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

Implementation of several neural network (and neural operator) architectures and numerical methods for solving kinetic equations (Boltzmann, Fokker-Planck-Landau, etc.)

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.

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

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

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

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.