Elucidating the Design Choice of Probability Paths in Flow Matching for Forecasting
NVIDIA PhysicsNeMo 中文教程书 · 8 章 + 4 附录 · PINN 与神经算子三种实现（裸 PyTorch / SDK / GPU production）
Published PyPI package for ArXiv embedding benchmarks, retrieval evaluation, and scientific RAG experiments.
Structured ecosystem of 190+ AI systems spanning foundation models, agentic reasoning, reinforcement learning, generative architectures, scientific ML, and self-evolving neural systems. A research-driven lab exploring scalable intelligence.
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.
Physics-informed neural network solving Maxwell's equations in ICP reactors. Hard BC ansatz, transfer learning across geometries, autograd sensitivity maps. 1700× faster than FEM. Built with NVIDIA Modulus + PyTorch.
This project measures whether Cl(3,0) geometric algebra layers bring anything beyond exact SO(3) equivariance on small synthetic 3D vector tasks.
live seismic-data ML pipeline
Parametric PINN surrogate for flow over cylinders — 95% accuracy, 95%+ CFD speedup, <50ms inference
Recovers Burgers viscosity from 32 sparse sensors at 3.4% error vs 34.9% closed-form, on 30 hard out-of-distribution cases. CPU PINN sandbox + live demo.
Solving 1D beam, 2D plate, and 3D solid mechanics problems using Physics-Informed Neural Networks (PINNs) with DeepXDE and TensorFlow. Results validated against ANSYS Mechanical 2025 R1. 2nd sem project
AI-assisted simulation pre-processing tool that validates STEP/IGES files for FEA readiness, auto-heals geometry, and generates mesh recommendations
CLI that applies the Clever Materials falsification test to materials-science ML datasets: detects confounding via a bibliographic fingerprint (year/author/journal leaking the target). Built on Jablonka 2026 (arXiv:2602.17730) + Huang&Cole / Nandy datasets (Scientific Data).
Physics-aware neural surrogate for black hole accretion flow (GRMHD-like) using Fourier Neural Operators.
ReAct-based economic dispatch agent using CVXPY tools for reliable optimization-grounded AI workflows.
GPU-accelerated, fault-tolerant Schlieren/PIV shock tracking with interactive ROI, 1-px edges, and resumable training.
Runge–Kutta Physics-Informed Neural Networks with time-discrete losses (Gauss/Radau/Lobatto) in PyTorch.
FNO vs UNet for parametric PDE surrogate modelling (1D Burgers): resolution-invariance, offline-first, zero-dep core + optional torch.
Cell tracking, lineage reconstruction and division-aware evaluation pipeline for the Biohub microscopy challenge.
Scientific ML project for aqueous solubility prediction with RDKit, FastAPI, Docker, and Azure ML deployment notes.
To associate your repository with the scientific-ml topic, visit your repo's landing page and select "manage topics."