stochastic-parrots/MNLP_M1_MCQA · Datasets at Hugging Face

question stringlengths 6 1.87k	choices sequencelengths 4 4	answer stringclasses 4 values
Which of the following are correct implementation for acquire function ? Assume 0 means UNLOCKED and 1 means LOCKED. Initially l->locked = 0.	[ "c \n void acquire(struct lock l)\n {\n for(;;)\n if(xchg(&l->locked, 1) == 0)\n return;\n }", "c \n void acquire(struct lock l)\n {\n if(cas(&l->locked, 0, 1) == 0)\n return;\n }", "c \n void acquire(struct lock l)\n {\n for(;;)\n if(cas(&l->locked, 1, 0) == 1)\n return;\n }", "c \n void acquire(struct lock l)\n {\n if(l->locked == 0) \n return;\n }" ]	A
Suppose a file system used only for reading immutable files in random fashion. What is the best block allocation strategy?	[ "Linked-list allocation", "Continuous allocation", "Index allocation with B-tree", "Index allocation with Hash-table" ]	B
Which flag prevents user programs from reading and writing kernel data?	[ "PTE_P", "PTE_U", "PTE_D", "PTE_W" ]	B
Which of the execution of an application are possible on a single-core machine?	[ "Concurrent execution", "Parallel execution", "Both concurrent and parallel execution", "Neither concurrent or parallel execution" ]	A
In JOS, suppose a value is passed between two Envs. What is the minimum number of executed system calls?	[ "1", "2", "3", "4" ]	B
What strace tool does?	[ "It prints out system calls for given program. These system calls are always called when executing the program.", "It prints out system calls for given program. These systems calls are called only for that particular instance of the program.", "To trace a symlink. I.e. to find where the symlink points to.", "To remove wildcards from the string." ]	B
What is a good distance metric to be used when you want to compute the similarity between documents independent of their length?A penalty will be applied for any incorrect answers.	[ "Cosine similarity", "Euclidean distance", "Manhattan distance", "Chi-squared distance" ]	A
Consider a matrix factorization problem of the form $\mathbf{X}=\mathbf{W Z}^{\top}$ to obtain an item-user recommender system where $x_{i j}$ denotes the rating given by $j^{\text {th }}$ user to the $i^{\text {th }}$ item . We use Root mean square error (RMSE) to gauge the quality of the factorization obtained. Select the correct option.	[ "Given a new item and a few ratings from existing users, we need to retrain the already trained recommender system from scratch to generate robust ratings for the user-item pairs containing this item.", "Regularization terms for $\\mathbf{W}$ and $\\mathbf{Z}$ in the form of their respective Frobenius norms are added to the RMSE so that the resulting objective function becomes convex.", "For obtaining a robust factorization of a matrix $\\mathbf{X}$ with $D$ rows and $N$ elements where $N \\ll D$, the latent dimension $\\mathrm{K}$ should lie somewhere between $D$ and $N$.", "None of the other options are correct." ]	D
You are doing your ML project. It is a regression task under a square loss. Your neighbor uses linear regression and least squares. You are smarter. You are using a neural net with 10 layers and activations functions $f(x)=3 x$. You have a powerful laptop but not a supercomputer. You are betting your neighbor a beer at Satellite who will have a substantially better scores. However, at the end it will essentially be a tie, so we decide to have two beers and both pay. What is the reason for the outcome of this bet?	[ "Because we use exactly the same scheme.", "Because it is almost impossible to train a network with 10 layers without a supercomputer.", "Because I should have used more layers.", "Because I should have used only one layer." ]	A
Which of the following is correct regarding Louvain algorithm?	[ "It creates a hierarchy of communities with a common root", "Clique is the only topology of nodes where the algorithm detects the same communities, independently of the starting point", "If n cliques of the same order are connected cyclically with n-1 edges, then the algorithm will always detect the same communities, independently of the starting point", "Modularity is always maximal for the communities found at the top level of the community hierarchy" ]	C
Let the first four retrieved documents be N N R R, where N denotes a non-relevant and R a relevant document. Then the MAP (Mean Average Precision) is:	[ "1/2", "5/12", "3/4", "7/24" ]	B
The inverse document frequency of a term can increase	[ "by adding the term to a document that contains the term", "by removing a document from the document collection that does not contain the term", "by adding a document to the document collection that contains the term", "by adding a document to the document collection that does not contain the term" ]	D
Which of the following is wrong regarding Ontologies?	[ "We can create more than one ontology that conceptualize the same real-world entities", "Ontologies help in the integration of data expressed in different models", "Ontologies support domain-specific vocabularies", "Ontologies dictate how semi-structured data are serialized" ]	D
In a Ranked Retrieval result, the result at position k is non-relevant and at k+1 is relevant. Which of the following is always true (P@k and R@k are the precision and recall of the result set consisting of the k top ranked documents)?	[ "P@k-1 > P@k+1", "P@k-1 = P@k+1", "R@k-1 < R@k+1", "R@k-1 = R@k+1" ]	C
What is true regarding Fagin's algorithm?	[ "It performs a complete scan over the posting files", "It provably returns the k documents with the largest aggregate scores", "Posting files need to be indexed by TF-IDF weights", "It never reads more than (kn)½ entries from a posting list" ]	B
Which of the following is WRONG for Ontologies?	[ "Different information systems need to agree on the same ontology in order to interoperate.", "They help in the integration of data expressed in different models.", "They give the possibility to specify schemas for different domains.", "They dictate how semi-structured data are serialized." ]	D
What is the benefit of LDA over LSI?	[ "LSI is sensitive to the ordering of the words in a document, whereas LDA is not", "LDA has better theoretical explanation, and its empirical results are in general better than LSI’s", "LSI is based on a model of how documents are generated, whereas LDA is not", "LDA represents semantic dimensions (topics, concepts) as weighted combinations of terms, whereas LSI does not" ]	B
Maintaining the order of document identifiers for vocabulary construction when partitioning the document collection is important	[ "in the index merging approach for single node machines", "in the map-reduce approach for parallel clusters", "in both", "in neither of the two" ]	A
Which of the following is correct regarding Crowdsourcing?	[ "Random Spammers give always the same answer for every question", "It is applicable only for binary classification problems", "Honey Pot discovers all the types of spammers but not the sloppy workers", "The output of Majority Decision can be equal to the one of Expectation-Maximization" ]	D
When computing PageRank iteratively, the computation ends when...	[ "The difference among the eigenvalues of two subsequent iterations falls below a predefined threshold", "The norm of the difference of rank vectors of two subsequent iterations falls below a predefined threshold", "All nodes of the graph have been visited at least once", "The probability of visiting an unseen node falls below a predefined threshold" ]	B
Suppose that an item in a leaf node N exists in every path. Which one is correct?	[ "N co-occurs with its prefix in every transaction.", "For every node P that is a parent of N in the fp tree, confidence(P->N) = 1", "N’s minimum possible support is equal to the number of paths.", "The item N exists in every candidate set." ]	C
In a Ranked Retrieval result, the result at position k is non-relevant and at k+1 is relevant. Which of the following is always true (P@k and R@k are the precision and recall of the result set consisting of the k top ranked documents)?	[ "P@k-1 > P@k+1", "P@k-1 = P@k+1", "R@k-1 < R@k+", "R@k-1 = R@k+1" ]	C
For the number of times the apriori algorithm and the FPgrowth algorithm for association rule mining are scanning the transaction database the following is true	[ "fpgrowth has always strictly fewer scans than apriori", "fpgrowth and apriori can have the same number of scans", "apriori cannot have fewer scans than fpgrowth", "all three above statements are false" ]	B
