physics-informed-neural-network

A physics-informed neural network method for head-related transfer function upsampling

Going through the tutorial on Physics-informed Neural Networks: https://github.com/madagra/basic-pinn

This repository contains all Assignments and Lecture Slides from the Physics Informed Machine learning course by Prof. Augustin Guibaud in Spring 2025 at NYU.

Physics informed neural networks (PINN) with a mathematically informed architecture for the nuclear decay equation

Physics-informed neural surrogate for structural analysis — 1000x faster than FEA with >99.9% accuracy

Physics-informed Bidirectional LSTM (DPRNet-Bi) for ventilator airway pressure prediction — deep learning + lung mechanics for biomedical time-series modeling. Validated on Kaggle Ventilator Pressure Prediction dataset with MIMIC-III cross-dataset evaluation.

Official PyTorch implementation of CPA-SHNN: Causality-Preserving Adaptive Symplectic Hamiltonian Neural Networks for chaotic autonomous and non-autonomous multi-body celestial gravitational dynamics.

Physics-Informed Neural Network framework for atmospheric CO2 flux inversion over Europe (ICOS network, 2019). Decoupled fossil/biospheric separation with HYSPLIT transport.

VICTOR (Variational Inference for Confined Tokamak Output Reconstruction) — A Physics-Informed Neural Network that reconstructs plasma emissivity on the WEST Tokamak in ~0.5ms, trained entirely without ground-truth labels. Enabling real-time fusion reactor diagnostics and disruption prevention.

JAX/Flax physics-informed neural network with jax2tf export — benchmark JAX vs PyTorch vs TensorFlow

An automated pipeline that leverages a TensorFlow Physics‑Informed Neural Network (PINN) and AGMA‑style equations to predict material properties and generate optimized 3D gear CAD models. Features a CLI, interactive GUI, and exports STEP/STL files for manufacturing.