[ICLR 2022] Accelerated Policy Learning with Parallel Differentiable Simulation

Reinforcement learning in differentiable multiphysics simulation with NVIDIA Warp.

A library for soft differentiable relaxations of common JAX functions.

Plug-and-Play Benchmarking of Reinforcement Learning Algorithms for Large-Scale Flow Control

This repo contains the differentiable physics simulation module in "PPR: Physically Plausible Reconstruction from Monocular Videos". ICCV 23.

A Python based framework for differentiable power system simulation and dynamic optimization

PyTorch and Taichi implementations of our paper on improving gradient computation

A differentiable approach to multi-scale brain modeling.

Simulation platform used in "Artificial Life" (CS302) at Northwestern University

Code for "Accelerated co-design of robots through morphological pretraining", Strgar & Kriegman, Proceedings of the International Conference on Learning Representations (ICLR) 2026

VirtualPatients is a differentiable biophysiological simulation platform that enables users to design and evaluate healthcare interventions for in-silico populations and general-purpose healthcare applications.

Modular Automatic Differentiation and Data-Enhanced Neural-network INteracting Graph (MADDENING). A pure JAX end-to-end autodifferentiable graph based framework for multiphysics simulation.

MIcrorobotics Multiphysics Engine (MIME), Differentiable physics for medical microrobots (currently focused on magnetically actuated and in confined biological flows). Stokeslet BEM, IB-LBM, and DiFVM inspired FVM-IBM (with optional GNN correction); built on MADDENING.