NetKet is an open-source project delivering advanced methods for the study of many-body quantum systems with artificial neural networks and machine learning techniques. It is the first open-source platform supporting collaborative developments in the field and aims to be a robust yet highly responsive reference implementation for both consolidated and new, more experimental, techniques.
One of the main features of this software is the ability to find the ground state of interacting Hamiltonians using neural network–based ansatz states for the many-body wave function. Because of the modular infrastructure of the library, it is possible to highly customize most of its components. For example, changing Hamiltonians, observables and other problem-dependent quantities is meant to be easy and does not require an in-depth knowledge of programming languages.
To stimulate a large-scale conceptual and practical development of the software, NetKet has introduced a series of “Challenges” tasks, to be tackled by researchers worldwide.