Given the following teleporting matrix (Ε) for nodes A, B and C:[0 ½ 0][0 0 0][0 ½ 1]and making no assumptions about the link matrix (R), which of the following is correct:(Reminder: columns are the probabilities to leave the respective node.)	[ "A random walker can never reach node A", "A random walker can never leave node A", "A random walker can always leave node C", "A random walker can always leave node B" ]	D
Which of the following methods does not exploit statistics on the co-occurrence of words in a text?	[ "Word embeddings\n\n\n", "Transformers\n\n\n", "Vector space retrieval\n\n\n", "Fasttext" ]	C
Which attribute gives the best split?A1PNa44b44A2PNx51y33A3PNt61j23	[ "A1", "A3", "A2", "All the same" ]	B
Suppose that q is density reachable from p. The chain of points that ensure this relationship are {t,u,g,r} Which one is FALSE?	[ "{t,u,g,r} have to be all core points.", "p and q will also be density-connected", "p has to be a core point", "q has to be a border point" ]	D
In User-Based Collaborative Filtering, which of the following is correct, assuming that all the ratings are positive?	[ "Pearson Correlation Coefficient and Cosine Similarity have different value range, but return the same similarity ranking for the users", "If the ratings of two users have both variance equal to 0, then their Cosine Similarity is maximized", "Pearson Correlation Coefficient and Cosine Similarity have the same value range, but can return different similarity ranking for the users", "If the variance of the ratings of one of the users is 0, then their Cosine Similarity is not computable" ]	B
The term frequency of a term is normalized	[ "by the maximal frequency of all terms in the document", "by the maximal frequency of the term in the document collection", "by the maximal frequency of any term in the vocabulary", "by the maximal term frequency of any document in the collection" ]	A
Data being classified as unstructured or structured depends on the:	[ "Degree of abstraction", "Level of human involvement", "Type of physical storage", "Amount of data " ]	A
Suppose you have a search engine that retrieves the top 100 documents and achieves 90% precision and 20% recall. You modify the search engine to retrieve the top 200 and mysteriously, the precision stays the same. Which one is CORRECT?	[ "The recall becomes 10%", "The number of relevant documents is 450", "The F-score stays the same", "This is not possible" ]	B
In the χ2 statistics for a binary feature, we obtain P(χ2 \| DF = 1) > 0.05. This means in this case, it is assumed:	[ "That the class labels depends on the feature", "That the class label is independent of the feature", "That the class label correlates with the feature", "None of the above" ]	B
Which of the following is correct regarding the use of Hidden Markov Models (HMMs) for entity recognition in text documents?	[ "The cost of learning the model is quadratic in the lengths of the text.", "The cost of predicting a word is linear in the lengths of the text preceding the word.", "An HMM model can be built using words enhanced with morphological features as input.", "The label of one word is predicted based on all the previous labels" ]	C
10 itemsets out of 100 contain item A, of which 5 also contain B. The rule A -> B has:	[ "5% support and 10% confidence", "10% support and 50% confidence", "5% support and 50% confidence", "10% support and 10% confidence" ]	C
Which of the following is correct regarding the use of Hidden Markov Models (HMMs) for entity recognition in text documents?	[ "HMMs cannot predict the label of a word that appears only in the test set", "If the smoothing parameter λ is equal to 1, the emission probabilities for all the words in the test set will be equal", "When computing the emission probabilities, a word can be replaced by a morphological feature (e.g., the number of uppercase first characters)", "The label of one word is predicted based on all the previous labels" ]	C
The number of non-zero entries in a column of a term-document matrix indicates:	[ "how many terms of the vocabulary a document contains", "how often a term of the vocabulary occurs in a document", "how relevant a term is for a document ", "none of the other responses is correct" ]	A
What is TRUE regarding Fagin's algorithm?	[ "Posting files need to be indexed by TF-IDF weights", "It performs a complete scan over the posting files", "It never reads more than (kn)1⁄2 entries from a posting list", "It provably returns the k documents with the largest aggregate scores" ]	D
A false negative in sampling can only occur for itemsets with support smaller than	[ "the threshold s", "ps", "pm", "None of the above" ]	D
Why is XML a document model?	[ "It supports application-specific markup", "It supports domain-specific schemas", "It has a serialized representation", "It uses HTML tags" ]	C
When indexing a document collection using an inverted file, the main space requirement is implied by	[ "The access structure", "The vocabulary", "The index file", "The postings file" ]	D
Which of the following is WRONG about inverted files? (Slide 24,28 Week 3)	[ "The space requirement for the postings file is O(n)", "Variable length compression is used to reduce the size of the index file", "The index file has space requirement of O(n^beta), where beta is about 1⁄2", "Storing differences among word addresses reduces the size of the postings file" ]	B
Which of the following is TRUE for Recommender Systems (RS)?	[ "The complexity of the Content-based RS depends on the number of users", "Item-based RS need not only the ratings but also the item features", "Matrix Factorization is typically robust to the cold-start problem.", "Matrix Factorization can predict a score for any user-item combination in the dataset." ]	D
Which of the following properties is part of the RDF Schema Language?	[ "Description", "Type", "Predicate", "Domain" ]	D
Which of the following is correct regarding crowdsourcing?	[ "Uniform spammers randomly select answers.", "The accuracy of majority voting is never equal to the one of Expectation Maximization.", "Honey pots can detect uniform spammers, random spammers and sloppy workers.", "Majority Decision and Expectation Maximization both give less weight to spammers’ answers." ]	C
When using bootstrapping in Random Forests, the number of different data items used to construct a single tree is:	[ "Smaller than the size of the training data set with high probability", "Of order square root of the size of the training set with high probability", "The same as the size of the training data set", "Depends on the outcome of the sampling process, and can be both smaller or larger than the training set" ]	A
To constrain an object of an RDF statement from being of an atomic type (e.g., String), one has to use the following RDF/RDFS property:	[ "rdf:type", "rdfs:range", "rdfs:domain", "rdfs:subClassOf" ]	B
What is a correct pruning strategy for decision tree induction?	[ "Apply Maximum Description Length principle", "Stop partitioning a node when either positive or negative samples dominate the samples of the other class", "Choose the model that maximizes L(M) + L(M\|D)", "Remove attributes with lowest information gain" ]	B
In the first pass over the database of the FP Growth algorithm	[ "Frequent itemsets are extracted", "A tree structure is constructed", "The frequency of items is computed", "Prefixes among itemsets are determined" ]	C
Your input is "Distributed Information Systems". Your model tries to predict "Distributed" and "Systems" by leveraging the fact that these words are in the neighborhood of "Information". This model can be:	[ "Bag of Words", "Word Embeddings", "LDA", "kNN" ]	B
Considering the transaction below, which one is WRONG? \|Transaction ID \|Items Bought\| \|--\|--\| \|1\|Tea\| \|2\|Tea, Yoghurt\| \|3\|Tea, Yoghurt, Kebap\| \|4 \|Kebap \| \|5\|Tea, Kebap\|	[ "{Yoghurt} -> {Kebab} has 50% confidence", "{Yoghurt, Kebap} has 20% support", "{Tea} has the highest support", "{Yoghurt} has the lowest support among all itemsets" ]	D
Which is an appropriate method for fighting skewed distributions of class labels in classification?	[ "Include an over-proportional number of samples from the larger class", "Use leave-one-out cross validation", "Construct the validation set such that the class label distribution approximately matches the global distribution of the class labels", "Generate artificial data points for the most frequent classes" ]	C
Consider the following set of frequent 3-itemsets: {1, 2, 3}, {1, 2, 4}, {1, 2, 5}, {1, 3, 4}, {2, 3, 4}, {2, 3, 5}, {3, 4, 5}. Which one is not a candidate 4-itemset?	[ "{1,2,4,5}", "{1,3,4,5} ", "{2,3,4,5}", "{1,2,3,4}" ]	B
Which of the following is true in the context of inverted files?	[ "Index merging compresses an inverted file index on disk and reduces the storage cost", "The trie structure used for index construction is also used as a data access structure to terms in the vocabulary", "The finer the addressing granularity used in documents, the smaller the posting file becomes", "Inverted files are optimized for supporting search on dynamic text collections" ]	B
Regarding Label Propagation, which of the following is false?	[ "The labels are inferred using the labels that are known apriori", "It can be interpreted as a random walk model", " Propagation of labels through high degree nodes are penalized by low abandoning probability", "Injection probability should be higher when labels are obtained from experts than by crowdworkers" ]	C
The type statement in RDF would be expressed in the relational data model by a table	[ "with one attribute", "with two attributes", "with three attributes", "cannot be expressed in the relational data model" ]	A
Given graph 1→2, 1→3, 2→3, 3→2, switching from Page Rank to Teleporting PageRank will have an influence on the value(s) of:	[ "All the nodes", "Node 1", "Node 2 and 3", "No nodes. The values will stay unchanged." ]	A
Which of the following is true regarding the random forest classification algorithm?	[ "It is not suitable for parallelization.", "It uses only a subset of features for learning in each weak learner.", "We compute a prediction by randomly selecting the decision of one weak learner.", "It produces a human interpretable model." ]	B
Which of the following properties is part of the RDF Schema Language?	[ "Type", "Predicate", "Description", "Domain" ]	D
When constructing a word embedding, negative samples are	[ "word - context word combinations that are not occurring in the document collection", "context words that are not part of the vocabulary of the document collection", "all less frequent words that do not occur in the context of a given word", "only words that never appear as context word" ]	A
Which of the following tasks would typically not be solved by clustering?	[ "Community detection in social networks", "Discretization of continuous features", "Spam detection in an email system", "Detection of latent topics in a document collection" ]	C
In general, what is true regarding Fagin's algorithm?	[ "It performs a complete scan over the posting files", "It provably returns the k documents with the largest aggregate scores", "Posting files need to be indexed by the TF-IDF weights", "It never reads more than (kn)½ entries from a posting list" ]	B
Which of the following statements is correct in the context of information extraction?	[ "A confidence measure that prunes too permissive patterns discovered with bootstrapping can help reducing semantic drift", "The bootstrapping technique requires a dataset where statements are labelled", "Distant supervision typically uses low-complexity features only, due to the lack of training data", "For supervised learning, sentences in which NER has detected no entities are used as negative samples" ]	A
In vector space retrieval each row of the matrix M corresponds to	[ "A document", "A concept", "A query", "A term" ]	D
Which of the following is correct regarding prediction models?	[ "Training error being less than test error means overfitting", "Training error being less than test error means underfitting", "Complex models tend to overfit, unless we feed them with more data", "Simple models have lower bias than complex models" ]	C
Applying SVD to a term-document matrix M. Each concept is represented in K	[ "as a singular value", "as a linear combination of terms of the vocabulary", "as a linear combination of documents in the document collection", "as a least squares approximation of the matrix M" ]	B
What is our final goal in machine learning? (One answer)	[ " Generalize ", " Underfit", " Overfit ", " Megafit " ]	A
A model predicts $\mathbf{\hat{y}} = [1, 0, 1, 1, 1]$. The ground truths are $\mathbf{y} = [1, 0, 0, 1, 1]$. What is the accuracy?	[ "0.5", "0.75", "0.8", "0.875" ]	C
K-Means:	[ "always converges to the same solution, no matter the initialization", "always converges, but not always to the same solution", "doesn't always converge", "can never converge" ]	B
What is the algorithm to perform optimization with gradient descent? Actions between Start loop and End loop are performed multiple times. (One answer)	[ "1 Start loop, 2 Initialize weights, 3 Compute gradients, 4 Update weights, 5 End loop", "1 Initialize weights, 2 Compute gradients, 3 Start loop, 4 Update weights, 5 End loop", "1 Initialize weights, 2 Start loop, 3 Update weights, 4 End loop, 5 Compute gradients ", "1 Initialize weights, 2 Start loop, 3 Compute gradients, 4 Update weights, 5 End Loop" ]	D
What is the mean squared error of $f$ for a sample, where $\textbf{x}$ is an input, $y$ a target and $f(\textbf{x},W)$ the mapping function ? (One answer)	[ " $\|\|y - f(\\textbf{x},W)\|\|^2 $ ", " $\|\|y - f(\\textbf{x},W)\|\| $", " $-\\log(P(y=i \| \\textbf{x})) = -\\log(\\frac{e^{\\textbf{f}_i(\\textbf{x},W)}}{\\sum_j e^{\\textbf{f}_j(\\textbf{x},W)}})$ ", " $P(y=i \|\\textbf{x}) = \\frac{e^{\\textbf{f}_i(\\textbf{x},W)}}{\\sum_j e^{\\textbf{f}_j(\\textbf{x},W)}}$ " ]	A
You write a Python code to optimize the weights of your linear regression with 10 features \textbf{using gradient descent} for 500 epochs. What is the minimum number of for-loops you need to perform your optimization?	[ "Two for-loops, one to iterate over the weights and the other to iterate over the epochs", "Only one for-loop to iterate over the epochs.", "Only one for-loop to iterate over the weights.", "No for-loop is really necessary. Everything can be vectorized" ]	B
Fill the missing line of code: (one answer)\\ \hspace{.5cm} \#code missing\\ \hspace{.5cm} np.mean(np.random.randn(1000))\\	[ "import np", "import numpy", "import numpy as np", "import np.mean\\\\\n\t\timport np.random" ]	C
What is the output of the following block of Python code? (one answer) \\ \verb\|my_string = `computational'\| \\ \verb\|print(my_string[1])\|\\ \verb\|print(my_string[3:5])\| \vspace{0.25cm}	[ "c\\\\mpu", "c\\\\mp", "o\\\\put", "o\\\\pu" ]	D
You are using a 3-layer fully-connected neural net with \textbf{ReLU activations}. Your input data has components in [0, 1]. \textbf{You initialize all your weights to -10}, and set all the bias terms to 0. You start optimizing using SGD. What will likely happen?	[ "The gradient is 0 so nothing happens", "The gradient is very large so the model can't converge", "Training is fine, but our neural net does only as well as a linear model", "Everything is fine" ]	A
You are using a 3-layer fully-connected neural, and you are using \textbf{$f(x) = 2x$ as your activation function} . Your input data has components in [0, 1]. \textbf{You initialize your weights using Kaiming (He) initialization}, and set all the bias terms to 0. You start optimizing using SGD. What will likely happen?	[ "The gradient is 0 so nothing happens", "The gradient is very large so the model can't converge", "Training is fine, but our neural net does only as well as a linear model", "Everything is fine" ]	C
What is a good representation for scores when classifying these three target classes: Car, Bike and Bus, in the context of logistic regression. (One or multiple answers)	[ "{Car: $(0,1,0)$,} {Bike: $(1,0,0)$,} {Bus: $(0,0,1)$}", "{Car: $(0,1)$,} {Bike: $(1,0)$,} {Bus: $(1,1)$}", "{Car: $1$,} {Bike: $2$,} {Bus: $3$}", "{Car: $(0,1)$,} {Bike: $(1,0)$,} {Bus: $(0.5,0.5)$}" ]	A
Mean Square Error loss:	[ "Minimizing the distance between the predicted point and the true point", "Maximizing the probability of the correct class", "Minimizing the score of false classes when they are close, or bigger than, the score of the true class", "Maximizing the accuracy" ]	A
You need to debug your Stochastic Gradient Descent update for a classification of three bridge types. Manually compute the model output for the feature vector $x=(1, 0, 0, 0, 0)$ and $W$ contains only zeros. The model is logistic regression, \textit{i.e.}, $\textrm{softmax}(Wx)$. Remember: \begin{equation} \textrm{softmax}_i(s) = \frac{e^{s_i}}{\sum_k e^{s_k}} \end{equation} (One answer!!!!!!)	[ "$(0, 0, 0)$", "$(\\frac{1}{3}, \\frac{1}{3}, \\frac{1}{3})$", "$(0, 0, 0, 0, 0)$", "$(\\frac{1}{5}, \\frac{1}{5}, \\frac{1}{5}, \\frac{1}{5}, \\frac{1}{5})$" ]	B
