb850aaa4070cc4675-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d27b95cac4c27feb850aaa4070cc4675-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d27b95cac4c27feb850aaa4070cc4675-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d27b95cac4c27feb850aaa4070cc4675-Supplemental.pdf | Weakly-supervised vision-language grounding aims to localize a target moment in a video or a specific region in an image according to the given sentence query, where only video-level or image-level sentence annotations are provided during training. Most existing approaches employ the MIL-based or reconstruction-based paradigms for the WSVLG task, but the former heavily depends on the quality of randomly-selected negative samples and the latter cannot directly optimize the visual-textual alignment score. In this paper, we propose a novel Counterfactual Contrastive Learning (CCL) to develop sufficient contrastive training between counterfactual positive and negative results, which are based on robust and destructive counterfactual transformations. Concretely, we design three counterfactual transformation strategies from the feature-, interaction- and relation-level, where the feature-level method damages the visual features of selected proposals, interaction-level approach confuses the vision-language interaction and relation-level strategy destroys the context clues in proposal relationships. Extensive experiments on five vision-language grounding datasets verify the effectiveness of our CCL paradigm. |
Decision trees as partitioning machines to characterize their generalization properties | https://papers.nips.cc/paper_files/paper/2020/hash/d2a10b0bd670e442b1d3caa3fbf9e695-Abstract.html | Jean-Samuel Leboeuf, Frédéric LeBlanc, Mario Marchand | https://papers.nips.cc/paper_files/paper/2020/hash/d2a10b0bd670e442b1d3caa3fbf9e695-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d2a10b0bd670e442b1d3caa3fbf9e695-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11246-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d2a10b0bd670e442b1d3caa3fbf9e695-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d2a10b0bd670e442b1d3caa3fbf9e695-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d2a10b0bd670e442b1d3caa3fbf9e695-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d2a10b0bd670e442b1d3caa3fbf9e695-Supplemental.zip | Decision trees are popular machine learning models that are simple to build and easy to interpret. Even though algorithms to learn decision trees date back to almost 50 years, key properties affecting their generalization error are still weakly bounded. Hence, we revisit binary decision trees on real-valued features from the perspective of partitions of the data. We introduce the notion of partitioning function, and we relate it to the growth function and to the VC dimension. Using this new concept, we are able to find the exact VC dimension of decision stumps, which is given by the largest integer $d$ such that $2\ell \ge \binom{d}{\floor{\frac{d}{2}}}$, where $\ell$ is the number of real-valued features. We provide a recursive expression to bound the partitioning functions, resulting in a upper bound on the growth function of any decision tree structure. This allows us to show that the VC dimension of a binary tree structure with $N$ internal nodes is of order $N \log(N\ell)$. Finally, we elaborate a pruning algorithm based on these results that performs better than the CART algorithm on a number of datasets, with the advantage that no cross-validation is required. |
Learning to Prove Theorems by Learning to Generate Theorems | https://papers.nips.cc/paper_files/paper/2020/hash/d2a27e83d429f0dcae6b937cf440aeb1-Abstract.html | Mingzhe Wang, Jia Deng | https://papers.nips.cc/paper_files/paper/2020/hash/d2a27e83d429f0dcae6b937cf440aeb1-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d2a27e83d429f0dcae6b937cf440aeb1-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11247-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d2a27e83d429f0dcae6b937cf440aeb1-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d2a27e83d429f0dcae6b937cf440aeb1-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d2a27e83d429f0dcae6b937cf440aeb1-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d2a27e83d429f0dcae6b937cf440aeb1-Supplemental.pdf | We consider the task of automated theorem proving, a key AI task. Deep learning has shown promise for training theorem provers, but there are limited human-written theorems and proofs available for supervised learning. To address this limitation, we propose to learn a neural generator that automatically synthesizes theorems and proofs for the purpose of training a theorem prover. Experiments on real-world tasks demonstrate that synthetic data from our approach improves the theorem prover and advances the state of the art of automated theorem proving in Metamath. |
3D Self-Supervised Methods for Medical Imaging | https://papers.nips.cc/paper_files/paper/2020/hash/d2dc6368837861b42020ee72b0896182-Abstract.html | Aiham Taleb, Winfried Loetzsch, Noel Danz, Julius Severin, Thomas Gaertner, Benjamin Bergner, Christoph Lippert | https://papers.nips.cc/paper_files/paper/2020/hash/d2dc6368837861b42020ee72b0896182-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d2dc6368837861b42020ee72b0896182-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11248-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d2dc6368837861b42020ee72b0896182-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d2dc6368837861b42020ee72b0896182-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d2dc6368837861b42020ee72b0896182-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d2dc6368837861b42020ee72b0896182-Supplemental.pdf | Self-supervised learning methods have witnessed a recent surge of interest after proving successful in multiple application fields.
In this work, we leverage these techniques, and we propose 3D versions for five different self-supervised methods, in the form of proxy tasks. Our methods facilitate neural network feature learning from unlabeled 3D images, aiming to reduce the required cost for expert annotation.
The developed algorithms are 3D Contrastive Predictive Coding, 3D Rotation prediction, 3D Jigsaw puzzles, Relative 3D patch location, and 3D Exemplar networks.
Our experiments show that pretraining models with our 3D tasks yields more powerful semantic representations, and enables solving downstream tasks more accurately and efficiently, compared to training the models from scratch and to pretraining them on 2D slices.
We demonstrate the effectiveness of our methods on three downstream tasks from the medical imaging domain: i) Brain Tumor Segmentation from 3D MRI, ii) Pancreas Tumor Segmentation from 3D CT, and iii) Diabetic Retinopathy Detection from 2D Fundus images. In each task, we assess the gains in data-efficiency, performance, and speed of convergence. Interestingly, we also find gains when transferring the learned representations, by our methods, from a large unlabeled 3D corpus to a small downstream-specific dataset.
We achieve results competitive to state-of-the-art solutions at a fraction of the computational expense.
We publish our implementations for the developed algorithms (both 3D and 2D versions) as an open-source library, in an effort to allow other researchers to apply and extend our methods on their datasets. |
Bayesian filtering unifies adaptive and non-adaptive neural network optimization methods | https://papers.nips.cc/paper_files/paper/2020/hash/d33174c464c877fb03e77efdab4ae804-Abstract.html | Laurence Aitchison | https://papers.nips.cc/paper_files/paper/2020/hash/d33174c464c877fb03e77efdab4ae804-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d33174c464c877fb03e77efdab4ae804-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11249-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d33174c464c877fb03e77efdab4ae804-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d33174c464c877fb03e77efdab4ae804-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d33174c464c877fb03e77efdab4ae804-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d33174c464c877fb03e77efdab4ae804-Supplemental.pdf | We formulate the problem of neural network optimization as Bayesian filtering, where the observations are backpropagated gradients. While neural network optimization has previously been studied using natural gradient methods which are closely related to Bayesian inference, they were unable to recover standard optimizers such as Adam and RMSprop with a root-mean-square gradient normalizer, instead getting a mean-square normalizer. To recover the root-mean-square normalizer, we find it necessary to account for the temporal dynamics of all the other parameters as they are optimized. The resulting optimizer, AdaBayes, adaptively transitions between SGD-like and Adam-like behaviour, automatically recovers AdamW, a state of the art variant of Adam with decoupled weight decay, and has generalisation performance competitive with SGD. |
Worst-Case Analysis for Randomly Collected Data | https://papers.nips.cc/paper_files/paper/2020/hash/d34a281acc62c6bec66425f0ad6dd645-Abstract.html | Justin Chen, Gregory Valiant, Paul Valiant | https://papers.nips.cc/paper_files/paper/2020/hash/d34a281acc62c6bec66425f0ad6dd645-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d34a281acc62c6bec66425f0ad6dd645-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11250-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d34a281acc62c6bec66425f0ad6dd645-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d34a281acc62c6bec66425f0ad6dd645-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d34a281acc62c6bec66425f0ad6dd645-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d34a281acc62c6bec66425f0ad6dd645-Supplemental.zip | We introduce a framework for statistical estimation that leverages knowledge of how samples are collected but makes no distributional assumptions on the data values. Specifically, we consider a population of elements [n]={1,...,n} with corresponding data values x1,...,xn. We observe the values for a "sample" set A \subset [n] and wish to estimate some statistic of the values for a "target" set B \subset [n] where B could be the entire set. Crucially, we assume that the sets A and B are drawn according to some known distribution P over pairs of subsets of [n]. A given estimation algorithm is evaluated based on its "worst-case, expected error" where the expectation is with respect to the distribution P from which the sample A and target sets B are drawn, and the worst-case is with respect to the data values x1,...,xn. Within this framework, we give an efficient algorithm for estimating the target mean that returns a weighted combination of the sample values–-where the weights are functions of the distribution P and the sample and target sets A, B--and show that the worst-case expected error achieved by this algorithm is at most a multiplicative pi/2 factor worse than the optimal of such algorithms. The algorithm and proof leverage a surprising connection to the Grothendieck problem. We also extend these results to the linear regression setting where each datapoint is not a scalar but a labeled vector (xi,yi). This framework, which makes no distributional assumptions on the data values but rather relies on knowledge of the data collection process via the distribution P, is a significant departure from the typical statistical estimation framework and introduces a uniform analysis for the many natural settings where membership in a sample may be correlated with data values, such as when individuals are recruited into a sample through their social networks as in "snowball/chain" sampling or when samples have chronological structure as in "selective prediction". |
Truthful Data Acquisition via Peer Prediction | https://papers.nips.cc/paper_files/paper/2020/hash/d35b05a832e2bb91f110d54e34e2da79-Abstract.html | Yiling Chen, Yiheng Shen, Shuran Zheng | https://papers.nips.cc/paper_files/paper/2020/hash/d35b05a832e2bb91f110d54e34e2da79-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d35b05a832e2bb91f110d54e34e2da79-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11251-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d35b05a832e2bb91f110d54e34e2da79-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d35b05a832e2bb91f110d54e34e2da79-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d35b05a832e2bb91f110d54e34e2da79-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d35b05a832e2bb91f110d54e34e2da79-Supplemental.pdf | We consider the problem of purchasing data for machine learning or statistical estimation. The data analyst has a budget to purchase datasets from multiple data providers. She does not have any test data that can be used to evaluate the collected data and can assign payments to data providers solely based on the collected datasets. We consider the problem in the standard Bayesian paradigm and in two settings: (1) data are only collected once; (2) data are collected repeatedly and each day's data are drawn independently from the same distribution. For both settings, our mechanisms guarantee that truthfully reporting one's dataset is always an equilibrium by adopting techniques from peer prediction: pay each provider the mutual information between his reported data and other providers' reported data. Depending on the data distribution, the mechanisms can also discourage misreports that would lead to inaccurate predictions. Our mechanisms also guarantee individual rationality and budget feasibility for certain underlying distributions in the first setting and for all distributions in the second setting. |
Learning Robust Decision Policies from Observational Data | https://papers.nips.cc/paper_files/paper/2020/hash/d3696cfb815ab692407d9362e6f06c28-Abstract.html | Muhammad Osama, Dave Zachariah, Peter Stoica | https://papers.nips.cc/paper_files/paper/2020/hash/d3696cfb815ab692407d9362e6f06c28-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d3696cfb815ab692407d9362e6f06c28-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11252-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d3696cfb815ab692407d9362e6f06c28-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d3696cfb815ab692407d9362e6f06c28-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d3696cfb815ab692407d9362e6f06c28-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d3696cfb815ab692407d9362e6f06c28-Supplemental.zip | We address the problem of learning a decision policy from observational data of past decisions in contexts with features and associated outcomes. The past policy maybe unknown and in safety-critical applications, such as medical decision support, it is of interest to learn robust policies that reduce the risk of outcomes with high costs. In this paper, we develop a method for learning policies that reduce tails of the cost distribution at a specified level and, moreover, provide a statistically valid bound on the cost of each decision. These properties are valid under finite samples -- even in scenarios with uneven or no overlap between features for different decisions in the observed data -- by building on recent results in conformal prediction. The performance and statistical properties of the proposed method are illustrated using both real and synthetic data. |
Byzantine Resilient Distributed Multi-Task Learning | https://papers.nips.cc/paper_files/paper/2020/hash/d37eb50d868361ea729bb4147eb3c1d8-Abstract.html | Jiani Li, Waseem Abbas, Xenofon Koutsoukos | https://papers.nips.cc/paper_files/paper/2020/hash/d37eb50d868361ea729bb4147eb3c1d8-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d37eb50d868361ea729bb4147eb3c1d8-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11253-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d37eb50d868361ea729bb4147eb3c1d8-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d37eb50d868361ea729bb4147eb3c1d8-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d37eb50d868361ea729bb4147eb3c1d8-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d37eb50d868361ea729bb4147eb3c1d8-Supplemental.pdf | Distributed multi-task learning provides significant advantages in multi-agent networks with heterogeneous data sources where agents aim to learn distinct but correlated models simultaneously. However, distributed algorithms for learning relatedness among tasks are not resilient in the presence of Byzantine agents. In this paper, we present an approach for Byzantine resilient distributed multi-task learning. We propose an efficient online weight assignment rule by measuring the accumulated loss using an agent’s data and its neighbors’ models. A small accumulated loss indicates a large similarity between the two tasks. In order to ensure the Byzantine resilience of the aggregation at a normal agent, we introduce a step for filtering out larger losses. We analyze the approach for convex models and show that normal agents converge resiliently towards their true targets. Further, an agent’s learning performance using the proposed weight assignment rule is guaranteed to be at least as good as in the non-cooperative case as measured by the expected regret. Finally, we demonstrate the approach using three case studies, including regression and classification problems, and show that our method exhibits good empirical performance for non-convex models, such as convolutional neural networks. |
Reinforcement Learning in Factored MDPs: Oracle-Efficient Algorithms and Tighter Regret Bounds for the Non-Episodic Setting | https://papers.nips.cc/paper_files/paper/2020/hash/d3b1fb02964aa64e257f9f26a31f72cf-Abstract.html | Ziping Xu, Ambuj Tewari | https://papers.nips.cc/paper_files/paper/2020/hash/d3b1fb02964aa64e257f9f26a31f72cf-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d3b1fb02964aa64e257f9f26a31f72cf-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11254-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d3b1fb02964aa64e257f9f26a31f72cf-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d3b1fb02964aa64e257f9f26a31f72cf-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d3b1fb02964aa64e257f9f26a31f72cf-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d3b1fb02964aa64e257f9f26a31f72cf-Supplemental.pdf | We study reinforcement learning in non-episodic factored Markov decision processes (FMDPs). We propose two near-optimal and oracle-efficient algorithms for FMDPs. Assuming oracle access to an FMDP planner, they enjoy a Bayesian and a frequentist regret bound respectively, both of which reduce to the near-optimal bound $O(DS\sqrt{AT})$ for standard non-factored MDPs. We propose a tighter connectivity measure, factored span, for FMDPs and prove a lower bound that depends on the factored span rather than the diameter $D$. In order to decrease the gap between lower and upper bounds, we propose an adaptation of the REGAL.C algorithm whose regret bound depends on the factored span. Our oracle-efficient algorithms outperform previously proposed near-optimal algorithms on computer network administration simulations. |
Improving model calibration with accuracy versus uncertainty optimization | https://papers.nips.cc/paper_files/paper/2020/hash/d3d9446802a44259755d38e6d163e820-Abstract.html | Ranganath Krishnan, Omesh Tickoo | https://papers.nips.cc/paper_files/paper/2020/hash/d3d9446802a44259755d38e6d163e820-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d3d9446802a44259755d38e6d163e820-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11255-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d3d9446802a44259755d38e6d163e820-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d3d9446802a44259755d38e6d163e820-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d3d9446802a44259755d38e6d163e820-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d3d9446802a44259755d38e6d163e820-Supplemental.pdf | Obtaining reliable and accurate quantification of uncertainty estimates from deep neural networks is important in safety-critical applications. A well-calibrated model should be accurate when it is certain about its prediction and indicate high uncertainty when it is likely to be inaccurate. Uncertainty calibration is a challenging problem as there is no ground truth available for uncertainty estimates. We propose an optimization method that leverages the relationship between accuracy and uncertainty as an anchor for uncertainty calibration. We introduce a differentiable accuracy versus uncertainty calibration (AvUC) loss function that allows a model to learn to provide well-calibrated uncertainties, in addition to improved accuracy. We also demonstrate the same methodology can be extended to post-hoc uncertainty calibration on pretrained models. We illustrate our approach with mean-field stochastic variational inference and compare with state-of-the-art methods. Extensive experiments demonstrate our approach yields better model calibration than existing methods on large-scale image classification tasks under distributional shift. |
