This a project aimed at replicating the results from the arXiV paper on using Quantum Autoencoders for Quantum Error Correction(qae-paper). The results in this paper was achieved by using DQNNs (Disspiative Quantum Neural Networks) with code written in MATLAB and they are numerical simulation results.
Desired Objectives for this project:
- Replicate the results but using Python. Insiped from DQNN Repo
- Make datasets and dataset generation code public.
- Run the learned encodings and decodings on actual hardware.
- Try gradient free methods for neural network training, trying an alternate solution for the Barren Plateau problem.
https://1drv.ms/p/s!AhzKZHA1xnhDiK9wIC2fFvj9QZDEKg?e=LzqNnt - Short presentation explaning the theory of the paper.
What is new in this repo, compared to the autoencoder implementation of DQNN Repo.
- methods to perform training for autoenoders in self inverse architecture of the autoencoders from qae-paper.
- datasets for training available as pandas DataFrames
- Circuit Implementation of the autoencoder as a VQA, with aim of making it hardware compatible.