Consider the following PyTorch code: class ThreeLayerNet (nn.Module): def __init__(): super().__init__() def forward(x): x = nn.Linear(100, 10)(x) x = nn.ReLU()(x) x = nn.Linear(10, 200)(x) x = nn.ReLU()(x) x = nn.Linear(200, 1)(x) return x Suppose that inputs are 100-dimensional, and outputs are 1-dimensional. What will happen if we try to train this network?	[ "There will be an error because we are re-using the variable x throughout the forward() method.", "There will be an error because the second layer has more neurons than the first. The number of neurons must never increase from one layer to the next.", "The model will not train properly. The performance will be the same at the beginning of the first epoch and at the end of the last epoch.", "Everything is fine." ]	C
You are using a 3-layer fully-connected neural net with \textbf{ReLU activations}. Your input data has components in [0, 1]. \textbf{You initialize your weights by sampling from $\mathcal{N}(-10, 0.1)$ (Gaussians of mean -10 and variance 0.1)}, and set all the bias terms to 0. You start optimizing using SGD. What will likely happen?	[ "The gradient is 0 so nothing happens", "The gradient is very large so the model can't converge", "Training is fine, but our neural net does only as well as a linear model", "Everything is fine" ]	A
We saw in class that we can quickly decrease the spatial size of the representation using pooling layers. Is there another way to do this without pooling?	[ "Yes, by increasing the amount of padding.", "Yes, by increasing the stride.", "Yes, by increasing the number of filters.", "No, pooling is necessary." ]	B
The \textbf{parameters} (weights \textbf{W}) are learned with ... (One answer)	[ " training ", " validation ", " test ", " all the data together " ]	A
The \textbf{hyperparameters} are learned with ... (One answer)	[ " training ", " validation ", " test ", " all the data together " ]	B
We report the final performance (e.g., accuracy) on the ... (One answer)	[ " training ", " validation ", " test ", " all the data together " ]	C
We consider a classification problem on linearly separable data. Our dataset had an outlier---a point that is very far from the other datapoints in distance (and also far from margins in SVM but still correctly classified by the SVM classifier). We trained the SVM, logistic regression and 1-nearest-neighbour models on this dataset. We tested trained models on a test set that comes from the same distribution as training set, but doesn't have any outlier points. For any vector $v \in \R^D$ let $\\|v\\|_2 := \sqrt{v_1^2 + \dots + v_D^2}$ denote the Euclidean norm. The hard-margin SVM problem for linearly separable points in $\R^D$ is to minimize the Euclidean norm $\\| \wv \\|_2$ under some constraints. What are the additional constraints for this optimization problem?	[ "$y_n \\ww^\top x_n \\geq 1 ~ \forall n \\in \\{1,\\cdots,N\\}$", "$\\ww^\top x_n \\geq 1 ~ \forall n \\in\\{1,\\cdots,N\\}$", "$y_n + \\ww^\top x_n \\geq 1 ~ \forall n \\in \\{1,\\cdots,N\\}$", "$\frac{y_n}{\\ww^\top x_n }\\geq 1 ~\forall n \\in \\{1,\\cdots,N\\}$" ]	A
Consider the function $f(x)=-x^{2}$. Which of the following statements are true regarding subgradients of $f(x)$ at $x=0$ ?	[ "A subgradient does not exist as $f(x)$ is differentiable at $x=0$.", "A subgradient exists but is not unique.", "A subgradient exists and is unique.", "A subgradient does not exist even though $f(x)$ is differentiable at $x=0$." ]	D
When constructing a word embedding, what is true regarding negative samples?	[ "They are words that do not appear as context words", "They are selected among words which are not stop words", "Their frequency is decreased down to its logarithm", "They are oversampled if less frequent" ]	D
If the top 100 documents contain 50 relevant documents	[ "the precision of the system at 50 is 0.25", "the precision of the system at 100 is 0.5", "the recall of the system is 0.5", "All of the above" ]	B
What is WRONG regarding the Transformer model?	[ "It uses a self-attention mechanism to compute representations of the input and output.", "Its computation cannot be parallelized compared to LSTMs and other sequential models.", "Its complexity is quadratic to the input size.", "It captures the semantic context of the input." ]	B
Which of the following statements about index merging (when constructing inverted files) is correct?	[ "While merging two partial indices on disk, the inverted lists of a term are concatenated without sorting", "Index merging is used when the vocabulary does no longer fit into the main memory", "The size of the final merged index file is O (n log2 (n) M )), where M is the size of the available memory", "While merging two partial indices on disk, the vocabularies are concatenated without sorting" ]	A
Which of the following statements on Latent Semantic Indexing (LSI) and Word Embeddings (WE) is false?	[ "The dimensions of LSI can be interpreted as concepts, whereas those of WE cannot", "LSI does not depend on the order of words in the document, whereas WE does", "LSI is deterministic (given the dimension), whereas WE is not", "LSI does take into account the frequency of words in the documents, whereas WE with negative sampling does not" ]	D
The number of non-zero entries in a column of a term-document matrix indicates:	[ "how many terms of the vocabulary a document contains", "how often a term of the vocabulary occurs in a document", "how relevant a term is for a document", "none of the other responses is correct" ]	D
Which of the following statements on Latent Semantic Indexing (LSI) and Word Embeddings (WE) is incorrect	[ "LSI is deterministic (given the dimension), whereas WE is not", "LSI does not take into account the order of words in the document, whereas WE does", "The dimensions of LSI can be interpreted as concepts, whereas those of WE cannot", "LSI does take into account the frequency of words in the documents, whereas WE does not." ]	D
Suppose that in a given FP Tree, an item in a leaf node N exists in every path. Which of the following is true?	[ "N co-occurs with its prefixes in every transaction", "{N}’s minimum possible support is equal to the number of paths", "For every node P that is a parent of N in the FP tree, confidence(P->N) = 1", "The item N exists in every candidate set" ]	B
Which of the following statements regarding topic models is false?	[ "Topic models map documents to dense vectors", "In LDA, topics are modeled as distributions over documents", "LDA assumes that each document is generated from a mixture of topics with a probability distribution", "Topics can serve as features for document classification" ]	B
Which of the following statements is correct concerning the use of Pearson’s Correlation for user- based collaborative filtering?	[ "It measures whether different users have similar preferences for the same items", "It measures how much a user’s ratings deviate from the average ratings I", "t measures how well the recommendations match the user’s preferences", "It measures whether a user has similar preferences for different items" ]	A
After the join step, the number of k+1-itemsets	[ "is equal to the number of frequent k-itemsets", "can be equal, lower or higher than the number of frequent k-itemsets", "is always higher than the number of frequent k-itemsets", "is always lower than the number of frequent k-itemsets" ]	B
Which is true about the use of entropy in decision tree induction?	[ "The entropy of the set of class labels of the samples from the training set at the leaf level is always 0", "We split on the attribute that has the highest entropy", "The entropy of the set of class labels of the samples from the training set at the leaf level can be 1", "We split on the attribute that has the lowest entropy" ]	C
Which of the following models for generating vector representations for text require to precompute the frequency of co-occurrence of words from the vocabulary in the document collection	[ "LSI", "CBOW", "Fasttext", "Glove" ]	D
Regarding the Expectation-Maximization algorithm, which one of the following false?	[ "Assigning equal weights to workers initially decreases the convergence time", "The label with the highest probability is assigned as the new label", "It distinguishes experts from normal workers", "In E step the labels change, in M step the weights of the workers change" ]	A