The Convolution Exponential and Generalized Sylvester Flows | https://papers.nips.cc/paper_files/paper/2020/hash/d3f06eef2ffac7faadbe3055a70682ac-Abstract.html | Emiel Hoogeboom, Victor Garcia Satorras, Jakub Tomczak, Max Welling | https://papers.nips.cc/paper_files/paper/2020/hash/d3f06eef2ffac7faadbe3055a70682ac-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d3f06eef2ffac7faadbe3055a70682ac-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11256-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d3f06eef2ffac7faadbe3055a70682ac-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d3f06eef2ffac7faadbe3055a70682ac-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d3f06eef2ffac7faadbe3055a70682ac-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d3f06eef2ffac7faadbe3055a70682ac-Supplemental.pdf | This paper introduces a new method to build linear flows, by taking the exponential of a linear transformation. This linear transformation does not need to be invertible itself, and the exponential has the following desirable properties: it is guaranteed to be invertible, its inverse is straightforward to compute and the log Jacobian determinant is equal to the trace of the linear transformation. An important insight is that the exponential can be computed implicitly, which allows the use of convolutional layers. Using this insight, we develop new invertible transformations named convolution exponentials and graph convolution exponentials, which retain the equivariance of their underlying transformations. In addition, we generalize Sylvester Flows and propose Convolutional Sylvester Flows which are based on the generalization and the convolution exponential as basis change. Empirically, we show that the convolution exponential outperforms other linear transformations in generative flows on CIFAR10 and the graph convolution exponential improves the performance of graph normalizing flows. In addition, we show that Convolutional Sylvester Flows improve performance over residual flows as a generative flow model measured in log-likelihood. |
An Improved Analysis of Stochastic Gradient Descent with Momentum | https://papers.nips.cc/paper_files/paper/2020/hash/d3f5d4de09ea19461dab00590df91e4f-Abstract.html | Yanli Liu, Yuan Gao, Wotao Yin | https://papers.nips.cc/paper_files/paper/2020/hash/d3f5d4de09ea19461dab00590df91e4f-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d3f5d4de09ea19461dab00590df91e4f-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11257-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d3f5d4de09ea19461dab00590df91e4f-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d3f5d4de09ea19461dab00590df91e4f-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d3f5d4de09ea19461dab00590df91e4f-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d3f5d4de09ea19461dab00590df91e4f-Supplemental.pdf | SGD with momentum (SGDM) has been widely applied in many machine learning tasks, and it is often applied with dynamic stepsizes and momentum weights tuned in a stagewise manner. Despite of its empirical advantage over SGD, the role of momentum is still unclear in general since previous analyses on SGDM either provide worse convergence bounds than those of SGD, or assume Lipschitz or quadratic objectives, which fail to hold in practice. Furthermore, the role of dynamic parameters has not been addressed. In this work, we show that SGDM converges as fast as SGD for smooth objectives under both strongly convex and nonconvex settings. We also prove that multistage strategy is beneficial for SGDM compared to using fixed parameters. Finally, we verify these theoretical claims by numerical experiments. |
Precise expressions for random projections: Low-rank approximation and randomized Newton | https://papers.nips.cc/paper_files/paper/2020/hash/d40d35b3063c11244fbf38e9b55074be-Abstract.html | Michal Derezinski, Feynman T. Liang, Zhenyu Liao, Michael W. Mahoney | https://papers.nips.cc/paper_files/paper/2020/hash/d40d35b3063c11244fbf38e9b55074be-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d40d35b3063c11244fbf38e9b55074be-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11258-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d40d35b3063c11244fbf38e9b55074be-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d40d35b3063c11244fbf38e9b55074be-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d40d35b3063c11244fbf38e9b55074be-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d40d35b3063c11244fbf38e9b55074be-Supplemental.pdf | It is often desirable to reduce the dimensionality of a large dataset
by projecting it onto a low-dimensional subspace. Matrix sketching
has emerged as a powerful technique for performing such dimensionality
reduction very efficiently. Even though there is an extensive
literature on the worst-case performance of sketching, existing
guarantees are typically very different from what is observed in
practice. We exploit recent developments in the spectral analysis of
random matrices to develop novel techniques that provide provably
accurate expressions for the expected value of random projection
matrices obtained via sketching. These expressions can be used to
characterize the performance of dimensionality reduction in a variety
of common machine learning tasks, ranging from low-rank approximation
to iterative stochastic optimization. Our results apply to several
popular sketching methods, including Gaussian and Rademacher
sketches, and they enable precise analysis of these methods in terms of
spectral properties of the data. Empirical results show that the
expressions we derive reflect the practical performance of these
sketching methods, down to lower-order effects and even constant factors. |
The MAGICAL Benchmark for Robust Imitation | https://papers.nips.cc/paper_files/paper/2020/hash/d464b5ac99e74462f321c06ccacc4bff-Abstract.html | Sam Toyer, Rohin Shah, Andrew Critch, Stuart Russell | https://papers.nips.cc/paper_files/paper/2020/hash/d464b5ac99e74462f321c06ccacc4bff-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d464b5ac99e74462f321c06ccacc4bff-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11259-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d464b5ac99e74462f321c06ccacc4bff-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d464b5ac99e74462f321c06ccacc4bff-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d464b5ac99e74462f321c06ccacc4bff-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d464b5ac99e74462f321c06ccacc4bff-Supplemental.pdf | Imitation Learning (IL) algorithms are typically evaluated in the same environment that was used to create demonstrations. This rewards precise reproduction of demonstrations in one particular environment, but provides little information about how robustly an algorithm can generalise the demonstrator's intent to substantially different deployment settings. This paper presents the MAGICAL benchmark suite, which permits systematic evaluation of generalisation by quantifying robustness to different kinds of distribution shift that an IL algorithm is likely to encounter in practice. Using the MAGICAL suite, we confirm that existing IL algorithms overfit significantly to the context in which demonstrations are provided. We also show that standard methods for reducing overfitting are effective at creating narrow perceptual invariances, but are not sufficient to enable transfer to contexts that require substantially different behaviour, which suggests that new approaches will be needed in order to robustly generalise demonstrator intent. Code and data for the MAGICAL suite is available at https://github.com/qxcv/magical/ |
X-CAL: Explicit Calibration for Survival Analysis | https://papers.nips.cc/paper_files/paper/2020/hash/d4a93297083a23cc099f7bd6a8621131-Abstract.html | Mark Goldstein, Xintian Han, Aahlad Puli, Adler Perotte, Rajesh Ranganath | https://papers.nips.cc/paper_files/paper/2020/hash/d4a93297083a23cc099f7bd6a8621131-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d4a93297083a23cc099f7bd6a8621131-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11260-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d4a93297083a23cc099f7bd6a8621131-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d4a93297083a23cc099f7bd6a8621131-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d4a93297083a23cc099f7bd6a8621131-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d4a93297083a23cc099f7bd6a8621131-Supplemental.pdf | Survival analysis models the distribution of time until an event of interest, such as discharge from the hospital or admission to the ICU. When a model’s predicted number of events within any time interval is similar to the observed number, it is called well-calibrated. A survival model’s calibration can be measured using, for instance, distributional calibration (D-CALIBRATION) [Haider et al., 2020] which computes the squared difference between the observed and predicted number of events within different time intervals. Classically, calibration is addressed in post-training analysis. We develop explicit calibration (X-CAL), which turns D-CALIBRATION into a differentiable objective that can be used in survival modeling alongside maximum likelihood estimation and other objectives. X-CAL allows us to directly optimize calibration and strike a desired trade-off between predictive power and calibration. In our experiments, we fit a variety of shallow and deep models on simulated data, a survival dataset based on MNIST, on length-of-stay prediction using MIMIC-III data, and on brain cancer data from The Cancer Genome Atlas. We show that the models we study can be miscalibrated. We give experimental evidence on these datasets that X-CAL improves D-CALIBRATION without a large decrease in concordance or likelihood. |
Decentralized Accelerated Proximal Gradient Descent | https://papers.nips.cc/paper_files/paper/2020/hash/d4b5b5c16df28e61124e13181db7774c-Abstract.html | Haishan Ye, Ziang Zhou, Luo Luo, Tong Zhang | https://papers.nips.cc/paper_files/paper/2020/hash/d4b5b5c16df28e61124e13181db7774c-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d4b5b5c16df28e61124e13181db7774c-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11261-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d4b5b5c16df28e61124e13181db7774c-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d4b5b5c16df28e61124e13181db7774c-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d4b5b5c16df28e61124e13181db7774c-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d4b5b5c16df28e61124e13181db7774c-Supplemental.pdf | Decentralized optimization has wide applications in machine learning, signal processing, and control.
In this paper, we study the decentralized composite
optimization problem with a non-smooth regularization term.
Many proximal gradient based decentralized algorithms have been proposed in the past.
However, these algorithms do not achieve near optimal
computational complexity and communication complexity.
In this paper, we propose a new method which establishes the
optimal computational
complexity and a near optimal communication complexity.
Our empirical study shows that the proposed algorithm outperforms existing state-of-the-art algorithms. |
Making Non-Stochastic Control (Almost) as Easy as Stochastic | https://papers.nips.cc/paper_files/paper/2020/hash/d4ca950da1d6fd954520c45ab19fef1c-Abstract.html | Max Simchowitz | https://papers.nips.cc/paper_files/paper/2020/hash/d4ca950da1d6fd954520c45ab19fef1c-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d4ca950da1d6fd954520c45ab19fef1c-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11262-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d4ca950da1d6fd954520c45ab19fef1c-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d4ca950da1d6fd954520c45ab19fef1c-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d4ca950da1d6fd954520c45ab19fef1c-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d4ca950da1d6fd954520c45ab19fef1c-Supplemental.pdf | Recent literature has made much progress in understanding \emph{online LQR}: a modern learning-theoretic take on the classical control problem where a learner attempts to optimally control an unknown linear dynamical system with fully observed state, perturbed by i.i.d. Gaussian noise. \iftoggle{nips}{The}{It is now understood that the} optimal regret over time horizon $T$ against the optimal control law scales as $\widetilde{\Theta}(\sqrt{T})$. In this paper, we show that the same regret rate (against a suitable benchmark) is attainable even in the considerably more general non-stochastic control model, where the system is driven by \emph{arbitrary adversarial} noise \citep{agarwal2019online}.
We attain the optimal $\widetilde{\mathcal{O}}(\sqrt{T})$ regret when the dynamics are unknown to the learner, and $\mathrm{poly}(\log T)$ regret when known, provided that the cost functions are strongly convex (as in LQR). Our algorithm is based on a novel variant of online Newton step \citep{hazan2007logarithmic}, which adapts to the geometry induced by adversarial disturbances, and our analysis hinges on generic regret bounds for certain structured losses in the OCO-with-memory framework \citep{anava2015online}. |
BERT Loses Patience: Fast and Robust Inference with Early Exit | https://papers.nips.cc/paper_files/paper/2020/hash/d4dd111a4fd973394238aca5c05bebe3-Abstract.html | Wangchunshu Zhou, Canwen Xu, Tao Ge, Julian McAuley, Ke Xu, Furu Wei | https://papers.nips.cc/paper_files/paper/2020/hash/d4dd111a4fd973394238aca5c05bebe3-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d4dd111a4fd973394238aca5c05bebe3-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11263-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d4dd111a4fd973394238aca5c05bebe3-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d4dd111a4fd973394238aca5c05bebe3-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d4dd111a4fd973394238aca5c05bebe3-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d4dd111a4fd973394238aca5c05bebe3-Supplemental.pdf | In this paper, we propose Patience-based Early Exit, a straightforward yet effective inference method that can be used as a plug-and-play technique to simultaneously improve the efficiency and robustness of a pretrained language model (PLM). To achieve this, our approach couples an internal-classifier with each layer of a PLM and dynamically stops inference when the intermediate predictions of the internal classifiers do not change for a pre-defined number of steps. Our approach improves inference efficiency as it allows the model to make a prediction with fewer layers. Meanwhile, experimental results with an ALBERT model show that our method can improve the accuracy and robustness of the model by preventing it from overthinking and exploiting multiple classifiers for prediction, yielding a better accuracy-speed trade-off compared to existing early exit methods. |
Optimal and Practical Algorithms for Smooth and Strongly Convex Decentralized Optimization | https://papers.nips.cc/paper_files/paper/2020/hash/d530d454337fb09964237fecb4bea6ce-Abstract.html | Dmitry Kovalev, Adil Salim, Peter Richtarik | https://papers.nips.cc/paper_files/paper/2020/hash/d530d454337fb09964237fecb4bea6ce-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d530d454337fb09964237fecb4bea6ce-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11264-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d530d454337fb09964237fecb4bea6ce-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d530d454337fb09964237fecb4bea6ce-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d530d454337fb09964237fecb4bea6ce-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d530d454337fb09964237fecb4bea6ce-Supplemental.pdf | We consider the task of decentralized minimization of the sum of smooth strongly convex functions stored across the nodes of a network. For this problem, lower bounds on the number of gradient computations and the number of communication rounds required to achieve $\varepsilon$ accuracy have recently been proven. We propose two new algorithms for this decentralized optimization problem and equip them with complexity guarantees. We show that our first method is optimal both in terms of the number of communication rounds and in terms of the number of gradient computations. Unlike existing optimal algorithms, our algorithm does not rely on the expensive evaluation of dual gradients. Our second algorithm is optimal in terms of the number of communication rounds, without a logarithmic factor. Our approach relies on viewing the two proposed algorithms as accelerated variants of the Forward Backward algorithm to solve monotone inclusions associated with the decentralized optimization problem. We also verify the efficacy of our methods against state-of-the-art algorithms through numerical experiments. |
BAIL: Best-Action Imitation Learning for Batch Deep Reinforcement Learning | https://papers.nips.cc/paper_files/paper/2020/hash/d55cbf210f175f4a37916eafe6c04f0d-Abstract.html | Xinyue Chen, Zijian Zhou, Zheng Wang, Che Wang, Yanqiu Wu, Keith Ross | https://papers.nips.cc/paper_files/paper/2020/hash/d55cbf210f175f4a37916eafe6c04f0d-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d55cbf210f175f4a37916eafe6c04f0d-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11265-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d55cbf210f175f4a37916eafe6c04f0d-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d55cbf210f175f4a37916eafe6c04f0d-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d55cbf210f175f4a37916eafe6c04f0d-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d55cbf210f175f4a37916eafe6c04f0d-Supplemental.pdf | There has recently been a surge in research in batch Deep Reinforcement Learning (DRL), which aims for learning a high-performing policy from a given dataset without additional interactions with the environment. We propose a new algorithm, Best-Action Imitation Learning (BAIL), which strives for both simplicity and performance. BAIL learns a V function, uses the V function to select actions it believes to be high-performing, and then uses those actions to train a policy network using imitation learning. For the MuJoCo benchmark, we provide a comprehensive experimental study of BAIL, comparing its performance to four other batch Q-learning and imitation-learning schemes for a large variety of batch datasets. Our experiments show that BAIL's performance is much higher than the other schemes, and is also computationally much faster than the batch Q-learning schemes. |
Regularizing Towards Permutation Invariance In Recurrent Models | https://papers.nips.cc/paper_files/paper/2020/hash/d58f36f7679f85784d8b010ff248f898-Abstract.html | Edo Cohen-Karlik, Avichai Ben David, Amir Globerson | https://papers.nips.cc/paper_files/paper/2020/hash/d58f36f7679f85784d8b010ff248f898-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d58f36f7679f85784d8b010ff248f898-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11266-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d58f36f7679f85784d8b010ff248f898-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d58f36f7679f85784d8b010ff248f898-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d58f36f7679f85784d8b010ff248f898-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d58f36f7679f85784d8b010ff248f898-Supplemental.pdf | In many machine learning problems the output should not depend on the order of the inputs. Such ``permutation invariant'' functions have been studied extensively recently. Here we argue that temporal architectures such as RNNs are highly relevant for such problems, despite the inherent dependence of RNNs on order. We show that RNNs can be regularized towards permutation invariance, and that this can result in compact models, as compared to non-recursive architectures.
