RL_SD

Open-source code for Reinforcement Learning and Stochastic Descent(see corresponding directories).

Requirements

The code uses heavily QuSpin (developed by one of the authors). QuSpin can be easily installed via an Anaconda Python 3+ environment, using:

conda install -c weinbe58 quspin

Citing

If you use this code in your research, you may consider citing one of our related papers.

Reinforcement learning in different phases of quantum control (PRX)

@article{PhysRevX.8.031086,
  title = {Reinforcement Learning in Different Phases of Quantum Control},
  author = {Bukov, Marin and Day, Alexandre G. R. and Sels, Dries and Weinberg, Phillip and Polkovnikov, Anatoli and Mehta, Pankaj},
  journal = {Phys. Rev. X},
  volume = {8},
  issue = {3},
  pages = {031086},
  numpages = {15},
  year = {2018},
  month = {Sep},
  publisher = {American Physical Society},
  doi = {10.1103/PhysRevX.8.031086},
  url = {https://link.aps.org/doi/10.1103/PhysRevX.8.031086}
}

The glassy phase of optimal quantum control (PRL)

@article{2018arXiv180310856D,
       author = {{Day}, Alexandre G.~R. and {Bukov}, Marin and {Weinberg}, Phillip and
        {Mehta}, Pankaj and {Sels}, Dries},
        title = "{The Glassy Phase of Optimal Quantum Control}",
      journal = {arXiv e-prints},
     keywords = {Quantum Physics, Condensed Matter - Statistical Mechanics, Mathematics -
        Optimization and Control},
         year = 2018,
        month = Mar,
          eid = {arXiv:1803.10856},
        pages = {arXiv:1803.10856},
archivePrefix = {arXiv},
       eprint = {1803.10856},
 primaryClass = {quant-ph},
       adsurl = {https://ui.adsabs.harvard.edu/\#abs/2018arXiv180310856D},
      adsnote = {Provided by the SAO/NASA Astrophysics Data System}
}

Broken symmetry in a two-qubit quantum control landscape (PRB)

@article{PhysRevA.97.052114,
  title = {Broken symmetry in a two-qubit quantum control landscape},
  author = {Bukov, Marin and Day, Alexandre G. R. and Weinberg, Phillip and Polkovnikov, Anatoli and Mehta, Pankaj and Sels, Dries},
  journal = {Phys. Rev. A},
  volume = {97},
  issue = {5},
  pages = {052114},
  numpages = {12},
  year = {2018},
  month = {May},
  publisher = {American Physical Society},
  doi = {10.1103/PhysRevA.97.052114},
  url = {https://link.aps.org/doi/10.1103/PhysRevA.97.052114}
}