Which of the following are correct implementation for acquire function ? Assume 0 means UNLOCKED and 1 means LOCKED. Initially l->locked = 0.

[ "c \n void acquire(struct lock *l)\n {\n for(;;)\n if(xchg(&l->locked, 1) == 0)\n return;\n }", "c \n void acquire(struct lock *l)\n {\n if(cas(&l->locked, 0, 1) == 0)\n return;\n }", "c \n void acquire(struct lock *l)\n {\n for(;;)\n if(cas(&l->locked, 1, 0) == 1)\n return;\n }", "c \n void acquire(struct lock *l)\n {\n if(l->locked == 0) \n return;\n }" ]

A

Suppose a file system used only for reading immutable files in random fashion. What is the best block allocation strategy?

[ "Linked-list allocation", "Continuous allocation", "Index allocation with B-tree", "Index allocation with Hash-table" ]

B

Which flag prevents user programs from reading and writing kernel data?

[ "PTE_P", "PTE_U", "PTE_D", "PTE_W" ]

B

Which of the execution of an application are possible on a single-core machine?

[ "Concurrent execution", "Parallel execution", "Both concurrent and parallel execution", "Neither concurrent or parallel execution" ]

A

In JOS, suppose a value is passed between two Envs. What is the minimum number of executed system calls?

[ "1", "2", "3", "4" ]

B

What strace tool does?

[ "It prints out system calls for given program. These system calls are always called when executing the program.", "It prints out system calls for given program. These systems calls are called only for that particular instance of the program.", "To trace a symlink. I.e. to find where the symlink points to.", "To remove wildcards from the string." ]

B

What is a good distance metric to be used when you want to compute the similarity between documents independent of their length?A penalty will be applied for any incorrect answers.

[ "Cosine similarity", "Euclidean distance", "Manhattan distance", "Chi-squared distance" ]

A

Consider a matrix factorization problem of the form $\mathbf{X}=\mathbf{W Z}^{\top}$ to obtain an item-user recommender system where $x_{i j}$ denotes the rating given by $j^{\text {th }}$ user to the $i^{\text {th }}$ item . We use Root mean square error (RMSE) to gauge the quality of the factorization obtained. Select the correct option.

[ "Given a new item and a few ratings from existing users, we need to retrain the already trained recommender system from scratch to generate robust ratings for the user-item pairs containing this item.", "Regularization terms for $\\mathbf{W}$ and $\\mathbf{Z}$ in the form of their respective Frobenius norms are added to the RMSE so that the resulting objective function becomes convex.", "For obtaining a robust factorization of a matrix $\\mathbf{X}$ with $D$ rows and $N$ elements where $N \\ll D$, the latent dimension $\\mathrm{K}$ should lie somewhere between $D$ and $N$.", "None of the other options are correct." ]

D

You are doing your ML project. It is a regression task under a square loss. Your neighbor uses linear regression and least squares. You are smarter. You are using a neural net with 10 layers and activations functions $f(x)=3 x$. You have a powerful laptop but not a supercomputer. You are betting your neighbor a beer at Satellite who will have a substantially better scores. However, at the end it will essentially be a tie, so we decide to have two beers and both pay. What is the reason for the outcome of this bet?

[ "Because we use exactly the same scheme.", "Because it is almost impossible to train a network with 10 layers without a supercomputer.", "Because I should have used more layers.", "Because I should have used only one layer." ]

A

Which of the following is correct regarding Louvain algorithm?

[ "It creates a hierarchy of communities with a common root", "Clique is the only topology of nodes where the algorithm detects the same communities, independently of the starting point", "If n cliques of the same order are connected cyclically with n-1 edges, then the algorithm will always detect the same communities, independently of the starting point", "Modularity is always maximal for the communities found at the top level of the community hierarchy" ]

C

Let the first four retrieved documents be N N R R, where N denotes a non-relevant and R a relevant document. Then the MAP (Mean Average Precision) is:

[ "1/2", "5/12", "3/4", "7/24" ]

B

The inverse document frequency of a term can increase

[ "by adding the term to a document that contains the term", "by removing a document from the document collection that does not contain the term", "by adding a document to the document collection that contains the term", "by adding a document to the document collection that does not contain the term" ]

D

Which of the following is wrong regarding Ontologies?

[ "We can create more than one ontology that conceptualize the same real-world entities", "Ontologies help in the integration of data expressed in different models", "Ontologies support domain-specific vocabularies", "Ontologies dictate how semi-structured data are serialized" ]

D

In a Ranked Retrieval result, the result at position k is non-relevant and at k+1 is relevant. Which of the following is always true (P@k and R@k are the precision and recall of the result set consisting of the k top ranked documents)?

[ "P@k-1 > P@k+1", "P@k-1 = P@k+1", "R@k-1 < R@k+1", "R@k-1 = R@k+1" ]

C

What is true regarding Fagin's algorithm?

[ "It performs a complete scan over the posting files", "It provably returns the k documents with the largest aggregate scores", "Posting files need to be indexed by TF-IDF weights", "It never reads more than (kn)½ entries from a posting list" ]

B

Which of the following is WRONG for Ontologies?

[ "Different information systems need to agree on the same ontology in order to interoperate.", "They help in the integration of data expressed in different models.", "They give the possibility to specify schemas for different domains.", "They dictate how semi-structured data are serialized." ]

D

What is the benefit of LDA over LSI?

[ "LSI is sensitive to the ordering of the words in a document, whereas LDA is not", "LDA has better theoretical explanation, and its empirical results are in general better than LSI’s", "LSI is based on a model of how documents are generated, whereas LDA is not", "LDA represents semantic dimensions (topics, concepts) as weighted combinations of terms, whereas LSI does not" ]

B

Maintaining the order of document identifiers for vocabulary construction when partitioning the document collection is important

[ "in the index merging approach for single node machines", "in the map-reduce approach for parallel clusters", "in both", "in neither of the two" ]

A

Which of the following is correct regarding Crowdsourcing?

[ "Random Spammers give always the same answer for every question", "It is applicable only for binary classification problems", "Honey Pot discovers all the types of spammers but not the sloppy workers", "The output of Majority Decision can be equal to the one of Expectation-Maximization" ]

D

When computing PageRank iteratively, the computation ends when...

[ "The difference among the eigenvalues of two subsequent iterations falls below a predefined threshold", "The norm of the difference of rank vectors of two subsequent iterations falls below a predefined threshold", "All nodes of the graph have been visited at least once", "The probability of visiting an unseen node falls below a predefined threshold" ]

B

Suppose that an item in a leaf node N exists in every path. Which one is correct?

[ "N co-occurs with its prefix in every transaction.", "For every node P that is a parent of N in the fp tree, confidence(P->N) = 1", "N’s minimum possible support is equal to the number of paths.", "The item N exists in every candidate set." ]

C

In a Ranked Retrieval result, the result at position k is non-relevant and at k+1 is relevant. Which of the following is always true (P@k and R@k are the precision and recall of the result set consisting of the k top ranked documents)?

[ "P@k-1 > P@k+1", "P@k-1 = P@k+1", "R@k-1 < R@k+", "R@k-1 = R@k+1" ]

C

For the number of times the apriori algorithm and the FPgrowth algorithm for association rule mining are scanning the transaction database the following is true

[ "fpgrowth has always strictly fewer scans than apriori", "fpgrowth and apriori can have the same number of scans", "apriori cannot have fewer scans than fpgrowth", "all three above statements are false" ]

B

Given the following teleporting matrix (Ε) for nodes A, B and C:[0 ½ 0][0 0 0][0 ½ 1]and making no assumptions about the link matrix (R), which of the following is correct:(Reminder: columns are the probabilities to leave the respective node.)