Existing solutions (e.g., DeepSets) mostly suggest restricting the learning problem to hypothesis classes which are permutation invariant by design. Our approach of enforcing permutation invariance via regularization gives rise to learning functions which are "semi permutation invariant", e.g. invariant to some permutations and not to others. Our approach relies on a novel form of stochastic regularization. We demonstrate that our method is beneficial compared to existing permutation invariant methods on synthetic and real world datasets. |
What Did You Think Would Happen? Explaining Agent Behaviour through Intended Outcomes | https://papers.nips.cc/paper_files/paper/2020/hash/d5ab8dc7ef67ca92e41d730982c5c602-Abstract.html | Herman Yau, Chris Russell, Simon Hadfield | https://papers.nips.cc/paper_files/paper/2020/hash/d5ab8dc7ef67ca92e41d730982c5c602-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d5ab8dc7ef67ca92e41d730982c5c602-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11267-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d5ab8dc7ef67ca92e41d730982c5c602-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d5ab8dc7ef67ca92e41d730982c5c602-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d5ab8dc7ef67ca92e41d730982c5c602-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d5ab8dc7ef67ca92e41d730982c5c602-Supplemental.zip | We present a novel form of explanation for Reinforcement Learning, based around the notion of intended outcome. These explanations describe the outcome an agent is trying to achieve by its actions. We provide a simple proof that general methods for post-hoc explanations of this nature are impossible in traditional reinforcement learning. Rather, the information needed for the explanations must be collected in conjunction with training the agent. We derive approaches designed to extract local explanations based on intention for several variants of Q-function approximation and prove consistency between the explanations and the Q-values learned. We demonstrate our method on multiple reinforcement learning problems, and provide code to help researchers introspecting their RL environments and algorithms. |
Batch normalization provably avoids ranks collapse for randomly initialised deep networks | https://papers.nips.cc/paper_files/paper/2020/hash/d5ade38a2c9f6f073d69e1bc6b6e64c1-Abstract.html | Hadi Daneshmand, Jonas Kohler, Francis Bach, Thomas Hofmann, Aurelien Lucchi | https://papers.nips.cc/paper_files/paper/2020/hash/d5ade38a2c9f6f073d69e1bc6b6e64c1-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d5ade38a2c9f6f073d69e1bc6b6e64c1-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11268-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d5ade38a2c9f6f073d69e1bc6b6e64c1-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d5ade38a2c9f6f073d69e1bc6b6e64c1-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d5ade38a2c9f6f073d69e1bc6b6e64c1-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d5ade38a2c9f6f073d69e1bc6b6e64c1-Supplemental.zip | Randomly initialized neural networks are known to become harder to train with increasing depth, unless architectural enhancements like residual connections and batch normalization are used. We here investigate this phenomenon by revisiting the connection between random initialization in deep networks and spectral instabilities in products of random matrices. Given the rich literature on random matrices, it is not surprising to find that the rank of the intermediate representations in unnormalized networks collapses quickly with depth. In this work we highlight the fact that batch normalization is an effective strategy to avoid rank collapse for both linear and ReLU networks. Leveraging tools from Markov chain theory, we derive a meaningful lower rank bound in deep linear networks. Empirically, we also demonstrate that this rank robustness generalizes to ReLU nets. Finally, we conduct an extensive set of experiments on real-world data sets, which confirm that rank stability is indeed a crucial condition for training modern-day deep neural architectures. |
Choice Bandits | https://papers.nips.cc/paper_files/paper/2020/hash/d5fcc35c94879a4afad61cacca56192c-Abstract.html | Arpit Agarwal, Nicholas Johnson, Shivani Agarwal | https://papers.nips.cc/paper_files/paper/2020/hash/d5fcc35c94879a4afad61cacca56192c-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d5fcc35c94879a4afad61cacca56192c-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11269-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d5fcc35c94879a4afad61cacca56192c-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d5fcc35c94879a4afad61cacca56192c-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d5fcc35c94879a4afad61cacca56192c-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d5fcc35c94879a4afad61cacca56192c-Supplemental.pdf | There has been much interest in recent years in the problem of dueling bandits, where on each round the learner plays a pair of arms and receives as feedback the outcome of a relative pairwise comparison between them. Here we study a natural generalization, that we term \emph{choice bandits}, where the learner plays a set of up to $k \geq 2$ arms and receives limited relative feedback in the form of a single multiway choice among the pulled arms, drawn from an underlying multiway choice model. We study choice bandits under a very general class of choice models that is characterized by the existence of a unique `best' arm (which we term generalized Condorcet winner), and includes as special cases the well-studied multinomial logit (MNL) and multinomial probit (MNP) choice models, and more generally, the class of random utility models with i.i.d. noise (IID-RUMs). We propose an algorithm for choice bandits, termed Winner Beats All (WBA), with distribution dependent $O(\log T)$ regret bound under all these choice models. The challenge in our setting is that the decision space is $\Theta(n^k)$, which is large for even moderate $k$. Our algorithm addresses this challenge by extracting just $O(n^2)$ statistics from multiway choices and exploiting the existence of a unique `best' arm to find arms that are competitive to this arm in order to construct sets with low regret. Since these statistics are extracted from the same choice observations, one needs a careful martingale analysis in order to show that these statistics are concentrated. We complement our upper bound result with a lower bound result, which shows that our upper bound is order-wise optimal. Our experiments demonstrate that for the special case of $k=2$, our algorithm is competitive when compared to previous dueling bandit algorithms, and for the more general case $k>2$, outperforms the recently proposed MaxMinUCB algorithm designed for the MNL model. |
What if Neural Networks had SVDs? | https://papers.nips.cc/paper_files/paper/2020/hash/d61e4bbd6393c9111e6526ea173a7c8b-Abstract.html | Alexander Mathiasen, Frederik Hvilshøj, Jakob Rødsgaard Jørgensen, Anshul Nasery, Davide Mottin | https://papers.nips.cc/paper_files/paper/2020/hash/d61e4bbd6393c9111e6526ea173a7c8b-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d61e4bbd6393c9111e6526ea173a7c8b-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11270-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d61e4bbd6393c9111e6526ea173a7c8b-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d61e4bbd6393c9111e6526ea173a7c8b-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d61e4bbd6393c9111e6526ea173a7c8b-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d61e4bbd6393c9111e6526ea173a7c8b-Supplemental.pdf | Various Neural Networks employ time-consuming matrix operations like matrix inversion.
Many such matrix operations are faster to compute given the Singular Value Decomposition (SVD).
Techniques from (Zhang et al., 2018; Mhammedi et al., 2017) allow using the SVD in Neural Networks without computing it.
In theory, the techniques can speed up matrix operations, however, in practice, they are not fast enough.
We present an algorithm that is fast enough to speed up several matrix operations.
The algorithm increases the degree of parallelism of an underlying matrix multiplication H*X where H is an orthogonal matrix represented by a product of Householder matrices. |
A Matrix Chernoff Bound for Markov Chains and Its Application to Co-occurrence Matrices | https://papers.nips.cc/paper_files/paper/2020/hash/d63fbf8c3173730f82b150c5ef38b8ff-Abstract.html | Jiezhong Qiu, Chi Wang, Ben Liao, Richard Peng, Jie Tang | https://papers.nips.cc/paper_files/paper/2020/hash/d63fbf8c3173730f82b150c5ef38b8ff-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d63fbf8c3173730f82b150c5ef38b8ff-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11271-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d63fbf8c3173730f82b150c5ef38b8ff-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d63fbf8c3173730f82b150c5ef38b8ff-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d63fbf8c3173730f82b150c5ef38b8ff-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d63fbf8c3173730f82b150c5ef38b8ff-Supplemental.pdf | Our matrix Chernoff bound for Markov chains can be applied to analyze the behavior of co-occurrence statistics for sequential data, which have been common and important data signals in machine learning. We show that given a regular Markov chain with n states and mixing time t, we need a trajectory of length O(t(log(n) + log(t))/e^2) to achieve an estimator of the co-occurrence matrix with error bound e. We conduct several experiments and the experimental results are consistent with the exponentially fast convergence rate from theoretical analysis. Our result gives the first bound on the convergence rate of the co-occurrence matrix and the first sample complexity analysis in graph representation learning. |
CoMIR: Contrastive Multimodal Image Representation for Registration | https://papers.nips.cc/paper_files/paper/2020/hash/d6428eecbe0f7dff83fc607c5044b2b9-Abstract.html | Nicolas Pielawski, Elisabeth Wetzer, Johan Öfverstedt, Jiahao Lu, Carolina Wählby, Joakim Lindblad, Natasa Sladoje | https://papers.nips.cc/paper_files/paper/2020/hash/d6428eecbe0f7dff83fc607c5044b2b9-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d6428eecbe0f7dff83fc607c5044b2b9-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11272-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d6428eecbe0f7dff83fc607c5044b2b9-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d6428eecbe0f7dff83fc607c5044b2b9-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d6428eecbe0f7dff83fc607c5044b2b9-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d6428eecbe0f7dff83fc607c5044b2b9-Supplemental.zip | We propose contrastive coding to learn shared, dense image representations, referred to as CoMIRs (Contrastive Multimodal Image Representations). CoMIRs enable the registration of multimodal images where existing registration methods often fail due to a lack of sufficiently similar image structures. CoMIRs reduce the multimodal registration problem to a monomodal one, in which general intensity-based, as well as feature-based, registration algorithms can be applied. The method involves training one neural network per modality on aligned images, using a contrastive loss based on noise-contrastive estimation (InfoNCE). Unlike other contrastive coding methods, used for, e.g., classification, our approach generates image-like representations that contain the information shared between modalities. We introduce a novel, hyperparameter-free modification to InfoNCE, to enforce rotational equivariance of the learnt representations, a property essential to the registration task. We assess the extent of achieved rotational equivariance and the stability of the representations with respect to weight initialization, training set, and hyperparameter settings, on a remote sensing dataset of RGB and near-infrared images. We evaluate the learnt representations through registration of a biomedical dataset of bright-field and second-harmonic generation microscopy images; two modalities with very little apparent correlation. The proposed approach based on CoMIRs significantly outperforms registration of representations created by GAN-based image-to-image translation, as well as a state-of-the-art, application-specific method which takes additional knowledge about the data into account. Code is available at: https://github.com/MIDA-group/CoMIR. |
Ensuring Fairness Beyond the Training Data | https://papers.nips.cc/paper_files/paper/2020/hash/d6539d3b57159babf6a72e106beb45bd-Abstract.html | Debmalya Mandal, Samuel Deng, Suman Jana, Jeannette Wing, Daniel J. Hsu | https://papers.nips.cc/paper_files/paper/2020/hash/d6539d3b57159babf6a72e106beb45bd-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d6539d3b57159babf6a72e106beb45bd-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11273-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d6539d3b57159babf6a72e106beb45bd-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d6539d3b57159babf6a72e106beb45bd-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d6539d3b57159babf6a72e106beb45bd-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d6539d3b57159babf6a72e106beb45bd-Supplemental.zip | We initiate the study of fair classifiers that are robust to perturbations in the training distribution. Despite recent progress, the literature on fairness has largely ignored the design of fair and robust classifiers. In this work, we develop classifiers that are fair not only with respect to the training distribution but also for a class of distributions that are weighted perturbations of the training samples. We formulate a min-max objective function whose goal is to minimize a distributionally robust training loss, and at the same time, find a classifier that is fair with respect to a class of distributions. We first reduce this problem to finding a fair classifier that is robust with respect to the class of distributions. Based on an online learning algorithm, we develop an iterative algorithm that provably converges to such a fair and robust solution.
