physics-informed

Physics-Informed Neural networks for Advanced modeling

Physics-constrained auto-regressive convolutional neural networks for dynamical PDEs

Dimensionless learning

Code for Rice et al. 2020 "Analyzing Koopman approaches to physics-informed machine learning for long-term sea-surface temperature forceasting"

Does your AI hydrologic model conserve mass, energy and momentum?

TensorFlow 2.0 implementation of Yibo Yang, Paris Perdikaris’s adversarial Uncertainty Quantification in Physics Informed Neural Networks (UQPINNs).

JAX/Flax neural operators for scientific machine learning, with benchmarks and pretrained weights.

WinDiNet: Pretrained Video Models as Differentiable Physics Simulators for Urban Wind Flows

Advisory water treatment for industrial cooling towers: risk indices, anomaly detection, forecasting and dose recommendations. A human authorizes every dose.

Source code of the Paper "Physics-Informed Generative Modeling of Wireless Channels" (ICML 25)

Ununennium (Element 119, the next alkali metal) represents the cutting edge of satellite imagery machine learning. This library provides a unified, GPU-first framework for end-to-end Earth observation workflows, from cloud-native data access through model training to deployment.

Official code for arXiv:2604.11807 - Physics-Informed State Space Models for Off-Grid Solar Forecasting

Spatiotemporal Gaussian process modeling for environmental data: non-stationary PDE prior, deep kernels, multi-fidelity fusion, and A-optimal sampling.非稳态 PDE + 核深度学习 + 多保真 Co-Kriging + 主动采样的物理约束克里金方法，用于复杂时空环境建模与预测

Open-source Physics-informed Turbojet Digital Twin Platform for Engine Health Monitoring, Prognostics, Explainable AI, and Interactive 3D Visualization.

Neural ODE-based State of Charge (SOC) estimation for Li-ion batteries using the NASA Battery Dataset. Built as a weekend project to explore learned dynamics for battery modeling, with visualizations designed for engineering audiences.

RVAV: a physics-informed PyTorch optimizer for energy-stable, high-LR training—with a simple closure API, tests, CI, and quickstart.

Differentiable Jacobian damping-ratio physics loss for power-grid trajectory NNs (CSD Engine A). ANDES-validated, repeated-root safe.

🔍 Explore and analyze ununennium, a Python package designed for efficient data manipulation and visualization. Streamline your data workflows today.

StormFormer: physics-informed hybrid Transformer-BiLSTM for 24h typhoon track prediction on IBTrACS

🌐 Unmix hyperspectral data using the DMTS-Net model, integrating a dual-stream architecture to enhance spectral variability analysis and model performance.