netket's Introduction

NetKet

NetKet is an open-source project delivering cutting-edge methods for the study of many-body quantum systems with artificial neural networks and machine learning techniques.

Major Features

Graphs
- Built-in Graphs
  - Hypercube
- Custom Graphs
  - Any Graph With Given Adjacency Matrix [from input file]
  - Any Graph With Given Edges [from input file]
- Symmetries
  - Automorphisms: pre-computed in built-in graphs, available through iGraph for custom graphs
Hamiltonians
- Built-in Hamiltonians
  - Transverse-field Ising
  - Heisenberg
  - Bose-Hubbard
- Custom Hamiltonians
  - General k-local Hamiltonians defined on Graphs
  - Any k-local Hamiltonian [from input file]
Learning
- Optimizers
  - Stochastic Gradient Descent
  - AdaMax, AdaDelta, AdaGrad, AMSGrad
  - RMSProp
  - Momentum
  - Gradient Clipping
- Ground-state Learning
  - Gradient Descent
  - Stochastic Reconfiguration Method
    - Direct Solver
    - Iterative Solver for Large Number of Parameters
Machines
- Restricted Boltzmann Machines
  - Standard
  - For Custom Local Hilbert Spaces
  - With Permutation Symmetry Using Graph Isomorphisms
- Feed-Forward Networks
  - For Custom Local Hilbert Spaces
  - Fully connected layer
  - Convnet layer for arbitrary underlying graph
  - Any Layer Satisfying Prototypes in AbstractLayer [extending C++ code]
- Jastrow wavefunction
  - Standard
  - With Permutation Symmetry Using Graph Isomorphisms
- Custom Machines
  - Any Machine Satisfying Prototypes in AbstractMachine [extending C++ code]
Observables
- Custom Observables
  - Any k-local Operator [from input file]
Sampling
- Local Metropolis Moves
  - Local Hilbert Space Sampling
  - Parallel Tempering Versions
- Hamiltonian Moves
  - Automatic Moves with Hamiltonian Symmetry
  - Parallel Tempering Versions
- Custom Sampling
  - Any k-local Stochastic Operator can be used to do Metropolis Sampling
Statistics
- Automatic Estimate of Correlation Times
I/O
- Python/JSON Interface