Experiments on standard machine learning fairness datasets suggest that, compared to the state-of-the-art fair classifiers, our classifier retains fairness guarantees and test accuracy for a large class of perturbations on the test set. Furthermore, our experiments show that there is an inherent trade-off between fairness robustness and accuracy of such classifiers. |
How do fair decisions fare in long-term qualification? | https://papers.nips.cc/paper_files/paper/2020/hash/d6d231705f96d5a35aeb3a76402e49a3-Abstract.html | Xueru Zhang, Ruibo Tu, Yang Liu, Mingyan Liu, Hedvig Kjellstrom, Kun Zhang, Cheng Zhang | https://papers.nips.cc/paper_files/paper/2020/hash/d6d231705f96d5a35aeb3a76402e49a3-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d6d231705f96d5a35aeb3a76402e49a3-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11274-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d6d231705f96d5a35aeb3a76402e49a3-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d6d231705f96d5a35aeb3a76402e49a3-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d6d231705f96d5a35aeb3a76402e49a3-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d6d231705f96d5a35aeb3a76402e49a3-Supplemental.zip | Although many fairness criteria have been proposed for decision making, their long-term impact on the well-being of a population remains unclear. In this work, we study the dynamics of population qualification and algorithmic decisions under a partially observed Markov decision problem setting. By characterizing the equilibrium of such dynamics, we analyze the long-term impact of static fairness constraints on the equality and improvement of group well-being. Our results show that static fairness constraints can either promote equality or exacerbate disparity depending on the driving factor of qualification transitions and the effect of sensitive attributes on feature distributions. We also consider possible interventions that can effectively improve group qualification or promote equality of group qualification. Our theoretical results and experiments on static real-world datasets with simulated dynamics show that our framework can be used to facilitate social science studies. |
Pre-training via Paraphrasing | https://papers.nips.cc/paper_files/paper/2020/hash/d6f1dd034aabde7657e6680444ceff62-Abstract.html | Mike Lewis, Marjan Ghazvininejad, Gargi Ghosh, Armen Aghajanyan, Sida Wang, Luke Zettlemoyer | https://papers.nips.cc/paper_files/paper/2020/hash/d6f1dd034aabde7657e6680444ceff62-Abstract.html | NIPS 2020 | null | https://papers.nips.cc/paper_files/paper/11275-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d6f1dd034aabde7657e6680444ceff62-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d6f1dd034aabde7657e6680444ceff62-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d6f1dd034aabde7657e6680444ceff62-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d6f1dd034aabde7657e6680444ceff62-Supplemental.pdf | We introduce MARGE, a pre-trained sequence-to-sequence model learned with an unsupervised multi-lingual multi-document paraphrasing objective. MARGE provides an alternative to the dominant masked language modeling paradigm, where we self-supervise the \emph{reconstruction} of target text by \emph{retrieving} a set of related texts (in many languages) and conditioning on them to maximize the likelihood of generating the original. We show it is possible to jointly learn to do retrieval and reconstruction, given only a random initialization. The objective noisily captures aspects of paraphrase, translation, multi-document summarization, and information retrieval, allowing for strong zero-shot performance on several tasks. For example, with no additional task-specific training we achieve BLEU scores of up to 35.8 for document translation. We further show that fine-tuning gives strong performance on a range of discriminative and generative tasks in many languages, making MARGE the most generally applicable pre-training method to date. |
GCN meets GPU: Decoupling “When to Sample” from “How to Sample” | https://papers.nips.cc/paper_files/paper/2020/hash/d714d2c5a796d5814c565d78dd16188d-Abstract.html | Morteza Ramezani, Weilin Cong, Mehrdad Mahdavi, Anand Sivasubramaniam, Mahmut Kandemir | https://papers.nips.cc/paper_files/paper/2020/hash/d714d2c5a796d5814c565d78dd16188d-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d714d2c5a796d5814c565d78dd16188d-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11276-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d714d2c5a796d5814c565d78dd16188d-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d714d2c5a796d5814c565d78dd16188d-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d714d2c5a796d5814c565d78dd16188d-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d714d2c5a796d5814c565d78dd16188d-Supplemental.pdf | Sampling-based methods promise scalability improvements when paired with stochastic gradient descent in training Graph Convolutional Networks (GCNs). While effective in alleviating the neighborhood explosion, due to bandwidth and memory bottlenecks, these methods lead to computational overheads in preprocessing and loading new samples in heterogeneous systems, which significantly deteriorate the sampling performance. By decoupling the frequency of sampling from the sampling strategy, we propose LazyGCN, a general yet effective framework that can be integrated with any sampling strategy to substantially improve the training time. The basic idea behind LazyGCN is to perform sampling periodically and effectively recycle the sampled nodes to mitigate data preparation overhead. We theoretically analyze the proposed algorithm and show that under a mild condition on the recycling size, by reducing the variance of inner layers, we are able to obtain the same convergence rate as the underlying sampling method. We also give corroborating empirical evidence on large real-world graphs, demonstrating that the proposed schema can significantly reduce the number of sampling steps and yield superior speedup without compromising the accuracy. |
Continual Learning of a Mixed Sequence of Similar and Dissimilar Tasks | https://papers.nips.cc/paper_files/paper/2020/hash/d7488039246a405baf6a7cbc3613a56f-Abstract.html | Zixuan Ke, Bing Liu, Xingchang Huang | https://papers.nips.cc/paper_files/paper/2020/hash/d7488039246a405baf6a7cbc3613a56f-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d7488039246a405baf6a7cbc3613a56f-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11277-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d7488039246a405baf6a7cbc3613a56f-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d7488039246a405baf6a7cbc3613a56f-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d7488039246a405baf6a7cbc3613a56f-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d7488039246a405baf6a7cbc3613a56f-Supplemental.pdf | Existing research on continual learning of a sequence of tasks focused on dealing with catastrophic forgetting, where the tasks are assumed to be dissimilar and have little shared knowledge. Some work has also been done to transfer previously learned knowledge to the new task when the tasks are similar and have shared knowledge. %However, in the most general case, a CL system not only should have the above two capabilities, but also the \textit{backward knowledge transfer} capability so that future tasks may help improve the past models whenever possible. To the best of our knowledge, no technique has been proposed to learn a sequence of mixed similar and dissimilar tasks that can deal with forgetting and also transfer knowledge forward and backward. This paper proposes such a technique to learn both types of tasks in the same network. For dissimilar tasks, the algorithm focuses on dealing with forgetting, and for similar tasks, the algorithm focuses on selectively transferring the knowledge learned from some similar previous tasks to improve the new task learning. Additionally, the algorithm automatically detects whether a new task is similar to any previous tasks. Empirical evaluation using sequences of mixed tasks demonstrates the effectiveness of the proposed model. |
All your loss are belong to Bayes | https://papers.nips.cc/paper_files/paper/2020/hash/d75320797f266ba9ed6dd6dc218cb1b5-Abstract.html | Christian Walder, Richard Nock | https://papers.nips.cc/paper_files/paper/2020/hash/d75320797f266ba9ed6dd6dc218cb1b5-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d75320797f266ba9ed6dd6dc218cb1b5-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11278-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d75320797f266ba9ed6dd6dc218cb1b5-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d75320797f266ba9ed6dd6dc218cb1b5-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d75320797f266ba9ed6dd6dc218cb1b5-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d75320797f266ba9ed6dd6dc218cb1b5-Supplemental.pdf | In this paper, we rely on a broader view of proper composite losses and a recent construct from information geometry, source functions, whose fitting alleviates constraints faced by canonical links. We introduce a trick on squared Gaussian Processes to obtain a random process whose paths are compliant source functions with many desirable properties in the context of link estimation. Experimental results demonstrate substantial improvements over the state of the art. |
HAWQ-V2: Hessian Aware trace-Weighted Quantization of Neural Networks | https://papers.nips.cc/paper_files/paper/2020/hash/d77c703536718b95308130ff2e5cf9ee-Abstract.html | Zhen Dong, Zhewei Yao, Daiyaan Arfeen, Amir Gholami, Michael W. Mahoney, Kurt Keutzer | https://papers.nips.cc/paper_files/paper/2020/hash/d77c703536718b95308130ff2e5cf9ee-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d77c703536718b95308130ff2e5cf9ee-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11279-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d77c703536718b95308130ff2e5cf9ee-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d77c703536718b95308130ff2e5cf9ee-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d77c703536718b95308130ff2e5cf9ee-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d77c703536718b95308130ff2e5cf9ee-Supplemental.pdf | Quantization is an effective method for reducing memory footprint and inference time of Neural Networks. However, ultra low precision quantization could lead to significant degradation in model accuracy. A promising method to address this is to perform mixed-precision quantization, where more sensitive layers are kept at higher precision. However, the search space for a mixed-precision quantization is exponential in the number of layers. Recent work has proposed a novel Hessian based framework, with the aim of reducing this exponential search space by using second-order information. While promising, this prior work has three major limitations: (i) they only use a heuristic metric based on top Hessian eigenvalue as a measure of sensitivity and do not consider the rest of the Hessian spectrum; (ii) their approach only provides relative sensitivity of different layers and therefore requires a manual selection of the mixed-precision setting; and (iii) they do not consider mixed-precision activation quantization. Here, we present HAWQ-V2 which addresses these shortcomings. For (i), we theoretically prove that the right sensitivity metric is the average Hessian trace, instead of just top Hessian eigenvalue. For (ii), we develop a Pareto frontier based method for automatic bit precision selection of different layers without any manual intervention. For (iii), we develop the first Hessian based analysis for mixed-precision activation quantization, which is very beneficial for object detection. We show that HAWQ-V2 achieves new state-of-the-art results for a wide range of tasks. In particular, we present quantization results for InceptionV3, ResNet50, and SqueezeNext, all without any manual bit selection. Furthermore, we present results for object detection on Microsoft COCO, where we achieve 2.6 higher mAP than direct uniform quantization and 1.6 higher mAP than the recently proposed method of FQN, with a smaller model size of 17.9MB. |
Sample-Efficient Reinforcement Learning of Undercomplete POMDPs | https://papers.nips.cc/paper_files/paper/2020/hash/d783823cc6284b929c2cd8df2167d212-Abstract.html | Chi Jin, Sham Kakade, Akshay Krishnamurthy, Qinghua Liu | https://papers.nips.cc/paper_files/paper/2020/hash/d783823cc6284b929c2cd8df2167d212-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d783823cc6284b929c2cd8df2167d212-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11280-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d783823cc6284b929c2cd8df2167d212-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d783823cc6284b929c2cd8df2167d212-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d783823cc6284b929c2cd8df2167d212-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d783823cc6284b929c2cd8df2167d212-Supplemental.pdf | Partial observability is a common challenge in many reinforcement learning applications, which requires an agent to maintain memory, infer latent states, and integrate this past information into exploration. This challenge leads to a number of computational and statistical hardness results for learning general Partially Observable Markov Decision Processes (POMDPs). This work shows that these hardness barriers do not preclude efficient reinforcement learning for rich and interesting subclasses of POMDPs. In particular, we present a sample-efficient algorithm, OOM-UCB, for episodic finite undercomplete POMDPs, where the number of observations is larger than the number of latent states and where exploration is essential for learning, thus distinguishing our results from prior works. OOM-UCB achieves an optimal sample complexity of $\tilde{\mathcal{O}}(1/\varepsilon^2)$ for finding an $\varepsilon$-optimal policy, along with being polynomial in all other relevant quantities. As an interesting special case, we also provide a computationally and statistically efficient algorithm for POMDPs with deterministic state transitions. |
