Welcome to the official GitHub home of AI4PhysSci Lab, based at the Hong Kong University of Science and Technology (HKUST)!
We are an interdisciplinary AI for Science lab aimming to build Simulators, Emulators, Predictors, and Evaluators for multiscale modeling workflows. Rooted in AI x quantum x theoretical chemistry, we have broad interests spanning machine learning, quantum computing, molecular simulation, property prediction, generative molecular and material design and experiment design. Our mission is to develop useful, not just fancy tools that revolutionize the paradigm of chemistry research & accelerate chemical discovery through the integration of data-driven models and physics-based principles.
We build general, practical, and scalable tools to solve real-world scientific problems in chemistry. Our current efforts span:
- 🧠 Deep Quantum Monte Carlo (Deep QMC): Wavefunction-based neural network solvers that deliver high-accuracy results for strongly correlated systems.
- 🧪 Orbital-based learning: Predicting physically constrained electron densities in real space.
- 🔍 AI for material design: Generative AI for material design problems.
- 🔬 AI for experiment deisgn: Bayesian optimization or reinforcement learning tools for guided experimental design and scientific exploration.
- 🌐 LLMs for Science: Understanding and guiding the use of foundation models in scientific discovery.
We welcome students, researchers, developers, and collaborators from all backgrounds—whether your home is in chemistry, physics, CS, or statistics.
If you're interested in contributing:
- Submit an issue or pull request to one of our repositories.
- Please make sure you run pytest (if any) locally before you PR your branch
- Code review is required for merging the PR.
- More contribution guide will be released
- Reach out us on GitHub or via email for any general questions.
- 🌍 Lab Homepage
- 👧🏻 Sherry's Homepage
- 🏘️ Department of Chemistry, HKUST
- 🧾 Recent Publication: