inverse-design

AI-driven inverse antenna design with real NEC2 + openEMS in the loop. Try the live in-browser playground.

torch-molecule is a deep learning package for molecular discovery, designed with an sklearn-style interface for property prediction, inverse design and representation learning.

AeroJAX: A differentiable, structure-preserving framework for real-time flow simulation, control, and inverse design. Architected for neural operator integration and latent-space acceleration. Built with JAX.

gRNAde is a Generative AI framework for inverse design of 3D RNA structure and function

Differentiable wave optics simulation library built on PyTorch

This repository hosts a simple demonstration of a deep learning approach for the inverse design of patch antennas. The goal is to explore energy-efficient designs and to significantly reduce simulation cost compared to conventional methods.

GPU-accelerated FDTD Python framework for fast, differentiable, large scale 2D/3D electromagnetic simulations and photonic inverse design.

[NeurIPS 2020] Diversity-Guided Efficient Multi-Objective Optimization With Batch Evaluations

Optimization and inverse design of photonic chips using deep reinforcement learning

Electromagnetics simulation library for moving point charges built on JAX

Efficient GPU-computing simulation for differentiable crystal plasticity finite element method

A Generic Framework for Optical Inverse Design

A PyTorch Implementation of "Optimization of Molecules via Deep Reinforcement Learning".

A collection of inverse design challenges

Silicon Photonics Design Tools.

Verifier-first transfer-matrix workbench for multilayer optics and differentiable inverse design

AI-assisted Photonic Device Inverse Design Framework, MAPS DATE 2025

Differentiable Flexible Mechanical Metamaterials

[ICLR 2024] CinDM uses compositional generative models to design boundaries and initial states significantly more complex than the ones seen in training for physical simulation

A Fortran-based neural network library for physics-based applications. Alongside standard neural network layer types, it also supports graph-based layers and physics informed neural networks.