Non-Convex SGD Learns Halfspaces with Adversarial Label Noise | https://papers.nips.cc/paper_files/paper/2020/hash/d785bf9067f8af9e078b93cf26de2b54-Abstract.html | Ilias Diakonikolas, Vasilis Kontonis, Christos Tzamos, Nikos Zarifis | https://papers.nips.cc/paper_files/paper/2020/hash/d785bf9067f8af9e078b93cf26de2b54-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d785bf9067f8af9e078b93cf26de2b54-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11281-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d785bf9067f8af9e078b93cf26de2b54-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d785bf9067f8af9e078b93cf26de2b54-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d785bf9067f8af9e078b93cf26de2b54-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d785bf9067f8af9e078b93cf26de2b54-Supplemental.pdf | We study the problem of agnostically learning homogeneous halfspaces in the distribution-specific PAC model.
For a broad family of structured distributions, including log-concave distributions, we show that non-convex SGD
efficiently converges to a solution with misclassification error $O(\opt)+\eps$, where $\opt$ is the misclassification
error of the best-fitting halfspace. In sharp contrast, we show that optimizing any convex surrogate inherently
leads to misclassification error of $\omega(\opt)$, even under Gaussian marginals.
|
A Tight Lower Bound and Efficient Reduction for Swap Regret | https://papers.nips.cc/paper_files/paper/2020/hash/d79c8788088c2193f0244d8f1f36d2db-Abstract.html | Shinji Ito | https://papers.nips.cc/paper_files/paper/2020/hash/d79c8788088c2193f0244d8f1f36d2db-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d79c8788088c2193f0244d8f1f36d2db-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11282-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d79c8788088c2193f0244d8f1f36d2db-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d79c8788088c2193f0244d8f1f36d2db-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d79c8788088c2193f0244d8f1f36d2db-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d79c8788088c2193f0244d8f1f36d2db-Supplemental.pdf | Swap regret, a generic performance measure of online decision-making algorithms, plays an important role in the theory of repeated games, along with a close connection to correlated equilibria in strategic games. This paper shows an $\Omega( \sqrt{T N\log{N}} )$-lower bound for swap regret, where $T$ and $N$ denote the numbers of time steps and available actions, respectively. Our lower bound is tight up to a constant, and resolves an open problem mentioned, e.g., in the book by Nisan et al. (2007). Besides, we present a computationally efficient reduction method that converts no-external-regret algorithms to no-swap-regret algorithms. This method can be applied not only to the full-information setting but also to the bandit setting and provides a better regret bound than previous results. |
DisCor: Corrective Feedback in Reinforcement Learning via Distribution Correction | https://papers.nips.cc/paper_files/paper/2020/hash/d7f426ccbc6db7e235c57958c21c5dfa-Abstract.html | Aviral Kumar, Abhishek Gupta, Sergey Levine | https://papers.nips.cc/paper_files/paper/2020/hash/d7f426ccbc6db7e235c57958c21c5dfa-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d7f426ccbc6db7e235c57958c21c5dfa-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11283-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d7f426ccbc6db7e235c57958c21c5dfa-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d7f426ccbc6db7e235c57958c21c5dfa-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d7f426ccbc6db7e235c57958c21c5dfa-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d7f426ccbc6db7e235c57958c21c5dfa-Supplemental.pdf | Deep reinforcement learning can learn effective policies for a wide range of tasks, but is notoriously difficult to use due to instability and sensitivity to hyperparameters. The reasons for this remain unclear. In this paper, we study how RL methods based on bootstrapping-based Q-learning can suffer from a pathological interaction between function approximation and the data distribution used to train the Q-function: with standard supervised learning, online data collection should induce corrective feedback, where new data corrects mistakes in old predictions. With dynamic programming methods like Q-learning, such feedback may be absent. This can lead to potential instability, sub-optimal convergence, and poor results when learning from noisy, sparse or delayed rewards. Based on these observations, we propose a new algorithm, DisCor, which explicitly optimizes for data distributions that can correct for accumulated errors in the value function. DisCor computes a tractable approximation to the distribution that optimally induces corrective feedback, which we show results in reweighting samples based on the estimated accuracy of their target values. Using this distribution for training, DisCor results in substantial improvements in a range of challenging RL settings, such as multi-task learning and learning from noisy reward signals. |
OTLDA: A Geometry-aware Optimal Transport Approach for Topic Modeling | https://papers.nips.cc/paper_files/paper/2020/hash/d800149d2f947ad4d64f34668f8b20f6-Abstract.html | Viet Huynh, He Zhao, Dinh Phung | https://papers.nips.cc/paper_files/paper/2020/hash/d800149d2f947ad4d64f34668f8b20f6-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d800149d2f947ad4d64f34668f8b20f6-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11284-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d800149d2f947ad4d64f34668f8b20f6-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d800149d2f947ad4d64f34668f8b20f6-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d800149d2f947ad4d64f34668f8b20f6-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d800149d2f947ad4d64f34668f8b20f6-Supplemental.pdf | We present an optimal transport framework for learning topics from textual data. While the celebrated Latent Dirichlet allocation (LDA) topic model and its variants have been applied to many disciplines, they mainly focus on word-occurrences and neglect to incorporate semantic regularities in language. Even though recent works have tried to exploit the semantic relationship between words to bridge this gap, however, these models which are usually extensions of LDA or Dirichlet Multinomial mixture (DMM) are tailored to deal effectively with either regular or short documents. The optimal transport distance provides an appealing tool to incorporate the geometry of word semantics into it. Moreover, recent developments on efficient computation of optimal transport distance also promote its application in topic modeling. In this paper we ground on optimal transport theory to naturally exploit the geometric structures of semantically related words in embedding spaces which leads to more interpretable learned topics. Comprehensive experiments illustrate that the proposed framework outperforms competitive approaches in terms of topic coherence on assorted text corpora which include both long and short documents. The representation of learned topic also leads to better accuracy on classification downstream tasks, which is considered as an extrinsic evaluation. |
Measuring Robustness to Natural Distribution Shifts in Image Classification | https://papers.nips.cc/paper_files/paper/2020/hash/d8330f857a17c53d217014ee776bfd50-Abstract.html | Rohan Taori, Achal Dave, Vaishaal Shankar, Nicholas Carlini, Benjamin Recht, Ludwig Schmidt | https://papers.nips.cc/paper_files/paper/2020/hash/d8330f857a17c53d217014ee776bfd50-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d8330f857a17c53d217014ee776bfd50-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11285-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d8330f857a17c53d217014ee776bfd50-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d8330f857a17c53d217014ee776bfd50-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d8330f857a17c53d217014ee776bfd50-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d8330f857a17c53d217014ee776bfd50-Supplemental.pdf | We study how robust current ImageNet models are to distribution shifts arising from natural variations in datasets. Most research on robustness focuses on synthetic image perturbations (noise, simulated weather artifacts, adversarial examples, etc.), which leaves open how robustness on synthetic distribution shift relates to distribution shift arising in real data. Informed by an evaluation of 204 ImageNet models in 213 different test conditions, we find that there is often little to no transfer of robustness from current synthetic to natural distribution shift. Moreover, most current techniques provide no robustness to the natural distribution shifts in our testbed. The main exception is training on larger and more diverse datasets, which in multiple cases increases robustness, but is still far from closing the performance gaps. Our results indicate that distribution shifts arising in real data are currently an open research problem. |
Can I Trust My Fairness Metric? Assessing Fairness with Unlabeled Data and Bayesian Inference | https://papers.nips.cc/paper_files/paper/2020/hash/d83de59e10227072a9c034ce10029c39-Abstract.html | Disi Ji, Padhraic Smyth, Mark Steyvers | https://papers.nips.cc/paper_files/paper/2020/hash/d83de59e10227072a9c034ce10029c39-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d83de59e10227072a9c034ce10029c39-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11286-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d83de59e10227072a9c034ce10029c39-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d83de59e10227072a9c034ce10029c39-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d83de59e10227072a9c034ce10029c39-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d83de59e10227072a9c034ce10029c39-Supplemental.zip | Group fairness is measured via parity of quantitative metrics across different protected demographic groups. In this paper, we investigate the problem of reliably assessing group fairness metrics when labeled examples are few but unlabeled examples are plentiful. We propose a general Bayesian framework that can augment labeled data with unlabeled data to produce more accurate and lower-variance estimates compared to methods based on labeled data alone. Our approach estimates calibrated scores (for unlabeled examples) of each group using a hierarchical latent variable model conditioned on labeled examples. This in turn allows for inference of posterior distributions for an array of group fairness metrics with a notion of uncertainty. We demonstrate that our approach leads to significant and consistent reductions in estimation error across multiple well-known fairness datasets, sensitive attributes, and predictive models. The results clearly show the benefits of using both unlabeled data and Bayesian inference in assessing whether a prediction model is fair or not. |
RandAugment: Practical Automated Data Augmentation with a Reduced Search Space | https://papers.nips.cc/paper_files/paper/2020/hash/d85b63ef0ccb114d0a3bb7b7d808028f-Abstract.html | Ekin Dogus Cubuk, Barret Zoph, Jon Shlens, Quoc Le | https://papers.nips.cc/paper_files/paper/2020/hash/d85b63ef0ccb114d0a3bb7b7d808028f-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d85b63ef0ccb114d0a3bb7b7d808028f-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11287-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d85b63ef0ccb114d0a3bb7b7d808028f-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d85b63ef0ccb114d0a3bb7b7d808028f-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d85b63ef0ccb114d0a3bb7b7d808028f-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d85b63ef0ccb114d0a3bb7b7d808028f-Supplemental.pdf | Recent work on automated data augmentation strategies has led to state-of-the-art results in image classification and object detection. An obstacle to a large-scale adoption of these methods is that they require a separate and expensive search phase. A common way to overcome the expense of the search phase was to use a smaller proxy task. However, it was not clear if the optimized hyperparameters found on the proxy task are also optimal for the actual task. In this work, we rethink the process of designing automated data augmentation strategies. We find that while previous work required searching for many augmentation parameters (e.g. magnitude and probability) independently for each augmentation operation, it is sufficient to only search for a single parameter that jointly controls all operations. Hence, we propose a search space that is vastly smaller (e.g. from 10^32 to 10^2 potential candidates). The smaller search space significantly reduces the computational expense of automated data augmentation and permits the removal of a separate proxy task. Despite the simplifications, our method achieves state-of-the-art performance on CIFAR-10, SVHN, and ImageNet. On EfficientNet-B7, we achieve 84.7% accuracy, a 1.0% increase over baseline augmentation and a 0.4% improvement over AutoAugment on the ImageNet dataset. On object detection, the same method used for classification leads to 1.0-1.3% improvement over the baseline augmentation method on COCO. Code is available online. |
Asymptotic normality and confidence intervals for derivatives of 2-layers neural network in the random features model | https://papers.nips.cc/paper_files/paper/2020/hash/d87ca511e2a8593c8039ef732f5bffed-Abstract.html | Yiwei Shen, Pierre C Bellec | https://papers.nips.cc/paper_files/paper/2020/hash/d87ca511e2a8593c8039ef732f5bffed-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d87ca511e2a8593c8039ef732f5bffed-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11288-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d87ca511e2a8593c8039ef732f5bffed-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d87ca511e2a8593c8039ef732f5bffed-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d87ca511e2a8593c8039ef732f5bffed-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d87ca511e2a8593c8039ef732f5bffed-Supplemental.zip | This paper studies two-layers Neural Networks (NN), where the first layer contains random weights, and the second layer is trained using Ridge regularization. This model has been the focus of numerous recent works, showing that despite its simplicity, it captures some of the empirically observed behaviors of NN in the overparametrized regime, such as the double-descent curve where the generalization error decreases as the number of weights increases to $+\infty$.
