News

Announcing NeurIPS 2026 RealPDE Competition! $21000 prize pool, 2 tracks, compete and develop in the first NeurIPS Scientific ML competition centered on data from real physical systems!
BuildArena 2.0 Challenge is now open! Collaborate with Agents to build rockets that can circle the Earth! $1000 prize pool! [Challenge homepage][Kaggle][知乎]
Our lab has 7 papers accepted at ICLR! [X]
西湖大学解锁 AI 新角色：基建工程师, 火箭桥梁都能造! 首次让语言驱动工程实现闭环 [DeepTech 深科技报道]
We release BuildArena, where LLM agents design, build, and test rockets, cars, and bridges in a physics simulator! With BuildArena, AIs turn words into designs, blueprints, and finally, 3D mechanisms tested in real-time physics!
实验室博士后魏龙将加入复旦大学人工智能创新与产业研究院任助理教授
Our CL-DiffyCon paper is nominated Outstanding Youth Paper Award at China Embodied AI Conference (CEAI 2025)
实验室博士后魏龙，荣获2024年度西湖大学优秀博士后
Co-organize NeurIPS 2024 Safe Generative AI Workshop
MIT Technology Review:西湖大学吴泰霖：用 AI 助力科学仿真和发现，可控权衡预测精度和计算效率
ICLR 2024 Spotlight for our paper “Compositional Generative Inverse Design”.
AAAI 2024 oral for our paper “Uncertainty Quantification for Forward and Inverse Problems of PDEs via Latent Global Evolution”
ICLR 2023 Spotlight for our paper “Learning Controllable Adaptive Simulation for Multi-resolution Physics”.
I have been organizing a bi-weekly Reading Group on Simulation and Inverse Design, with >20 participants (Ph.Ds, postdocs, assistant professor) from CS, EE, ME, and Energy departments from Stanford, MIT, and Caltech. Each time we spend 1h discussing a relevant paper deeply (Thursday 4-5pm PT, Friday 7am Beijing time). Contact me if you want to participate!
Invited talk at Lawrence Livermore National Laboratory (LLNL)’s DDPS seminar, “Learning to accelerate large-scale physical simulations”.
Guest lecture at Caltech CS159, “AI Physicist and Machine Learning for Simulations”
Organized “Deep Learning for Simulation” workshop in ICLR 2021
Our paper Toward an AI Physicist for Unsupervised Learning published at Physical Review E as part of a PRE Spotlight on Machine Learning in Physics.
Our paper Graph Information Bottleneck was featured in Synced AI Technology & Industry Review (机器之心).
Best Poster Award at ICML Time Series Workshop