[ "A random walker can never reach node A", "A random walker can never leave node A", "A random walker can always leave node C", "A random walker can always leave node B" ]

D

Which of the following methods does not exploit statistics on the co-occurrence of words in a text?

[ "Word embeddings\n\n\n", "Transformers\n\n\n", "Vector space retrieval\n\n\n", "Fasttext" ]

C

Which attribute gives the best split?A1PNa44b44A2PNx51y33A3PNt61j23

[ "A1", "A3", "A2", "All the same" ]

B

Suppose that q is density reachable from p. The chain of points that ensure this relationship are {t,u,g,r} Which one is FALSE?

[ "{t,u,g,r} have to be all core points.", "p and q will also be density-connected", "p has to be a core point", "q has to be a border point" ]

D

In User-Based Collaborative Filtering, which of the following is correct, assuming that all the ratings are positive?

[ "Pearson Correlation Coefficient and Cosine Similarity have different value range, but return the same similarity ranking for the users", "If the ratings of two users have both variance equal to 0, then their Cosine Similarity is maximized", "Pearson Correlation Coefficient and Cosine Similarity have the same value range, but can return different similarity ranking for the users", "If the variance of the ratings of one of the users is 0, then their Cosine Similarity is not computable" ]

B

The term frequency of a term is normalized

[ "by the maximal frequency of all terms in the document", "by the maximal frequency of the term in the document collection", "by the maximal frequency of any term in the vocabulary", "by the maximal term frequency of any document in the collection" ]

A

Data being classified as unstructured or structured depends on the:

[ "Degree of abstraction", "Level of human involvement", "Type of physical storage", "Amount of data " ]

A

Suppose you have a search engine that retrieves the top 100 documents and achieves 90% precision and 20% recall. You modify the search engine to retrieve the top 200 and mysteriously, the precision stays the same. Which one is CORRECT?

[ "The recall becomes 10%", "The number of relevant documents is 450", "The F-score stays the same", "This is not possible" ]

B

In the χ2 statistics for a binary feature, we obtain P(χ2 | DF = 1) > 0.05. This means in this case, it is assumed:

[ "That the class labels depends on the feature", "That the class label is independent of the feature", "That the class label correlates with the feature", "None of the above" ]

B

Which of the following is correct regarding the use of Hidden Markov Models (HMMs) for entity recognition in text documents?

[ "The cost of learning the model is quadratic in the lengths of the text.", "The cost of predicting a word is linear in the lengths of the text preceding the word.", "An HMM model can be built using words enhanced with morphological features as input.", "The label of one word is predicted based on all the previous labels" ]

C

10 itemsets out of 100 contain item A, of which 5 also contain B. The rule A -> B has:

[ "5% support and 10% confidence", "10% support and 50% confidence", "5% support and 50% confidence", "10% support and 10% confidence" ]

C

Which of the following is correct regarding the use of Hidden Markov Models (HMMs) for entity recognition in text documents?

[ "HMMs cannot predict the label of a word that appears only in the test set", "If the smoothing parameter λ is equal to 1, the emission probabilities for all the words in the test set will be equal", "When computing the emission probabilities, a word can be replaced by a morphological feature (e.g., the number of uppercase first characters)", "The label of one word is predicted based on all the previous labels" ]

C

The number of non-zero entries in a column of a term-document matrix indicates:

[ "how many terms of the vocabulary a document contains", "how often a term of the vocabulary occurs in a document", "how relevant a term is for a document ", "none of the other responses is correct" ]

A

What is TRUE regarding Fagin's algorithm?

[ "Posting files need to be indexed by TF-IDF weights", "It performs a complete scan over the posting files", "It never reads more than (kn)1⁄2 entries from a posting list", "It provably returns the k documents with the largest aggregate scores" ]

D

A false negative in sampling can only occur for itemsets with support smaller than

[ "the threshold s", "p*s", "p*m", "None of the above" ]

D

Why is XML a document model?

[ "It supports application-specific markup", "It supports domain-specific schemas", "It has a serialized representation", "It uses HTML tags" ]

C

When indexing a document collection using an inverted file, the main space requirement is implied by

[ "The access structure", "The vocabulary", "The index file", "The postings file" ]

D

Which of the following is WRONG about inverted files? (Slide 24,28 Week 3)

[ "The space requirement for the postings file is O(n)", "Variable length compression is used to reduce the size of the index file", "The index file has space requirement of O(n^beta), where beta is about 1⁄2", "Storing differences among word addresses reduces the size of the postings file" ]

B

Which of the following is TRUE for Recommender Systems (RS)?

[ "The complexity of the Content-based RS depends on the number of users", "Item-based RS need not only the ratings but also the item features", "Matrix Factorization is typically robust to the cold-start problem.", "Matrix Factorization can predict a score for any user-item combination in the dataset." ]

D

Which of the following properties is part of the RDF Schema Language?

[ "Description", "Type", "Predicate", "Domain" ]

D

Which of the following is correct regarding crowdsourcing?

[ "Uniform spammers randomly select answers.", "The accuracy of majority voting is never equal to the one of Expectation Maximization.", "Honey pots can detect uniform spammers, random spammers and sloppy workers.", "Majority Decision and Expectation Maximization both give less weight to spammers’ answers." ]

C

When using bootstrapping in Random Forests, the number of different data items used to construct a single tree is:

[ "Smaller than the size of the training data set with high probability", "Of order square root of the size of the training set with high probability", "The same as the size of the training data set", "Depends on the outcome of the sampling process, and can be both smaller or larger than the training set" ]

A

To constrain an object of an RDF statement from being of an atomic type (e.g., String), one has to use the following RDF/RDFS property:

[ "rdf:type", "rdfs:range", "rdfs:domain", "rdfs:subClassOf" ]

B

What is a correct pruning strategy for decision tree induction?

[ "Apply Maximum Description Length principle", "Stop partitioning a node when either positive or negative samples dominate the samples of the other class", "Choose the model that maximizes L(M) + L(M|D)", "Remove attributes with lowest information gain" ]

B

In the first pass over the database of the FP Growth algorithm

[ "Frequent itemsets are extracted", "A tree structure is constructed", "The frequency of items is computed", "Prefixes among itemsets are determined" ]

C

Your input is "Distributed Information Systems". Your model tries to predict "Distributed" and "Systems" by leveraging the fact that these words are in the neighborhood of "Information". This model can be:

[ "Bag of Words", "Word Embeddings", "LDA", "kNN" ]

B

Considering the transaction below, which one is WRONG? |Transaction ID |Items Bought| |--|--| |1|Tea| |2|Tea, Yoghurt| |3|Tea, Yoghurt, Kebap| |4 |Kebap | |5|Tea, Kebap|

[ "{Yoghurt} -> {Kebab} has 50% confidence", "{Yoghurt, Kebap} has 20% support", "{Tea} has the highest support", "{Yoghurt} has the lowest support among all itemsets" ]

D

Which is an appropriate method for fighting skewed distributions of class labels in classification?

[ "Include an over-proportional number of samples from the larger class", "Use leave-one-out cross validation", "Construct the validation set such that the class label distribution approximately matches the global distribution of the class labels", "Generate artificial data points for the most frequent classes" ]

C

Consider the following set of frequent 3-itemsets: {1, 2, 3}, {1, 2, 4}, {1, 2, 5}, {1, 3, 4}, {2, 3, 4}, {2, 3, 5}, {3, 4, 5}. Which one is not a candidate 4-itemset?

[ "{1,2,4,5}", "{1,3,4,5} ", "{2,3,4,5}", "{1,2,3,4}" ]

B

Which of the following is true in the context of inverted files?

[ "Index merging compresses an inverted file index on disk and reduces the storage cost", "The trie structure used for index construction is also used as a data access structure to terms in the vocabulary", "The finer the addressing granularity used in documents, the smaller the posting file becomes", "Inverted files are optimized for supporting search on dynamic text collections" ]

B

Regarding Label Propagation, which of the following is false?

[ "The labels are inferred using the labels that are known apriori", "It can be interpreted as a random walk model", " Propagation of labels through high degree nodes are penalized by low abandoning probability", "Injection probability should be higher when labels are obtained from experts than by crowdworkers" ]

C

The type statement in RDF would be expressed in the relational data model by a table

[ "with one attribute", "with two attributes", "with three attributes", "cannot be expressed in the relational data model" ]

A

Given graph 1→2, 1→3, 2→3, 3→2, switching from Page Rank to Teleporting PageRank will have an influence on the value(s) of:

[ "All the nodes", "Node 1", "Node 2 and 3", "No nodes. The values will stay unchanged." ]

A

Which of the following is true regarding the random forest classification algorithm?

[ "It is not suitable for parallelization.", "It uses only a subset of features for learning in each weak learner.", "We compute a prediction by randomly selecting the decision of one weak learner.", "It produces a human interpretable model." ]

B

Which of the following properties is part of the RDF Schema Language?

[ "Type", "Predicate", "Description", "Domain" ]

D

When constructing a word embedding, negative samples are

[ "word - context word combinations that are not occurring in the document collection", "context words that are not part of the vocabulary of the document collection", "all less frequent words that do not occur in the context of a given word", "only words that never appear as context word" ]

A

Which of the following tasks would typically not be solved by clustering?

[ "Community detection in social networks", "Discretization of continuous features", "Spam detection in an email system", "Detection of latent topics in a document collection" ]

C

In general, what is true regarding Fagin's algorithm?

[ "It performs a complete scan over the posting files", "It provably returns the k documents with the largest aggregate scores", "Posting files need to be indexed by the TF-IDF weights", "It never reads more than (kn)½ entries from a posting list" ]

B

Which of the following statements is correct in the context of information extraction?

[ "A confidence measure that prunes too permissive patterns discovered with bootstrapping can help reducing semantic drift", "The bootstrapping technique requires a dataset where statements are labelled", "Distant supervision typically uses low-complexity features only, due to the lack of training data", "For supervised learning, sentences in which NER has detected no entities are used as negative samples" ]

A

In vector space retrieval each row of the matrix M corresponds to

[ "A document", "A concept", "A query", "A term" ]

D

Which of the following is correct regarding prediction models?

[ "Training error being less than test error means overfitting", "Training error being less than test error means underfitting", "Complex models tend to overfit, unless we feed them with more data", "Simple models have lower bias than complex models" ]

C

Applying SVD to a term-document matrix M. Each concept is represented in K

[ "as a singular value", "as a linear combination of terms of the vocabulary", "as a linear combination of documents in the document collection", "as a least squares approximation of the matrix M" ]

B

What is our final goal in machine learning? (One answer)

[ " Generalize ", " Underfit", " Overfit ", " Megafit " ]

A

A model predicts $\mathbf{\hat{y}} = [1, 0, 1, 1, 1]$. The ground truths are $\mathbf{y} = [1, 0, 0, 1, 1]$. What is the accuracy?

[ "0.5", "0.75", "0.8", "0.875" ]

C

K-Means:

[ "always converges to the same solution, no matter the initialization", "always converges, but not always to the same solution", "doesn't always converge", "can never converge" ]

B

What is the algorithm to perform optimization with gradient descent? Actions between Start loop and End loop are performed multiple times. (One answer)

[ "1 Start loop, 2 Initialize weights, 3 Compute gradients, 4 Update weights, 5 End loop", "1 Initialize weights, 2 Compute gradients, 3 Start loop, 4 Update weights, 5 End loop", "1 Initialize weights, 2 Start loop, 3 Update weights, 4 End loop, 5 Compute gradients ", "1 Initialize weights, 2 Start loop, 3 Compute gradients, 4 Update weights, 5 End Loop" ]

D

What is the mean squared error of $f$ for a sample, where $\textbf{x}$ is an input, $y$ a target and $f(\textbf{x},W)$ the mapping function ? (One answer)

[ " $||y - f(\\textbf{x},W)||^2 $ ", " $||y - f(\\textbf{x},W)|| $", " $-\\log(P(y=i | \\textbf{x})) = -\\log(\\frac{e^{\\textbf{f}_i(\\textbf{x},W)}}{\\sum_j e^{\\textbf{f}_j(\\textbf{x},W)}})$ ", " $P(y=i |\\textbf{x}) = \\frac{e^{\\textbf{f}_i(\\textbf{x},W)}}{\\sum_j e^{\\textbf{f}_j(\\textbf{x},W)}}$ " ]

A

You write a Python code to optimize the weights of your linear regression with 10 features \textbf{using gradient descent} for 500 epochs. What is the minimum number of for-loops you need to perform your optimization?

[ "Two for-loops, one to iterate over the weights and the other to iterate over the epochs", "Only one for-loop to iterate over the epochs.", "Only one for-loop to iterate over the weights.", "No for-loop is really necessary. Everything can be vectorized" ]

B

Fill the missing line of code: (one answer)\\ \hspace*{.5cm} \#code missing\\ \hspace*{.5cm} np.mean(np.random.randn(1000))\\

[ "import np", "import numpy", "import numpy as np", "import np.mean\\\\\n\t\timport np.random" ]

C

What is the output of the following block of Python code? (one answer) \\ \verb|my_string = `computational'| \\ \verb|print(my_string[1])|\\ \verb|print(my_string[3:5])| \vspace{0.25cm}

[ "c\\\\mpu", "c\\\\mp", "o\\\\put", "o\\\\pu" ]

D

You are using a 3-layer fully-connected neural net with \textbf{ReLU activations}. Your input data has components in [0, 1]. \textbf{You initialize all your weights to -10}, and set all the bias terms to 0. You start optimizing using SGD. What will likely happen?

[ "The gradient is 0 so nothing happens", "The gradient is very large so the model can't converge", "Training is fine, but our neural net does only as well as a linear model", "Everything is fine" ]

A

You are using a 3-layer fully-connected neural, and you are using \textbf{$f(x) = 2x$ as your activation function} . Your input data has components in [0, 1]. \textbf{You initialize your weights using Kaiming (He) initialization}, and set all the bias terms to 0. You start optimizing using SGD. What will likely happen?

[ "The gradient is 0 so nothing happens", "The gradient is very large so the model can't converge", "Training is fine, but our neural net does only as well as a linear model", "Everything is fine" ]

C

What is a good representation for scores when classifying these three target classes: Car, Bike and Bus, in the context of logistic regression. (One or multiple answers)

[ "{Car: $(0,1,0)$,} {Bike: $(1,0,0)$,} {Bus: $(0,0,1)$}", "{Car: $(0,1)$,} {Bike: $(1,0)$,} {Bus: $(1,1)$}", "{Car: $1$,} {Bike: $2$,} {Bus: $3$}", "{Car: $(0,1)$,} {Bike: $(1,0)$,} {Bus: $(0.5,0.5)$}" ]

A

Mean Square Error loss:

[ "Minimizing the distance between the predicted point and the true point", "Maximizing the probability of the correct class", "Minimizing the score of false classes when they are close, or bigger than, the score of the true class", "Maximizing the accuracy" ]

A

You need to debug your Stochastic Gradient Descent update for a classification of three bridge types. Manually compute the model output for the feature vector $x=(1, 0, 0, 0, 0)$ and $W$ contains only zeros. The model is logistic regression, \textit{i.e.}, $\textrm{softmax}(Wx)$. Remember: \begin{equation} \textrm{softmax}_i(s) = \frac{e^{s_i}}{\sum_k e^{s_k}} \end{equation} (One answer!!!!!!)