This paper establishes asymptotic distribution results for this 2-layers NN model in the regime where the ratios $\frac p n$ and $\frac d n$ have finite limits, where $n$ is the sample size, $p$ the ambient dimension and $d$ is the width of the first layer. We show that a weighted average of the derivatives of the trained NN at the observed data is asymptotically normal, in a setting with Lipschitz activation functions in a linear regression response with Gaussian features under possibly non-linear perturbations. We then leverage this asymptotic normality result to construct confidence intervals (CIs) for single components of the unknown regression vector.
The novelty of our results are threefold:
(1) Despite the nonlinearity induced by the activation function, we characterize the asymptotic distribution of a weighted average of the gradients of the network after training;
(2) It provides the first frequentist uncertainty quantification guarantees, in the form of valid ($1\text{-}\alpha$)-CIs, based on NN estimates;
(3) It shows that the double-descent phenomenon occurs in terms of the length of the CIs, with the length increasing and then decreasing as $\frac d n\nearrow +\infty$ for certain fixed values of $\frac p n$. We also provide a toolbox to predict the length of CIs numerically, which lets us compare activation functions and other parameters in terms of CI length. |
DisARM: An Antithetic Gradient Estimator for Binary Latent Variables | https://papers.nips.cc/paper_files/paper/2020/hash/d880e783834172e5ebd1868d84463d93-Abstract.html | Zhe Dong, Andriy Mnih, George Tucker | https://papers.nips.cc/paper_files/paper/2020/hash/d880e783834172e5ebd1868d84463d93-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d880e783834172e5ebd1868d84463d93-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11289-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d880e783834172e5ebd1868d84463d93-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d880e783834172e5ebd1868d84463d93-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d880e783834172e5ebd1868d84463d93-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d880e783834172e5ebd1868d84463d93-Supplemental.pdf | Training models with discrete latent variables is challenging due to the difficulty of estimating the gradients accurately. Much of the recent progress has been achieved by taking advantage of continuous relaxations of the system, which are not always available or even possible. The Augment-REINFORCE-Merge (ARM) estimator provides an alternative that, instead of relaxation, uses continuous augmentation. Applying antithetic sampling over the augmenting variables yields a relatively low-variance and unbiased estimator applicable to any model with binary latent variables. However, while antithetic sampling reduces variance, the augmentation process increases variance. We show that ARM can be improved by analytically integrating out the randomness introduced by the augmentation process, guaranteeing substantial variance reduction. Our estimator, DisARM, is simple to implement and has the same computational cost as ARM. We evaluate DisARM on several generative modeling benchmarks and show that it consistently outperforms ARM and a strong independent sample baseline in terms of both variance and log-likelihood. Furthermore, we propose a local version of DisARM designed for optimizing the multi-sample variational bound, and show that it outperforms VIMCO, the current state-of-the-art method. |
Variational Inference for Graph Convolutional Networks in the Absence of Graph Data and Adversarial Settings | https://papers.nips.cc/paper_files/paper/2020/hash/d882050bb9eeba930974f596931be527-Abstract.html | Pantelis Elinas, Edwin V. Bonilla, Louis Tiao | https://papers.nips.cc/paper_files/paper/2020/hash/d882050bb9eeba930974f596931be527-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d882050bb9eeba930974f596931be527-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11290-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d882050bb9eeba930974f596931be527-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d882050bb9eeba930974f596931be527-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d882050bb9eeba930974f596931be527-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d882050bb9eeba930974f596931be527-Supplemental.pdf | We propose a framework that lifts the capabilities of graph convolutional networks (GCNs) to scenarios where no input graph is given and increases their robustness to adversarial attacks. We formulate a joint probabilistic model that considers a prior distribution over graphs along with a GCN-based likelihood and develop a stochastic variational inference algorithm to estimate the graph posterior and the GCN parameters jointly. To address the problem of propagating gradients through latent variables drawn from discrete distributions, we use their continuous relaxations known as Concrete distributions. We show that, on real datasets, our approach can outperform state-of-the-art Bayesian and non-Bayesian graph neural network algorithms on the task of semi-supervised classification in the absence of graph data and when the network structure is subjected to adversarial perturbations. |
Supervised Contrastive Learning | https://papers.nips.cc/paper_files/paper/2020/hash/d89a66c7c80a29b1bdbab0f2a1a94af8-Abstract.html | Prannay Khosla, Piotr Teterwak, Chen Wang, Aaron Sarna, Yonglong Tian, Phillip Isola, Aaron Maschinot, Ce Liu, Dilip Krishnan | https://papers.nips.cc/paper_files/paper/2020/hash/d89a66c7c80a29b1bdbab0f2a1a94af8-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d89a66c7c80a29b1bdbab0f2a1a94af8-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11291-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d89a66c7c80a29b1bdbab0f2a1a94af8-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d89a66c7c80a29b1bdbab0f2a1a94af8-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d89a66c7c80a29b1bdbab0f2a1a94af8-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d89a66c7c80a29b1bdbab0f2a1a94af8-Supplemental.pdf | Contrastive learning applied to self-supervised representation learning has seen a resurgence in recent years, leading to state of the art performance in the unsupervised training of deep image models. Modern batch contrastive approaches subsume or significantly outperform traditional contrastive losses such as triplet, max-margin and the N-pairs loss. In this work, we extend the self-supervised batch contrastive approach to the fully-supervised setting, allowing us to effectively leverage label information. Clusters of points belonging to the same class are pulled together in embedding space, while simultaneously pushing apart clusters of samples from different classes. We analyze two possible versions of the supervised contrastive (SupCon) loss, identifying the best-performing formulation of the loss. On ResNet-200, we achieve top-1 accuracy of 81.4% on the ImageNet dataset, which is 0.8% above the best number reported for this architecture. We show consistent outperformance over cross-entropy on other datasets and two ResNet variants. The loss shows benefits for robustness to natural corruptions, and is more stable to hyperparameter settings such as optimizers and data augmentations. In reduced data settings, it outperforms cross-entropy significantly. Our loss function is simple to implement and reference TensorFlow code is released at https://t.ly/supcon. |
Learning Optimal Representations with the Decodable Information Bottleneck | https://papers.nips.cc/paper_files/paper/2020/hash/d8ea5f53c1b1eb087ac2e356253395d8-Abstract.html | Yann Dubois, Douwe Kiela, David J. Schwab, Ramakrishna Vedantam | https://papers.nips.cc/paper_files/paper/2020/hash/d8ea5f53c1b1eb087ac2e356253395d8-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d8ea5f53c1b1eb087ac2e356253395d8-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11292-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d8ea5f53c1b1eb087ac2e356253395d8-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d8ea5f53c1b1eb087ac2e356253395d8-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d8ea5f53c1b1eb087ac2e356253395d8-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d8ea5f53c1b1eb087ac2e356253395d8-Supplemental.pdf | We address the question of characterizing and finding optimal representations for supervised learning. Traditionally, this question has been tackled using the Information Bottleneck, which compresses the inputs while retaining information about the targets, in a decoder-agnostic fashion. In machine learning, however, our goal is not compression but rather generalization, which is intimately linked to the predictive family or decoder of interest (e.g. linear classifier). We propose the Decodable Information Bottleneck (DIB) that considers information retention and compression from the perspective of the desired predictive family. As a result, DIB gives rise to representations that are optimal in terms of expected test performance and can be estimated with guarantees. Empirically, we show that the framework can be used to enforce a small generalization gap on downstream classifiers and to predict the generalization ability of neural networks. |
Meta-trained agents implement Bayes-optimal agents | https://papers.nips.cc/paper_files/paper/2020/hash/d902c3ce47124c66ce615d5ad9ba304f-Abstract.html | Vladimir Mikulik, Grégoire Delétang, Tom McGrath, Tim Genewein, Miljan Martic, Shane Legg, Pedro Ortega | https://papers.nips.cc/paper_files/paper/2020/hash/d902c3ce47124c66ce615d5ad9ba304f-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d902c3ce47124c66ce615d5ad9ba304f-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11293-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d902c3ce47124c66ce615d5ad9ba304f-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d902c3ce47124c66ce615d5ad9ba304f-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d902c3ce47124c66ce615d5ad9ba304f-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d902c3ce47124c66ce615d5ad9ba304f-Supplemental.pdf | Memory-based meta-learning is a powerful technique to build agents that adapt fast to any task within a target distribution. A previous theoretical study has argued that this remarkable performance is because the meta-training protocol incentivises agents to behave Bayes-optimally. We empirically investigate this claim on a number of prediction and bandit tasks. Inspired by ideas from theoretical computer science, we show that meta-learned and Bayes-optimal agents not only behave alike, but they even share a similar computational structure, in the sense that one agent system can approximately simulate the other. Furthermore, we show that Bayes-optimal agents are fixed points of the meta-learning dynamics. Our results suggest that memory-based meta-learning is a general technique for numerically approximating Bayes-optimal agents; that is, even for task distributions for which we currently don't possess tractable models. |
Learning Agent Representations for Ice Hockey | https://papers.nips.cc/paper_files/paper/2020/hash/d90e5b6628b4291225cba0bdc643c295-Abstract.html | Guiliang Liu, Oliver Schulte, Pascal Poupart, Mike Rudd, Mehrsan Javan | https://papers.nips.cc/paper_files/paper/2020/hash/d90e5b6628b4291225cba0bdc643c295-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d90e5b6628b4291225cba0bdc643c295-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11294-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d90e5b6628b4291225cba0bdc643c295-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d90e5b6628b4291225cba0bdc643c295-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d90e5b6628b4291225cba0bdc643c295-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d90e5b6628b4291225cba0bdc643c295-Supplemental.zip | Team sports is a new application domain for agent modeling with high real-world impact. A fundamental challenge for modeling professional players is their large number (over 1K), which includes many bench players with sparse participation in a game season. The diversity and sparsity of player observations make it difficult to extend previous agent representation models to the sports domain. This paper develops a new approach for agent representations, based on a Markov game model, that is tailored towards applications in professional ice hockey. We introduce a novel player representation via player generation framework where a variational encoder embeds player information with latent variables. The encoder learns a context-specific shared prior to induce a shrinkage effect for the posterior player representations, allowing it to share statistical information across players with different participations. To model the play dynamics in sequential sports data, we design a Variational Recurrent Ladder Agent Encoder (VaRLAE). It learns a contextualized player representation with a hierarchy of latent variables that effectively prevents latent posterior collapse. We validate our player representations in major sports analytics tasks. Our experimental results, based on a large dataset that contains over 4.5M events, show state-of-the-art performance for our VarLAE on facilitating 1) identifying the acting player, 2) estimating expected goals, and 3) predicting the final score difference. |
Weak Form Generalized Hamiltonian Learning | https://papers.nips.cc/paper_files/paper/2020/hash/d93c96e6a23fff65b91b900aaa541998-Abstract.html | Kevin Course, Trefor Evans, Prasanth Nair | https://papers.nips.cc/paper_files/paper/2020/hash/d93c96e6a23fff65b91b900aaa541998-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d93c96e6a23fff65b91b900aaa541998-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11295-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d93c96e6a23fff65b91b900aaa541998-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d93c96e6a23fff65b91b900aaa541998-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d93c96e6a23fff65b91b900aaa541998-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d93c96e6a23fff65b91b900aaa541998-Supplemental.pdf | We present a method for learning generalized Hamiltonian decompositions of ordinary differential equations given a set of noisy time series measurements. Our method simultaneously learns a continuous time model and a scalar energy function for a general dynamical system. Learning predictive models in this form allows one to place strong, high-level, physics inspired priors onto the form of the learnt governing equations for general dynamical systems. Moreover, having shown how our method extends and unifies some previous work in deep learning with physics inspired priors, we present a novel method for learning continuous time models from the weak form of the governing equations which is less computationally taxing than standard adjoint methods. |
Neural Non-Rigid Tracking | https://papers.nips.cc/paper_files/paper/2020/hash/d93ed5b6db83be78efb0d05ae420158e-Abstract.html | Aljaz Bozic, Pablo Palafox, Michael Zollhöfer, Angela Dai, Justus Thies, Matthias Niessner | https://papers.nips.cc/paper_files/paper/2020/hash/d93ed5b6db83be78efb0d05ae420158e-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d93ed5b6db83be78efb0d05ae420158e-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11296-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d93ed5b6db83be78efb0d05ae420158e-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d93ed5b6db83be78efb0d05ae420158e-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d93ed5b6db83be78efb0d05ae420158e-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d93ed5b6db83be78efb0d05ae420158e-Supplemental.zip | We introduce a novel, end-to-end learnable, differentiable non-rigid tracker that enables state-of-the-art non-rigid reconstruction by a learned robust optimization. Given two input RGB-D frames of a non-rigidly moving object, we employ a convolutional neural network to predict dense correspondences and their confidences. These correspondences are used as constraints in an as-rigid-as-possible (ARAP) optimization problem. By enabling gradient back-propagation through the weighted non-linear least squares solver, we are able to learn correspondences and confidences in an end-to-end manner such that they are optimal for the task of non-rigid tracking. Under this formulation, correspondence confidences can be learned via self-supervision, informing a learned robust optimization, where outliers and wrong correspondences are automatically down-weighted to enable effective tracking. Compared to state-of-the-art approaches, our algorithm shows improved reconstruction performance, while simultaneously achieving 85 times faster correspondence prediction than comparable deep-learning based methods. |
Collegial Ensembles | https://papers.nips.cc/paper_files/paper/2020/hash/d958628e70134d9e1e17499a9d815a71-Abstract.html | Etai Littwin, Ben Myara, Sima Sabah, Joshua Susskind, Shuangfei Zhai, Oren Golan | https://papers.nips.cc/paper_files/paper/2020/hash/d958628e70134d9e1e17499a9d815a71-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d958628e70134d9e1e17499a9d815a71-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11297-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d958628e70134d9e1e17499a9d815a71-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d958628e70134d9e1e17499a9d815a71-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d958628e70134d9e1e17499a9d815a71-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d958628e70134d9e1e17499a9d815a71-Supplemental.pdf | Modern neural network performance typically improves as model size increases. A recent line of research on the Neural Tangent Kernel (NTK) of over-parameterized networks indicates that the improvement with size increase is a product of a better conditioned loss landscape. In this work, we investigate a form of over-parameterization achieved through ensembling, where we define collegial ensembles (CE) as the aggregation of multiple independent models with identical architectures, trained as a single model. We show that the optimization dynamics of CE simplify dramatically when the number of models in the ensemble is large, resembling the dynamics of wide models, yet scale much more favorably. We use recent theoretical results on the finite width corrections of the NTK to perform efficient architecture search in a space of finite width CE that aims to either minimize capacity, or maximize trainability under a set of constraints. The resulting ensembles can be efficiently implemented in practical architectures using group convolutions and block diagonal layers. Finally, we show how our framework can be used to analytically derive optimal group convolution modules originally found using expensive grid searches, without having to train a single model. |
ICNet: Intra-saliency Correlation Network for Co-Saliency Detection | https://papers.nips.cc/paper_files/paper/2020/hash/d961e9f236177d65d21100592edb0769-Abstract.html | Wen-Da Jin, Jun Xu, Ming-Ming Cheng, Yi Zhang, Wei Guo | https://papers.nips.cc/paper_files/paper/2020/hash/d961e9f236177d65d21100592edb0769-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d961e9f236177d65d21100592edb0769-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11298-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d961e9f236177d65d21100592edb0769-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d961e9f236177d65d21100592edb0769-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d961e9f236177d65d21100592edb0769-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d961e9f236177d65d21100592edb0769-Supplemental.pdf | Intra-saliency and inter-saliency cues have been extensively studied for co-saliency detection (Co-SOD). Model-based methods produce coarse Co-SOD results due to hand-crafted intra- and inter-saliency features. Current data-driven models exploit inter-saliency cues, but undervalue the potential power of intra-saliency cues. In this paper, we propose an Intra-saliency Correlation Network (ICNet) to extract intra-saliency cues from the single image saliency maps (SISMs) predicted by any off-the-shelf SOD method, and obtain inter-saliency cues by correlation techniques. Specifically, we adopt normalized masked average pooling (NMAP) to extract latent intra-saliency categories from the SISMs and semantic features as intra cues. Then we employ a correlation fusion module (CFM) to obtain inter cues by exploiting correlations between the intra cues and single-image features. To improve Co-SOD performance, we propose a category-independent rearranged self-correlation feature (RSCF) strategy. Experiments on three benchmarks show that our ICNet outperforms previous state-of-the-art methods on Co-SOD. Ablation studies validate the effectiveness of our contributions. The PyTorch code is available at https://github.com/blanclist/ICNet. |
Improved Variational Bayesian Phylogenetic Inference with Normalizing Flows | https://papers.nips.cc/paper_files/paper/2020/hash/d96409bf894217686ba124d7356686c9-Abstract.html | Cheng Zhang | https://papers.nips.cc/paper_files/paper/2020/hash/d96409bf894217686ba124d7356686c9-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d96409bf894217686ba124d7356686c9-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11299-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d96409bf894217686ba124d7356686c9-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d96409bf894217686ba124d7356686c9-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d96409bf894217686ba124d7356686c9-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d96409bf894217686ba124d7356686c9-Supplemental.pdf | Variational Bayesian phylogenetic inference (VBPI) provides a promising general variational framework for efficient estimation of phylogenetic posteriors. However, the current diagonal Lognormal branch length approximation would significantly restrict the quality of the approximating distributions. In this paper, we propose a new type of VBPI, VBPI-NF, as a first step to empower phylogenetic posterior estimation with deep learning techniques. By handling the non-Euclidean branch length space of phylogenetic models with carefully designed permutation equivariant transformations, VBPI-NF uses normalizing flows to provide a rich family of flexible branch length distributions that generalize across different tree topologies. We show that VBPI-NF significantly improves upon the vanilla VBPI on a benchmark of challenging real data Bayesian phylogenetic inference problems. Further investigation also reveals that the structured parameterization in those permutation equivariant transformations can provide additional amortization benefit. |
Deep Metric Learning with Spherical Embedding | https://papers.nips.cc/paper_files/paper/2020/hash/d9812f756d0df06c7381945d2e2c7d4b-Abstract.html | Dingyi Zhang, Yingming Li, Zhongfei Zhang | https://papers.nips.cc/paper_files/paper/2020/hash/d9812f756d0df06c7381945d2e2c7d4b-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d9812f756d0df06c7381945d2e2c7d4b-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11300-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d9812f756d0df06c7381945d2e2c7d4b-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d9812f756d0df06c7381945d2e2c7d4b-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d9812f756d0df06c7381945d2e2c7d4b-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d9812f756d0df06c7381945d2e2c7d4b-Supplemental.pdf | Deep metric learning has attracted much attention in recent years, due to seamlessly combining the distance metric learning and deep neural network. Many endeavors are devoted to design different pair-based angular loss functions, which decouple the magnitude and direction information for embedding vectors and ensure the training and testing measure consistency. However, these traditional angular losses cannot guarantee that all the sample embeddings are on the surface of the same hypersphere during the training stage, which would result in unstable gradient in batch optimization and may influence the quick convergence of the embedding learning. In this paper, we first investigate the effect of the embedding norm for deep metric learning with angular distance, and then propose a spherical embedding constraint (SEC) to regularize the distribution of the norms. SEC adaptively adjusts the embeddings to fall on the same hypersphere and performs more balanced direction update. Extensive experiments on deep metric learning, face recognition, and contrastive self-supervised learning show that the SEC-based angular space learning strategy significantly improves the performance of the state-of-the-art. |
Preference-based Reinforcement Learning with Finite-Time Guarantees | https://papers.nips.cc/paper_files/paper/2020/hash/d9d3837ee7981e8c064774da6cdd98bf-Abstract.html | Yichong Xu, Ruosong Wang, Lin Yang, Aarti Singh, Artur Dubrawski | https://papers.nips.cc/paper_files/paper/2020/hash/d9d3837ee7981e8c064774da6cdd98bf-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d9d3837ee7981e8c064774da6cdd98bf-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11301-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d9d3837ee7981e8c064774da6cdd98bf-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d9d3837ee7981e8c064774da6cdd98bf-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d9d3837ee7981e8c064774da6cdd98bf-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d9d3837ee7981e8c064774da6cdd98bf-Supplemental.pdf | Preference-based Reinforcement Learning (PbRL) replaces reward values in traditional reinforcement learning by preferences to better elicit human opinion on the target objective, especially when numerical reward values are hard to design or interpret.
Despite promising results in applications, the theoretical understanding of PbRL is still in its infancy.
In this paper, we present the first finite-time analysis for general PbRL problems.
We first show that a unique optimal policy may not exist if preferences over trajectories are deterministic for PbRL.
If preferences are stochastic, and the preference probability relates to the hidden reward values, we present algorithms for PbRL, both with and without a simulator, that are able to identify the best policy up to accuracy $\varepsilon$ with high probability. Our method explores the state space by navigating to under-explored states, and solves PbRL using a combination of dueling bandits and policy search.
Experiments show the efficacy of our method when it is applied to real-world problems. |
AdaBelief Optimizer: Adapting Stepsizes by the Belief in Observed Gradients | https://papers.nips.cc/paper_files/paper/2020/hash/d9d4f495e875a2e075a1a4a6e1b9770f-Abstract.html | Juntang Zhuang, Tommy Tang, Yifan Ding, Sekhar C. Tatikonda, Nicha Dvornek, Xenophon Papademetris, James Duncan | https://papers.nips.cc/paper_files/paper/2020/hash/d9d4f495e875a2e075a1a4a6e1b9770f-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d9d4f495e875a2e075a1a4a6e1b9770f-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11302-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d9d4f495e875a2e075a1a4a6e1b9770f-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d9d4f495e875a2e075a1a4a6e1b9770f-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d9d4f495e875a2e075a1a4a6e1b9770f-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d9d4f495e875a2e075a1a4a6e1b9770f-Supplemental.zip | Most popular optimizers for deep learning can be broadly categorized as adaptive methods (e.g.~Adam) and accelerated schemes (e.g.~stochastic gradient descent (SGD) with momentum).
For many models such as convolutional neural networks (CNNs), adaptive methods typically converge faster but generalize worse compared to SGD; for complex settings such as generative adversarial networks (GANs), adaptive methods are typically the default because of their stability. We propose AdaBelief to simultaneously achieve three goals: fast convergence as in adaptive methods, good generalization as in SGD, and training stability. The intuition for AdaBelief is to adapt the stepsize according to the "belief" in the current gradient direction.
Viewing the exponential moving average (EMA) of the noisy gradient as the prediction of the gradient at the next time step, if the observed gradient greatly deviates from the prediction, we distrust the current observation and take a small step; if the observed gradient is close to the prediction, we trust it and take a large step.
