This repository focus on reconstructing density matrices using measurements.

For reproducibility's sake, every notebook has a google colab link that you can click on and run the notebook on your browser. In the future I plan on doing a requirements.txt for pip instalation to run on a local machine.

In the notebooks you will find the results for several methods for reconstruction of the density matrices. For example:

• Reconstruction_MSE.ipynb : We used the Mean Squared Error of measurements to reconstruct.

• Reconstruction DM.ipynb : We used the trace distance between the reconstructed density matrix and original density matrix.

• Analysis Results.ipynb : Used for analysing the results obtained by other notebooks.

• Autoencoder Benchmark.ipynb : We use Autoencoders to benchmark our reconstruction results.

• Autoencoder Benchmark.ipynb : We use Autoencoders to benchmark our reconstruction results.

• CreateMeasurements.py : Python file to create the dataset.

• Data: (Data used)

  • Measurements: Measurement data with X_train.txt and X_test.txt.

• Results: (Results obtained)

  • AE: Results for the Autoencoder with MSE as a loss function.
  • TAE: Results for the Autoencoder with Trace Distance as a loss function.
  • TVAE: Results for the Variational Autoencoder with Trace Distance as a loss function.

• Utils:

  • Dataset.py : Creating and handling the dataset.
  • Plotter.py : Plotting during training.
  • QutipUtils.py : Qutip helper functions.
  • QMeasures.py : Tensorflow implementation of quantum metrics: trace distance,etc.

• Models: (Deep Learning Models used)

  • VAE_Keras.py : Variational Autoencoder implementation on keras.
  • TVAE.py : Variational Autoencoder implementation using trace distance as a loss function.
  • EVAE.py : Variational Autoencoder implementation using quantum entropy measurements as a loss function. (Not completed)
  • AE.py : Autoencoder implementation on keras.
  • TAE.py : Autoencoder implementation using trace distance as a loss function.