physics-informed-machine-learning

Code accompanying my blog post: So, what is a physics-informed neural network?

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.

Introductory workshop on PINNs using the harmonic oscillator

A carefully curated collection of high-quality libraries, projects, tutorials, research papers, and other essential resources focused on Physics-Informed Machine Learning (PIML) and Physics-Informed Neural Networks (PINNs).

Workshop on designing efficient and scalable physics-informed neural networks with JAX

Official implementation of Neural Lithography (SIGGRAPH Asia 2023)

Official repo for separable operator networks -- extreme-scale operator learning for parametric PDEs.

A physics-informed deep learning (DL)-based constitutive model for investigating epoxy based composites under different ambient conditions.

AI-native open-source finite-element platform connecting engineering, computation, data, and AI.

Curso teórico práctico sobre inversión geofísica en Python: SimPEG, gravimetría, magnetometría, MT 1D y FWI.

Implementation of BubbleOKAN and the Two-Step DeepOKAN framework — a physics-informed neural operator for high-frequency bubble dynamics (Rayleigh–Plesset, Keller–Miksis). Uses continual learning to overcome spectral bias.

Implementation of Anant-Net, a scalable, interpretable physics-informed neural surrogate for solving high-dimensional PDEs at O(Nd) cost via input-space partitioning. Demonstrated for MLPs and KANs.

"Bayesian Enhanced AoA Estimator: A Physics-Informed Machine Learning Approach for Accurate Angle of Arrival Estimation". This repository is an AoA estimator for passive UHF RFID based on Bayesian regression and classical antenna array signal processing. Combines physics-informed analysis with Pyro-based uncertainty quantification.

📕Code for paper Parallelizable Complex Neural Dynamics Models for PMSM Temperature Estimation with Hardware Acceleration.

Falsification-first scientific formula discovery with LLMs, symbolic regression, dimensional analysis, Pareto selection, and LLM-SRBench evaluation. Codex API by default.

Main codes for half-cell model, PINN and co-kriging implemented for physics-informed degradation diagnostics project: https://doi.org/10.1016/j.ensm.2024.103343

Physics-Informed Machine Learning for Precision UAV Control: Adaptive Transformers with Safety Guarantees

NeedForHat Diagnosis: physics informed machine learning models for the residential energy transition

LieNLSD: explicit nonlinear Lie symmetry discovery from dynamical-system and PDE data; discovers Lie algebra dimension and symbolic infinitesimal generators. ICML 2025.

A comparative analysis of DeepONet and FNO architectures, benchmarking their performance on Function-to-Function (Heat Equation) vs. Parameter-to-Function (Elastic Bar) PDE problems to motivate hybrid operator designs.