We validate AdaBelief in extensive experiments, showing that it outperforms other methods with fast convergence and high accuracy on image classification and language modeling. Specifically, on ImageNet, AdaBelief achieves comparable accuracy to SGD. Furthermore, in the training of a GAN on Cifar10, AdaBelief demonstrates high stability and improves the quality of generated samples compared to a well-tuned Adam optimizer. |
Interpretable Sequence Learning for Covid-19 Forecasting | https://papers.nips.cc/paper_files/paper/2020/hash/d9dbc51dc534921589adf460c85cd824-Abstract.html | Sercan Arik, Chun-Liang Li, Jinsung Yoon, Rajarishi Sinha, Arkady Epshteyn, Long Le, Vikas Menon, Shashank Singh, Leyou Zhang, Martin Nikoltchev, Yash Sonthalia, Hootan Nakhost, Elli Kanal, Tomas Pfister | https://papers.nips.cc/paper_files/paper/2020/hash/d9dbc51dc534921589adf460c85cd824-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/d9dbc51dc534921589adf460c85cd824-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11303-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/d9dbc51dc534921589adf460c85cd824-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/d9dbc51dc534921589adf460c85cd824-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/d9dbc51dc534921589adf460c85cd824-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/d9dbc51dc534921589adf460c85cd824-Supplemental.pdf | We propose a novel approach that integrates machine learning into compartmental disease modeling (e.g., SEIR) to predict the progression of COVID-19. Our model is explainable by design as it explicitly shows how different compartments evolve and it uses interpretable encoders to incorporate covariates and improve performance. Explainability is valuable to ensure that the model's forecasts are credible to epidemiologists and to instill confidence in end-users such as policy makers and healthcare institutions. Our model can be applied at different geographic resolutions, and we demonstrate it for states and counties in the United States. We show that our model provides more accurate forecasts compared to the alternatives, and that it provides qualitatively meaningful explanatory insights. |
Off-policy Policy Evaluation For Sequential Decisions Under Unobserved Confounding | https://papers.nips.cc/paper_files/paper/2020/hash/da21bae82c02d1e2b8168d57cd3fbab7-Abstract.html | Hongseok Namkoong, Ramtin Keramati, Steve Yadlowsky, Emma Brunskill | https://papers.nips.cc/paper_files/paper/2020/hash/da21bae82c02d1e2b8168d57cd3fbab7-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/da21bae82c02d1e2b8168d57cd3fbab7-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11304-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/da21bae82c02d1e2b8168d57cd3fbab7-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/da21bae82c02d1e2b8168d57cd3fbab7-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/da21bae82c02d1e2b8168d57cd3fbab7-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/da21bae82c02d1e2b8168d57cd3fbab7-Supplemental.zip | When observed decisions depend only on observed features, off-policy policy evaluation (OPE) methods for sequential decision problems can estimate the performance of evaluation policies before deploying them. However, this assumption is frequently violated due to unobserved confounders, unrecorded variables that impact both the decisions and their outcomes. We assess robustness of OPE methods under unobserved confounding by developing worst-case bounds on the performance of an evaluation policy. When unobserved confounders can affect every decision in an episode, we demonstrate that even small amounts of per-decision confounding can heavily bias OPE methods. Fortunately, in a number of important settings found in healthcare, policy-making, and technology, unobserved confounders may directly affect only one of the many decisions made, and influence future decisions/rewards only through the directly affected decision. Under this less pessimistic model of one-decision confounding, we propose an efficient loss-minimization-based procedure for computing worst-case bounds, and prove its statistical consistency. On simulated healthcare examples---management of sepsis and interventions for autistic children---where this is a reasonable model, we demonstrate that our method invalidates non-robust results and provides meaningful certificates of robustness, allowing reliable selection of policies under unobserved confounding. |
Modern Hopfield Networks and Attention for Immune Repertoire Classification | https://papers.nips.cc/paper_files/paper/2020/hash/da4902cb0bc38210839714ebdcf0efc3-Abstract.html | Michael Widrich, Bernhard Schäfl, Milena Pavlović, Hubert Ramsauer, Lukas Gruber, Markus Holzleitner, Johannes Brandstetter, Geir Kjetil Sandve, Victor Greiff, Sepp Hochreiter, Günter Klambauer | https://papers.nips.cc/paper_files/paper/2020/hash/da4902cb0bc38210839714ebdcf0efc3-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/da4902cb0bc38210839714ebdcf0efc3-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11305-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/da4902cb0bc38210839714ebdcf0efc3-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/da4902cb0bc38210839714ebdcf0efc3-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/da4902cb0bc38210839714ebdcf0efc3-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/da4902cb0bc38210839714ebdcf0efc3-Supplemental.zip | A central mechanism in machine learning is to identify, store, and recognize patterns. How to learn, access, and retrieve such patterns is crucial in Hopfield networks and the more recent transformer architectures. We show that the attention mechanism of transformer architectures is actually the update rule of modern Hopfield networks that can store exponentially many patterns. We exploit this high storage capacity of modern Hopfield networks to solve a challenging multiple instance learning (MIL) problem in computational biology: immune repertoire classification. In immune repertoire classification, a vast number of immune receptors are used to predict the immune status of an individual. This constitutes a MIL problem with an unprecedentedly massive number of instances, two orders of magnitude larger than currently considered problems, and with an extremely low witness rate. Accurate and interpretable machine learning methods solving this problem could pave the way towards new vaccines and therapies, which is currently a very relevant research topic intensified by the COVID-19 crisis. In this work, we present our novel method DeepRC that integrates transformer-like attention, or equivalently modern Hopfield networks, into deep learning architectures for massive MIL such as immune repertoire classification. We demonstrate that DeepRC outperforms all other methods with respect to predictive performance on large-scale experiments including simulated and real-world virus infection data and enables the extraction of sequence motifs that are connected to a given disease class. Source code and datasets: https://github.com/ml-jku/DeepRC |
One Ring to Rule Them All: Certifiably Robust Geometric Perception with Outliers | https://papers.nips.cc/paper_files/paper/2020/hash/da6ea77475918a3d83c7e49223d453cc-Abstract.html | Heng Yang, Luca Carlone | https://papers.nips.cc/paper_files/paper/2020/hash/da6ea77475918a3d83c7e49223d453cc-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/da6ea77475918a3d83c7e49223d453cc-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11306-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/da6ea77475918a3d83c7e49223d453cc-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/da6ea77475918a3d83c7e49223d453cc-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/da6ea77475918a3d83c7e49223d453cc-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/da6ea77475918a3d83c7e49223d453cc-Supplemental.pdf | We propose the first general and practical framework to design certifiable algorithms for robust geometric perception in the presence of a large amount of outliers. We investigate the use of a truncated least squares (TLS) cost function, which is known to be robust to outliers, but leads to hard, nonconvex, and nonsmooth optimization problems. Our first contribution is to show that –for a broad class of geometric perception problems– TLS estimation can be reformulated as an optimization over the ring of polynomials and Lasserre’s hierarchy of convex moment relaxations is empirically tight at the minimum relaxation order (i.e., certifiably obtains the global minimum of the nonconvex TLS problem). Our second contribution is to exploit the structural sparsity of the objective and constraint polynomials and leverage basis reduction to significantly reduce the size of the semidefinite program (SDP) resulting from the moment relaxation, without compromising its tightness. Our third contribution is to develop scalable dual optimality certifiers from the lens of sums-of-squares (SOS) relaxation, that can compute the suboptimality gap and possibly certify global optimality of any candidate solution (e.g., returned by fast heuristics such as RANSAC or graduated non-convexity). Our dual certifiers leverage Douglas-Rachford Splitting to solve a convex feasibility SDP. Numerical experiments across different perception problems, including single rotation averaging, shape alignment, 3D point cloud and mesh registration, and high-integrity satellite pose estimation, demonstrate the tightness of our relaxations, the correctness of the certification, and the scalability of the proposed dual certifiers to large problems, beyond the reach of current SDP solvers. |
Task-Robust Model-Agnostic Meta-Learning | https://papers.nips.cc/paper_files/paper/2020/hash/da8ce53cf0240070ce6c69c48cd588ee-Abstract.html | Liam Collins, Aryan Mokhtari, Sanjay Shakkottai | https://papers.nips.cc/paper_files/paper/2020/hash/da8ce53cf0240070ce6c69c48cd588ee-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/da8ce53cf0240070ce6c69c48cd588ee-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11307-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/da8ce53cf0240070ce6c69c48cd588ee-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/da8ce53cf0240070ce6c69c48cd588ee-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/da8ce53cf0240070ce6c69c48cd588ee-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/da8ce53cf0240070ce6c69c48cd588ee-Supplemental.zip | Meta-learning methods have shown an impressive ability to train models that rapidly learn new tasks. However, these methods only aim to perform well in expectation over tasks coming from some particular distribution that is typically equivalent across meta-training and meta-testing, rather than considering worst-case task performance. In this work we introduce the notion of ``task-robustness'' by reformulating the popular Model-Agnostic Meta-Learning (MAML) objective \citep{finn2017model} such that the goal is to minimize the maximum loss over the observed meta-training tasks. The solution to this novel formulation is task-robust in the sense that it places equal importance on even the most difficult and/or rare tasks. This also means that it performs well over all distributions of the observed tasks, making it robust to shifts in the task distribution between meta-training and meta-testing. We present an algorithm to solve the proposed min-max problem, and show that it converges to an $\epsilon$-accurate point at the optimal rate of $\mathcal{O}(1/\epsilon^2)$ in the convex setting and to an $(\epsilon, \delta)$-stationary point at the rate of $\mathcal{O}(\max\{1/\epsilon^5, 1/\delta^5\})$ in nonconvex settings. We also provide an upper bound on the new task generalization error that captures the advantage of minimizing the worst-case task loss, and demonstrate this advantage in sinusoid regression and image classification experiments. |
R-learning in actor-critic model offers a biologically relevant mechanism for sequential decision-making | https://papers.nips.cc/paper_files/paper/2020/hash/da97f65bd113e490a5fab20c4a69f586-Abstract.html | Sergey Shuvaev, Sarah Starosta, Duda Kvitsiani, Adam Kepecs, Alexei Koulakov | https://papers.nips.cc/paper_files/paper/2020/hash/da97f65bd113e490a5fab20c4a69f586-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/da97f65bd113e490a5fab20c4a69f586-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11308-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/da97f65bd113e490a5fab20c4a69f586-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/da97f65bd113e490a5fab20c4a69f586-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/da97f65bd113e490a5fab20c4a69f586-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/da97f65bd113e490a5fab20c4a69f586-Supplemental.pdf | In real-world settings, we repeatedly decide whether to pursue better conditions or to keep things unchanged. Examples include time investment, employment, entertainment preferences etc. How do we make such decisions? To address this question, the field of behavioral ecology has developed foraging paradigms – the model settings in which human and non-human subjects decided when to leave depleting food resources. Foraging theory, represented by the marginal value theorem (MVT), provided accurate average-case stay-or-leave rules consistent with behaviors of subjects towards depleting resources. Yet, the algorithms underlying individual choices and ways to learn such algorithms remained unclear. In this work, we build interpretable deep actor-critic models to show that R-learning – a reinforcement learning (RL) approach balancing short-term and long-term rewards – is consistent with the way real-life agents may learn making stay-or-leave decisions. Specifically we show that deep R-learning predicts choice patterns consistent with behavior of mice in foraging tasks; its TD error, the training signal in our model, correlates with dopamine activity of ventral tegmental area (VTA) neurons in the brain. Our theoretical and experimental results show that deep R-learning agents leave depleting reward resources when reward intake rates fall below their exponential averages over past trials. This individual-case decision rule, learned within RL and matching the MVT on average, bridges the gap between these major approaches to sequential decision-making. We further argue that our proposed decision rule, resulting from R-learning and consistent with animals’ behavior, is Bayes optimal in dynamic real-world environments. Overall, our work links available sequential decision-making theories including the MVT, RL, and Bayesian approaches to propose the learning mechanism and an optimal decision rule for sequential stay-or-leave choices in natural environments. |
Revisiting Frank-Wolfe for Polytopes: Strict Complementarity and Sparsity | https://papers.nips.cc/paper_files/paper/2020/hash/da9e6a4a4aeca98588e4dd77ceb37695-Abstract.html | Dan Garber | https://papers.nips.cc/paper_files/paper/2020/hash/da9e6a4a4aeca98588e4dd77ceb37695-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/da9e6a4a4aeca98588e4dd77ceb37695-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11309-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/da9e6a4a4aeca98588e4dd77ceb37695-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/da9e6a4a4aeca98588e4dd77ceb37695-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/da9e6a4a4aeca98588e4dd77ceb37695-Review.html | null | In recent years it was proved that simple modifications of the classical Frank-Wolfe algorithm (aka conditional gradient algorithm) for smooth convex minimization over convex and compact polytopes, converge with linear rate, assuming the objective function has the quadratic growth property. However, the rate of these methods depends explicitly on the dimension of the problem which cannot explain their empirical success for large scale problems. In this paper we first demonstrate that already for very simple problems and even when the optimal solution lies on a low-dimensional face of the polytope, such dependence on the dimension cannot be avoided in worst case. We then revisit the addition of a strict complementarity assumption already considered in Wolfe's classical book \cite{Wolfe1970}, and prove that under this condition, the Frank-Wolfe method with away-steps and line-search converges linearly with rate that depends explicitly only on the dimension of the optimal face, hence providing a significant improvement in case the optimal solution is sparse. We motivate this strict complementarity condition by proving that it implies sparsity-robustness of optimal solutions to noise. |
Fast Convergence of Langevin Dynamics on Manifold: Geodesics meet Log-Sobolev | https://papers.nips.cc/paper_files/paper/2020/hash/dab10c50dc668cd8560df444ff3a4227-Abstract.html | Xiao Wang, Qi Lei, Ioannis Panageas | https://papers.nips.cc/paper_files/paper/2020/hash/dab10c50dc668cd8560df444ff3a4227-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/dab10c50dc668cd8560df444ff3a4227-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11310-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/dab10c50dc668cd8560df444ff3a4227-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/dab10c50dc668cd8560df444ff3a4227-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/dab10c50dc668cd8560df444ff3a4227-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/dab10c50dc668cd8560df444ff3a4227-Supplemental.pdf | Sampling is a fundamental and arguably very important task with numerous applications in Machine Learning. One approach to sample from a high dimensional distribution $e^{-f}$ for some function $f$ is the Langevin Algorithm (LA). Recently, there has been a lot of progress in showing fast convergence of LA even in cases where $f$ is non-convex, notably \cite{VW19}, \cite{MoritaRisteski} in which the former paper focuses on functions $f$ defined in $\mathbb{R}^n$ and the latter paper focuses on functions with symmetries (like matrix completion type objectives) with manifold structure. Our work generalizes the results of \cite{VW19} where $f$ is defined on a manifold $M$ rather than $\mathbb{R}^n$. From technical point of view, we show that KL decreases in a geometric rate whenever the distribution $e^{-f}$ satisfies a log-Sobolev inequality on $M$. |
Tensor Completion Made Practical | https://papers.nips.cc/paper_files/paper/2020/hash/dab1263d1e6a88c9ba5e7e294def5e8b-Abstract.html | Allen Liu, Ankur Moitra | https://papers.nips.cc/paper_files/paper/2020/hash/dab1263d1e6a88c9ba5e7e294def5e8b-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/dab1263d1e6a88c9ba5e7e294def5e8b-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11311-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/dab1263d1e6a88c9ba5e7e294def5e8b-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/dab1263d1e6a88c9ba5e7e294def5e8b-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/dab1263d1e6a88c9ba5e7e294def5e8b-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/dab1263d1e6a88c9ba5e7e294def5e8b-Supplemental.pdf | Tensor completion is a natural higher-order generalization of matrix completion where the goal is to recover a low-rank tensor from sparse observations of its entries. Existing algorithms are either heuristic without provable guarantees, based on solving large semidefinite programs which are impractical to run, or make strong assumptions such as requiring the factors to be nearly orthogonal. In this paper we introduce a new variant of alternating minimization, which in turn is inspired by understanding how the progress measures that guide convergence of alternating minimization in the matrix setting need to be adapted to the tensor setting. We show strong provable guarantees, including showing that our algorithm converges linearly to the true tensors even when the factors are highly correlated and can be implemented in nearly linear time. Moreover our algorithm is also highly practical and we show that we can complete third order tensors with a thousand dimensions from observing a tiny fraction of its entries. In contrast, and somewhat surprisingly, we show that the standard version of alternating minimization, without our new twist, can converge at a drastically slower rate in practice. |
Optimization and Generalization Analysis of Transduction through Gradient Boosting and Application to Multi-scale Graph Neural Networks | https://papers.nips.cc/paper_files/paper/2020/hash/dab49080d80c724aad5ebf158d63df41-Abstract.html | Kenta Oono, Taiji Suzuki | https://papers.nips.cc/paper_files/paper/2020/hash/dab49080d80c724aad5ebf158d63df41-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/dab49080d80c724aad5ebf158d63df41-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11312-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/dab49080d80c724aad5ebf158d63df41-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/dab49080d80c724aad5ebf158d63df41-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/dab49080d80c724aad5ebf158d63df41-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/dab49080d80c724aad5ebf158d63df41-Supplemental.pdf | It is known that the current graph neural networks (GNNs) are difficult to make themselves deep due to the problem known as over-smoothing. Multi-scale GNNs are a promising approach for mitigating the over-smoothing problem. However, there is little explanation of why it works empirically from the viewpoint of learning theory. In this study, we derive the optimization and generalization guarantees of transductive learning algorithms that include multi-scale GNNs. Using the boosting theory, we prove the convergence of the training error under weak learning-type conditions. By combining it with generalization gap bounds in terms of transductive Rademacher complexity, we show that a test error bound of a specific type of multi-scale GNNs that decreases corresponding to the number of node aggregations under some conditions. Our results offer theoretical explanations for the effectiveness of the multi-scale structure against the over-smoothing problem. We apply boosting algorithms to the training of multi-scale GNNs for real-world node prediction tasks. We confirm that its performance is comparable to existing GNNs, and the practical behaviors are consistent with theoretical observations. Code is available at https://github.com/delta2323/GB-GNN. |
Content Provider Dynamics and Coordination in Recommendation Ecosystems | https://papers.nips.cc/paper_files/paper/2020/hash/dabd8d2ce74e782c65a973ef76fd540b-Abstract.html | Omer Ben-Porat, Itay Rosenberg, Moshe Tennenholtz | https://papers.nips.cc/paper_files/paper/2020/hash/dabd8d2ce74e782c65a973ef76fd540b-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/dabd8d2ce74e782c65a973ef76fd540b-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11313-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/dabd8d2ce74e782c65a973ef76fd540b-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/dabd8d2ce74e782c65a973ef76fd540b-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/dabd8d2ce74e782c65a973ef76fd540b-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/dabd8d2ce74e782c65a973ef76fd540b-Supplemental.pdf | Recommendation Systems like YouTube are vibrant ecosystems with two types of users: Content consumers (those who watch videos) and content providers (those who create videos). While the computational task of recommending relevant content is largely solved, designing a system that guarantees high social welfare for \textit{all} stakeholders is still in its infancy. In this work, we investigate the dynamics of content creation using a game-theoretic lens. Employing a stylized model that was recently suggested by other works, we show that the dynamics will always converge to a pure Nash Equilibrium (PNE), but the convergence rate can be exponential. We complement the analysis by proposing an efficient PNE computation algorithm via a combinatorial optimization problem that is of independent interest. |
Almost Surely Stable Deep Dynamics | https://papers.nips.cc/paper_files/paper/2020/hash/daecf755df5b1d637033bb29b319c39a-Abstract.html | Nathan Lawrence, Philip Loewen, Michael Forbes, Johan Backstrom, Bhushan Gopaluni | https://papers.nips.cc/paper_files/paper/2020/hash/daecf755df5b1d637033bb29b319c39a-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/daecf755df5b1d637033bb29b319c39a-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11314-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/daecf755df5b1d637033bb29b319c39a-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/daecf755df5b1d637033bb29b319c39a-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/daecf755df5b1d637033bb29b319c39a-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/daecf755df5b1d637033bb29b319c39a-Supplemental.pdf | We introduce a method for learning provably stable deep neural network based dynamic models from observed data. Specifically, we consider discrete-time stochastic dynamic models, as they are of particular interest in practical applications such as estimation and control. However, these aspects exacerbate the challenge of guaranteeing stability. Our method works by embedding a Lyapunov neural network into the dynamic model, thereby inherently satisfying the stability criterion. To this end, we propose two approaches and apply them in both the deterministic and stochastic settings: one exploits convexity of the Lyapunov function, while the other enforces stability through an implicit output layer. We demonstrate the utility of each approach through numerical examples. |
Experimental design for MRI by greedy policy search | https://papers.nips.cc/paper_files/paper/2020/hash/daed210307f1dbc6f1dd9551408d999f-Abstract.html | Tim Bakker, Herke van Hoof, Max Welling | https://papers.nips.cc/paper_files/paper/2020/hash/daed210307f1dbc6f1dd9551408d999f-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/daed210307f1dbc6f1dd9551408d999f-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11315-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/daed210307f1dbc6f1dd9551408d999f-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/daed210307f1dbc6f1dd9551408d999f-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/daed210307f1dbc6f1dd9551408d999f-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/daed210307f1dbc6f1dd9551408d999f-Supplemental.pdf | In today’s clinical practice, magnetic resonance imaging (MRI) is routinely accelerated through subsampling of the associated Fourier domain. Currently, the construction of these subsampling strategies - known as experimental design - relies primarily on heuristics. We propose to learn experimental design strategies for accelerated MRI with policy gradient methods. Unexpectedly, our experiments show that a simple greedy approximation of the objective leads to solutions nearly on-par with the more general non-greedy approach. We offer a partial explanation for this phenomenon rooted in greater variance in the non-greedy objective's gradient estimates, and experimentally verify that this variance hampers non-greedy models in adapting their policies to individual MR images. We empirically show that this adaptivity is key to improving subsampling designs. |
Expert-Supervised Reinforcement Learning for Offline Policy Learning and Evaluation | https://papers.nips.cc/paper_files/paper/2020/hash/daf642455364613e2120c636b5a1f9c7-Abstract.html | Aaron Sonabend, Junwei Lu, Leo Anthony Celi, Tianxi Cai, Peter Szolovits | https://papers.nips.cc/paper_files/paper/2020/hash/daf642455364613e2120c636b5a1f9c7-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/daf642455364613e2120c636b5a1f9c7-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11316-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/daf642455364613e2120c636b5a1f9c7-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/daf642455364613e2120c636b5a1f9c7-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/daf642455364613e2120c636b5a1f9c7-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/daf642455364613e2120c636b5a1f9c7-Supplemental.zip | Offline Reinforcement Learning (RL) is a promising approach for learning optimal
policies in environments where direct exploration is expensive or unfeasible. However, the adoption of such policies in practice is often challenging, as they are hard
to interpret within the application context, and lack measures of uncertainty for the
learned policy value and its decisions. To overcome these issues, we propose an
Expert-Supervised RL (ESRL) framework which uses uncertainty quantification
for offline policy learning. In particular, we have three contributions: 1) the method
can learn safe and optimal policies through hypothesis testing, 2) ESRL allows for
different levels of risk averse implementations tailored to the application context,
and finally, 3) we propose a way to interpret ESRL’s policy at every state through
posterior distributions, and use this framework to compute off-policy value function
posteriors. We provide theoretical guarantees for our estimators and regret bounds
consistent with Posterior Sampling for RL (PSRL). Sample efficiency of ESRL
is independent of the chosen risk aversion threshold and quality of the behavior
policy. |
ColdGANs: Taming Language GANs with Cautious Sampling Strategies | https://papers.nips.cc/paper_files/paper/2020/hash/db261d4f615f0e982983be499e57ccda-Abstract.html | Thomas Scialom, Paul-Alexis Dray, Sylvain Lamprier, Benjamin Piwowarski, Jacopo Staiano | https://papers.nips.cc/paper_files/paper/2020/hash/db261d4f615f0e982983be499e57ccda-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/db261d4f615f0e982983be499e57ccda-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11317-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/db261d4f615f0e982983be499e57ccda-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/db261d4f615f0e982983be499e57ccda-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/db261d4f615f0e982983be499e57ccda-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/db261d4f615f0e982983be499e57ccda-Supplemental.pdf | We report experimental results obtained on three tasks: unconditional text generation, question generation, and abstractive summarization. For the first time, to the best of our knowledge, the proposed language GANs compare favorably to MLE, and obtain improvements over the state-of-the-art on the considered tasks. |
Hedging in games: Faster convergence of external and swap regrets | https://papers.nips.cc/paper_files/paper/2020/hash/db346ccb62d491029b590bbbf0f5c412-Abstract.html | Xi Chen, Binghui Peng | https://papers.nips.cc/paper_files/paper/2020/hash/db346ccb62d491029b590bbbf0f5c412-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/db346ccb62d491029b590bbbf0f5c412-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11318-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/db346ccb62d491029b590bbbf0f5c412-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/db346ccb62d491029b590bbbf0f5c412-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/db346ccb62d491029b590bbbf0f5c412-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/db346ccb62d491029b590bbbf0f5c412-Supplemental.pdf | We consider the setting where players run the Hedge algorithm or its optimistic variant \cite{syrgkanis2015fast} to play an n-action game repeatedly for T rounds.
1) For two-player games, we show that the regret of optimistic Hedge decays at \tilde{O}( 1/T ^{5/6} ), improving the previous bound O(1/T^{3/4}) by \cite{syrgkanis2015fast}.
2) In contrast, we show that the convergence rate of vanilla Hedge is no better than \tilde{\Omega}(1/ \sqrt{T})}, addressing an open question posted in \cite{syrgkanis2015fast}.
For general m-player games, we show that the swap regret of each player decays at rate \tilde{O}(m^{1/2} (n/T)^{3/4}) when they combine optimistic Hedge with the classical external-to-internal reduction of Blum and Mansour \cite{blum2007external}. The algorithm can also be modified to achieve the same rate against itself and a rate of \tilde{O}(\sqrt{n/T}) against adversaries. Via standard connections, our upper bounds also imply faster convergence to coarse correlated equilibria in two-player games and to correlated equilibria in multiplayer games. |
The Origins and Prevalence of Texture Bias in Convolutional Neural Networks | https://papers.nips.cc/paper_files/paper/2020/hash/db5f9f42a7157abe65bb145000b5871a-Abstract.html | Katherine Hermann, Ting Chen, Simon Kornblith | https://papers.nips.cc/paper_files/paper/2020/hash/db5f9f42a7157abe65bb145000b5871a-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/db5f9f42a7157abe65bb145000b5871a-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11319-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/db5f9f42a7157abe65bb145000b5871a-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/db5f9f42a7157abe65bb145000b5871a-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/db5f9f42a7157abe65bb145000b5871a-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/db5f9f42a7157abe65bb145000b5871a-Supplemental.pdf | Recent work has indicated that, unlike humans, ImageNet-trained CNNs tend to classify images by texture rather than by shape. How pervasive is this bias, and where does it come from? We find that, when trained on datasets of images with conflicting shape and texture, CNNs learn to classify by shape at least as easily as by texture. What factors, then, produce the texture bias in CNNs trained on ImageNet? Different unsupervised training objectives and different architectures have small but significant and largely independent effects on the level of texture bias. However, all objectives and architectures still lead to models that make texture-based classification decisions a majority of the time, even if shape information is decodable from their hidden representations. The effect of data augmentation is much larger. By taking less aggressive random crops at training time and applying simple, naturalistic augmentation (color distortion, noise, and blur), we train models that classify ambiguous images by shape a majority of the time, and outperform baselines on out-of-distribution test sets. Our results indicate that apparent differences in the way humans and ImageNet-trained CNNs process images may arise not primarily from differences in their internal workings, but from differences in the data that they see. |
Time-Reversal Symmetric ODE Network | https://papers.nips.cc/paper_files/paper/2020/hash/db8419f41d890df802dca330e6284952-Abstract.html | In Huh, Eunho Yang, Sung Ju Hwang, Jinwoo Shin | https://papers.nips.cc/paper_files/paper/2020/hash/db8419f41d890df802dca330e6284952-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/db8419f41d890df802dca330e6284952-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11320-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/db8419f41d890df802dca330e6284952-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/db8419f41d890df802dca330e6284952-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/db8419f41d890df802dca330e6284952-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/db8419f41d890df802dca330e6284952-Supplemental.zip | Time-reversal symmetry, which requires that the dynamics of a system should not change with the reversal of time axis, is a fundamental property that frequently holds in classical and quantum mechanics. In this paper, we propose a novel loss function that measures how well our ordinary differential equation (ODE) networks comply with this time-reversal symmetry; it is formally defined by the discrepancy in the time evolutions of ODE networks between forward and backward dynamics. Then, we design a new framework, which we name as Time-Reversal Symmetric ODE Networks (TRS-ODENs), that can learn the dynamics of physical systems more sample-efficiently by learning with the proposed loss function. We evaluate TRS-ODENs on several classical dynamics, and find they can learn the desired time evolution from observed noisy and complex trajectories. We also show that, even for systems that do not possess the full time-reversal symmetry, TRS-ODENs can achieve better predictive performances over baselines. |
Provable Overlapping Community Detection in Weighted Graphs | https://papers.nips.cc/paper_files/paper/2020/hash/db957c626a8cd7a27231adfbf51e20eb-Abstract.html | Jimit Majmudar, Stephen Vavasis | https://papers.nips.cc/paper_files/paper/2020/hash/db957c626a8cd7a27231adfbf51e20eb-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/db957c626a8cd7a27231adfbf51e20eb-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11321-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/db957c626a8cd7a27231adfbf51e20eb-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/db957c626a8cd7a27231adfbf51e20eb-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/db957c626a8cd7a27231adfbf51e20eb-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/db957c626a8cd7a27231adfbf51e20eb-Supplemental.zip | Community detection is a widely-studied unsupervised learning problem in which the task is to group similar entities together based on observed pairwise entity interactions. This problem has applications in diverse domains such as social network analysis and computational biology. There is a significant amount of literature studying this problem under the assumption that the communities do not overlap. When the communities are allowed to overlap, often a \textit{pure nodes} assumption is made, i.e. each community has a node that belongs exclusively to that community. This assumption, however, may not always be satisfied in practice. In this paper, we provide a provable method to detect overlapping communities in weighted graphs without explicitly making the pure nodes assumption. Moreover, contrary to most existing algorithms, our approach is based on convex optimization, for which many useful theoretical properties are already known. We demonstrate the success of our algorithm on artificial and real-world datasets. |
Fast Unbalanced Optimal Transport on a Tree | https://papers.nips.cc/paper_files/paper/2020/hash/dba31bb5c75992690f20c2d3b370ec7c-Abstract.html | Ryoma Sato, Makoto Yamada, Hisashi Kashima | https://papers.nips.cc/paper_files/paper/2020/hash/dba31bb5c75992690f20c2d3b370ec7c-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/dba31bb5c75992690f20c2d3b370ec7c-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11322-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/dba31bb5c75992690f20c2d3b370ec7c-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/dba31bb5c75992690f20c2d3b370ec7c-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/dba31bb5c75992690f20c2d3b370ec7c-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/dba31bb5c75992690f20c2d3b370ec7c-Supplemental.pdf | This study examines the time complexities of the unbalanced optimal transport problems from an algorithmic perspective for the first time. We reveal which problems in unbalanced optimal transport can/cannot be solved efficiently. Specifically, we prove that the Kantorovich Rubinstein distance and optimal partial transport in Euclidean metric cannot be computed in strongly subquadratic time under the strong exponential time hypothesis. Then, we propose an algorithm that solves a more general unbalanced optimal transport problem exactly in quasi-linear time on a tree metric. The proposed algorithm processes a tree with one million nodes in less than one second. Our analysis forms a foundation for the theoretical study of unbalanced optimal transport algorithms and opens the door to the applications of unbalanced optimal transport to million-scale datasets. |
Acceleration with a Ball Optimization Oracle | https://papers.nips.cc/paper_files/paper/2020/hash/dba4c1a117472f6aca95211285d0587e-Abstract.html | Yair Carmon, Arun Jambulapati, Qijia Jiang, Yujia Jin, Yin Tat Lee, Aaron Sidford, Kevin Tian | https://papers.nips.cc/paper_files/paper/2020/hash/dba4c1a117472f6aca95211285d0587e-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/dba4c1a117472f6aca95211285d0587e-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11323-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/dba4c1a117472f6aca95211285d0587e-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/dba4c1a117472f6aca95211285d0587e-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/dba4c1a117472f6aca95211285d0587e-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/dba4c1a117472f6aca95211285d0587e-Supplemental.pdf | Consider an oracle which takes a point x and returns the minimizer of a convex function f in an l2 ball of radius r around x. It is straightforward to show that roughly r^{-1}\log(1/epsilon) calls to the oracle suffice to find an \epsilon-approximate minimizer of f in an l2 unit ball. Perhaps surprisingly, this is not optimal: we design an accelerated algorithm which attains an epsilon-approximate minimizer with roughly r^{-2/3} \log(1/epsilon) oracle queries, and give a matching lower bound. Further, we implement ball optimization oracles for functions with a locally stable Hessian using a variant of Newton's method and, in certain cases, stochastic first-order methods. The resulting algorithms apply to a number of problems of practical and theoretical import, improving upon previous results for logistic and
linfinity regression and achieving guarantees comparable to the
state-of-the-art for lp regression. |
Avoiding Side Effects By Considering Future Tasks | https://papers.nips.cc/paper_files/paper/2020/hash/dc1913d422398c25c5f0b81cab94cc87-Abstract.html | Victoria Krakovna, Laurent Orseau, Richard Ngo, Miljan Martic, Shane Legg | https://papers.nips.cc/paper_files/paper/2020/hash/dc1913d422398c25c5f0b81cab94cc87-Abstract.html | NIPS 2020 | https://papers.nips.cc/paper_files/paper/2020/file/dc1913d422398c25c5f0b81cab94cc87-AuthorFeedback.pdf | https://papers.nips.cc/paper_files/paper/11324-/bibtex | https://papers.nips.cc/paper_files/paper/2020/file/dc1913d422398c25c5f0b81cab94cc87-MetaReview.html | https://papers.nips.cc/paper_files/paper/2020/file/dc1913d422398c25c5f0b81cab94cc87-Paper.pdf | https://papers.nips.cc/paper_files/paper/2020/file/dc1913d422398c25c5f0b81cab94cc87-Review.html | https://papers.nips.cc/paper_files/paper/2020/file/dc1913d422398c25c5f0b81cab94cc87-Supplemental.pdf | Designing reward functions is difficult: the designer has to specify what to do (what it means to complete the task) as well as what not to do (side effects that should be avoided while completing the task). To alleviate the burden on the reward designer, we propose an algorithm to automatically generate an auxiliary reward function that penalizes side effects. This auxiliary objective rewards the ability to complete possible future tasks, which decreases if the agent causes side effects during the current task. The future task reward can also give the agent an incentive to interfere with events in the environment that make future tasks less achievable, such as irreversible actions by other agents. To avoid this interference incentive, we introduce a baseline policy that represents a default course of action (such as doing nothing), and use it to filter out future tasks that are not achievable by default. We formally define interference incentives and show that the future task approach with a baseline policy avoids these incentives in the deterministic case. Using gridworld environments that test for side effects and interference, we show that our method avoids interference and is more effective for avoiding side effects than the common approach of penalizing irreversible actions. |
Subsets and Splits
No community queries yet
The top public SQL queries from the community will appear here once available.