[ "$(0, 0, 0)$", "$(\\frac{1}{3}, \\frac{1}{3}, \\frac{1}{3})$", "$(0, 0, 0, 0, 0)$", "$(\\frac{1}{5}, \\frac{1}{5}, \\frac{1}{5}, \\frac{1}{5}, \\frac{1}{5})$" ]

B

Consider the following PyTorch code: class ThreeLayerNet (nn.Module): def __init__(): super().__init__() def forward(x): x = nn.Linear(100, 10)(x) x = nn.ReLU()(x) x = nn.Linear(10, 200)(x) x = nn.ReLU()(x) x = nn.Linear(200, 1)(x) return x Suppose that inputs are 100-dimensional, and outputs are 1-dimensional. What will happen if we try to train this network?

[ "There will be an error because we are re-using the variable x throughout the forward() method.", "There will be an error because the second layer has more neurons than the first. The number of neurons must never increase from one layer to the next.", "The model will not train properly. The performance will be the same at the beginning of the first epoch and at the end of the last epoch.", "Everything is fine." ]

C

You are using a 3-layer fully-connected neural net with \textbf{ReLU activations}. Your input data has components in [0, 1]. \textbf{You initialize your weights by sampling from $\mathcal{N}(-10, 0.1)$ (Gaussians of mean -10 and variance 0.1)}, and set all the bias terms to 0. You start optimizing using SGD. What will likely happen?

[ "The gradient is 0 so nothing happens", "The gradient is very large so the model can't converge", "Training is fine, but our neural net does only as well as a linear model", "Everything is fine" ]

A

We saw in class that we can quickly decrease the spatial size of the representation using pooling layers. Is there another way to do this without pooling?

[ "Yes, by increasing the amount of padding.", "Yes, by increasing the stride.", "Yes, by increasing the number of filters.", "No, pooling is necessary." ]

B

The \textbf{parameters} (weights \textbf{W}) are learned with ... (One answer)

[ " training ", " validation ", " test ", " all the data together " ]

A

The \textbf{hyperparameters} are learned with ... (One answer)

[ " training ", " validation ", " test ", " all the data together " ]

B

We report the final performance (e.g., accuracy) on the ... (One answer)

[ " training ", " validation ", " test ", " all the data together " ]

C

We consider a classification problem on linearly separable data. Our dataset had an outlier---a point that is very far from the other datapoints in distance (and also far from margins in SVM but still correctly classified by the SVM classifier). We trained the SVM, logistic regression and 1-nearest-neighbour models on this dataset. We tested trained models on a test set that comes from the same distribution as training set, but doesn't have any outlier points. For any vector $v \in \R^D$ let $\|v\|_2 := \sqrt{v_1^2 + \dots + v_D^2}$ denote the Euclidean norm. The hard-margin SVM problem for linearly separable points in $\R^D$ is to minimize the Euclidean norm $\| \wv \|_2$ under some constraints. What are the additional constraints for this optimization problem?

[ "$y_n \\ww^\top x_n \\geq 1 ~ \forall n \\in \\{1,\\cdots,N\\}$", "$\\ww^\top x_n \\geq 1 ~ \forall n \\in\\{1,\\cdots,N\\}$", "$y_n + \\ww^\top x_n \\geq 1 ~ \forall n \\in \\{1,\\cdots,N\\}$", "$\frac{y_n}{\\ww^\top x_n }\\geq 1 ~\forall n \\in \\{1,\\cdots,N\\}$" ]

A

Consider the function $f(x)=-x^{2}$. Which of the following statements are true regarding subgradients of $f(x)$ at $x=0$ ?

[ "A subgradient does not exist as $f(x)$ is differentiable at $x=0$.", "A subgradient exists but is not unique.", "A subgradient exists and is unique.", "A subgradient does not exist even though $f(x)$ is differentiable at $x=0$." ]

D

When constructing a word embedding, what is true regarding negative samples?

[ "They are words that do not appear as context words", "They are selected among words which are not stop words", "Their frequency is decreased down to its logarithm", "They are oversampled if less frequent" ]

D

If the top 100 documents contain 50 relevant documents

[ "the precision of the system at 50 is 0.25", "the precision of the system at 100 is 0.5", "the recall of the system is 0.5", "All of the above" ]

B

What is WRONG regarding the Transformer model?

[ "It uses a self-attention mechanism to compute representations of the input and output.", "Its computation cannot be parallelized compared to LSTMs and other sequential models.", "Its complexity is quadratic to the input size.", "It captures the semantic context of the input." ]

B

Which of the following statements about index merging (when constructing inverted files) is correct?

[ "While merging two partial indices on disk, the inverted lists of a term are concatenated without sorting", "Index merging is used when the vocabulary does no longer fit into the main memory", "The size of the final merged index file is O (n log2 (n) M )), where M is the size of the available memory", "While merging two partial indices on disk, the vocabularies are concatenated without sorting" ]

A

Which of the following statements on Latent Semantic Indexing (LSI) and Word Embeddings (WE) is false?

[ "The dimensions of LSI can be interpreted as concepts, whereas those of WE cannot", "LSI does not depend on the order of words in the document, whereas WE does", "LSI is deterministic (given the dimension), whereas WE is not", "LSI does take into account the frequency of words in the documents, whereas WE with negative sampling does not" ]

D

The number of non-zero entries in a column of a term-document matrix indicates:

[ "how many terms of the vocabulary a document contains", "how often a term of the vocabulary occurs in a document", "how relevant a term is for a document", "none of the other responses is correct" ]

D

Which of the following statements on Latent Semantic Indexing (LSI) and Word Embeddings (WE) is incorrect

[ "LSI is deterministic (given the dimension), whereas WE is not", "LSI does not take into account the order of words in the document, whereas WE does", "The dimensions of LSI can be interpreted as concepts, whereas those of WE cannot", "LSI does take into account the frequency of words in the documents, whereas WE does not." ]

D

Suppose that in a given FP Tree, an item in a leaf node N exists in every path. Which of the following is true?

[ "N co-occurs with its prefixes in every transaction", "{N}’s minimum possible support is equal to the number of paths", "For every node P that is a parent of N in the FP tree, confidence(P->N) = 1", "The item N exists in every candidate set" ]

B

Which of the following statements regarding topic models is false?

[ "Topic models map documents to dense vectors", "In LDA, topics are modeled as distributions over documents", "LDA assumes that each document is generated from a mixture of topics with a probability distribution", "Topics can serve as features for document classification" ]

B

Which of the following statements is correct concerning the use of Pearson’s Correlation for user- based collaborative filtering?

[ "It measures whether different users have similar preferences for the same items", "It measures how much a user’s ratings deviate from the average ratings I", "t measures how well the recommendations match the user’s preferences", "It measures whether a user has similar preferences for different items" ]

A

After the join step, the number of k+1-itemsets

[ "is equal to the number of frequent k-itemsets", "can be equal, lower or higher than the number of frequent k-itemsets", "is always higher than the number of frequent k-itemsets", "is always lower than the number of frequent k-itemsets" ]

B

Which is true about the use of entropy in decision tree induction?

[ "The entropy of the set of class labels of the samples from the training set at the leaf level is always 0", "We split on the attribute that has the highest entropy", "The entropy of the set of class labels of the samples from the training set at the leaf level can be 1", "We split on the attribute that has the lowest entropy" ]

C

Which of the following models for generating vector representations for text require to precompute the frequency of co-occurrence of words from the vocabulary in the document collection

[ "LSI", "CBOW", "Fasttext", "Glove" ]

D

Regarding the Expectation-Maximization algorithm, which one of the following false?

[ "Assigning equal weights to workers initially decreases the convergence time", "The label with the highest probability is assigned as the new label", "It distinguishes experts from normal workers", "In E step the labels change, in M step the weights of the workers change" ]

A