QDisc

Interpretable Machine Learning • Quantum Physics • Scientific Discovery

Get started | Documentation | Tutorials

QDisc is a library for discovering quantum phenomena from raw quantum data via interpretable machine learning. Without any prior knowledge on the input data, the pipeline aims to uncover unknown regimes and find order parameters that can help us identify interesting physics in a human readable form.

The pipeline is composed of two main components: first, a variational autoencoder (VAE) learns representations that compress the input into the minimal set of physical parameters needed to understand the data. Second, and based on the learned representations, a symbolic regression (SR) module finds closed form equations able to distinguish the different identified regimes.

Getting Started

To begin using the library, we recommend starting with the tutorial notebooks. We cover the two pillars of the Qdisc pipeline:

1. Using variational autoencoders to extract interpretable representations (link)
2. Finding closed forms for order paramters via symbolic regression (link)

Additional example notebooks are used to illustrate specific applications, as those covered in the manuscript (TODO). For these examples, the quantum data are not provided:

1. Measurement snapshots of Rydberg atom systems (link).
2. Classical shadows of the cluster Ising model (link).
3. Hybrid data from a fermionic system (link).

Usage

Installation

Install latest from the GitHub repository:

$ pip install git+https://github.com/PaulinDS/qdisc.git

or from pypi

$ pip install qdisc

Documentation

Documentation can be found hosted on this GitHub repository’s pages. Additionally you can find package manager specific guidelines on conda and pypi respectively.

Citing

If you find the librairy usefull in your projects, please cite the accompanying paper:

TODO add the citation