Update README.md
Browse files
README.md
CHANGED
@@ -1,4 +1,90 @@
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
pipeline_tag: tabular-regression
|
4 |
-
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
|
1 |
---
|
2 |
license: apache-2.0
|
3 |
pipeline_tag: tabular-regression
|
4 |
+
datasets:
|
5 |
+
- Allanatrix/QST
|
6 |
+
tags:
|
7 |
+
- Physics
|
8 |
+
---
|
9 |
+
# NexaQST: Quantum State Tomography with Physics-Informed Neural Networks
|
10 |
+
|
11 |
+
**NexaQST** is a Quantum State Tomography model built using a Physics-Informed Neural Network (PINN) trained on synthetic 2-qubit experiments. This model leverages quantum mechanical priors such as positivity, Hermiticity, and trace constraints to ensure physically plausible reconstructions of quantum states.
|
12 |
+
|
13 |
+
---
|
14 |
+
|
15 |
+
## Model Overview
|
16 |
+
|
17 |
+
- **Task**: Reconstruct quantum states (density matrices) from tomographic measurement traces
|
18 |
+
- **System**: Simulated 2-qubit experiments
|
19 |
+
- **Architecture**: Physics-Informed Neural Network (PINN)
|
20 |
+
- **Constraints Embedded**:
|
21 |
+
- Positivity
|
22 |
+
- Hermiticity
|
23 |
+
- Trace normalization (Tr(Ο) = 1)
|
24 |
+
|
25 |
+
---
|
26 |
+
|
27 |
+
## Dataset Generation
|
28 |
+
|
29 |
+
- **Simulation Toolkit**: [`QuTiP`](https://qutip.org/)
|
30 |
+
- **Process**:
|
31 |
+
1. Generated full 2-qubit tomography experiments via simulation
|
32 |
+
2. Extracted and structured measurement traces
|
33 |
+
3. Derived backward synthetic traces from known density matrices
|
34 |
+
4. Created supervised training pairs: `(trace β density matrix)`
|
35 |
+
5. Enforced physical constraints directly inside the model during training
|
36 |
+
|
37 |
+
---
|
38 |
+
|
39 |
+
## Model Input/Output
|
40 |
+
|
41 |
+
- **Input**: Vector of tomographic measurement traces
|
42 |
+
- **Output**: Reconstructed density matrix (complex-valued 4x4 for 2 qubits)
|
43 |
+
|
44 |
+
---
|
45 |
+
|
46 |
+
## Example Usage
|
47 |
+
|
48 |
+
```python
|
49 |
+
# Pseudocode for usage
|
50 |
+
|
51 |
+
from model import NexaQSTModel
|
52 |
+
import torch
|
53 |
+
|
54 |
+
model = NexaQSTModel()
|
55 |
+
model.load_state_dict(torch.load("nexaqst_model.pt"))
|
56 |
+
model.eval()
|
57 |
+
|
58 |
+
traces = load_measurement_vector("qst_trace.npy") # shape: (N,)
|
59 |
+
rho_pred = model.predict_density_matrix(traces)
|
60 |
+
|
61 |
+
# rho_pred is a 4x4 complex-valued matrix satisfying physical constraints
|
62 |
+
````
|
63 |
+
|
64 |
+
---
|
65 |
+
|
66 |
+
## Applications
|
67 |
+
|
68 |
+
* Quantum error correction diagnostics
|
69 |
+
* Quantum system identification
|
70 |
+
* Educational simulation of QST techniques
|
71 |
+
* Physically consistent state estimation for quantum simulations
|
72 |
+
|
73 |
+
---
|
74 |
+
|
75 |
+
## Licensing & Citation
|
76 |
+
|
77 |
+
* If used in publications or products, please credit the Nexa Scientific Suite.
|
78 |
+
|
79 |
+
---
|
80 |
+
|
81 |
+
## π Related Tools
|
82 |
+
|
83 |
+
* [π Nexa Data Studio](https://huggingface.co/spaces/Allanatrix/NexaDataStudio)
|
84 |
+
* [π¬ Nexa R\&D Hub](https://huggingface.co/spaces/Allanatrix/NexaR&D)
|
85 |
+
* [π§ Nexa MOE Models](https://huggingface.co/collections/Allanatrix/nexa-models)
|
86 |
+
* [π Nexa Hub (Main Portal)](https://huggingface.co/spaces/Allanatrix/NexaHub)
|
87 |
+
|
88 |
+
---
|
89 |
+
|
90 |
+
*Created by [Allan](https://huggingface.co/Allanatrix), independent quantum systems architect and ML researcher. Part of the Nexa scientific computing ecosystem.*
|