We work on Machine Learning techniques for Quantum Physics, in particular we are developing Neural Quantum States Variational Monte Carlo approaches. We are a team of researchers based in Paris, working between Ecole Polytechnique and College de France.
We are the main contributors behind NetKet, but we also develop additional codes that do not make their way into NetKet directly.
You can find here some of our public works
- NQXPack : A simple way to save and load flax-based NN models. It provides an easy to use solution to save and load NetKet's variational states.
- PTVMC Systematic Study : State-of-the-art implementation of Infidelity optimization for NetKet, as well as infidelity-based dynamics, with autotuning of hyperparameters.
- ConvNext NNQS : Repository accompanying the article "Design principles of deep translationally-symmetric neural quantum states for frustrated magnets", implementing a ConvNext-based NQS in Flax.