JAX, M.D. A framework for differentiable physics*

DOI10.1088/1742-5468/AC3AE9MaRDI QIDQ5020064zbMATH OpenOpenAlexFDO

Authors Samuel S. Schoenholz, Ekin D. Cubuk

Publication date 3 January 2022

Published in Journal of Statistical Mechanics: Theory and Experiment (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1912.04232

zbMATH Keywords

machine learningnumerical simulationsmolecular dynamics

Mathematics Subject Classification ID

Statistical mechanics, structure of matter (82-XX)

Recommendations

• Differentiable coarse graining
• JAX-FEM: a differentiable GPU-accelerated 3D finite element solver for automatic inverse design and mechanistic data science
• An abstract Lagrangian framework for computing shape derivatives
• Deep hidden physics models: deep learning of nonlinear partial differential equations
• Discrete mechanics and variational integrators
• Natural discrete differential calculus in physics
• Hidden physics models: machine learning of nonlinear partial differential equations
• Differentiable manifolds. A theoretical physics approach
• Differentiable manifolds. A theoretical physics approach

Cites work

• scientific article; zbMATH DE number 6982909 (Why is no real title available?)
• Automatic differentiation of explicit Runge-Kutta methods for optimal control
• Efficient adjoint derivatives: application to the meteorological model meso-nh
• Fast parallel algorithms for short-range molecular dynamics
• General purpose molecular dynamics simulations fully implemented on graphics processing units
• Human-level concept learning through probabilistic program induction
• On the performance of discrete adjoint CFD codes using automatic differentiation
• The frontier of simulation-based inference

Cited in

(6)

• JAX-fluids 2.0: towards HPC for differentiable CFD of compressible two-phase flows
• JAX-FEM: a differentiable GPU-accelerated 3D finite element solver for automatic inverse design and mechanistic data science
• A Model-Constrained Tangent Slope Learning Approach for Dynamical Systems
• Learning Lagrangian fluid mechanics with E(3)-equivariant graph neural networks
• Differentiability in unrolled training of neural physics simulators on transient dynamics
• JAX MD

Describes a project that uses

Uses Software

• Torch
• Theano
• NumPy
• TensorFlow
• IPython
• DiffSharp
• OpenMM
• PyTorch
• Autograd
• DeepChem
• JAX
• Omniglot
• GitHub
• Boids
• Haiku
• Taichi
• DiffTaichi
• Zygote
• HOOMD-TF
• SchNet

This page was built for publication: JAX, M.D. A framework for differentiable physics*

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5020064)