Yimeng Min

Research Interests

My work sits at the intersection of artificial intelligence and its applications in the natural sciences. I draw on physics, geometry, and dynamical systems to build more efficient generative models and to understand why they work, and I am interested in AI systems that are physically grounded rather than purely data-driven. In AI for Science and Sustainability, I use AI-guided discovery to find new functional materials, validated by real experiments.

Some current research directions include:

Discovery Artificial intelligence for scientific discovery and sustainability
Generative Efficient generative AI models and how they work
Geometric Geometric deep learning and neural combinatorial optimization

Selected Publications

Miao Zhong^*, Kevin Tran^*, Yimeng Min^*, Chuanhao Wang^*, et al. "Accelerated Discovery of CO₂ Electrocatalysts Using Active Machine Learning." Nature, 2020.

Details

We show that machine learning can accelerate the search. Using computer models and theoretical data, algorithms can toss out worst options and point the way toward more promising candidates. The new catalyst we found is the first one for CO₂-to-ethylene conversion to have been designed in part through the use of AI.

Yimeng Min, Carla P. Gomes. "Structure As Search: Unsupervised Permutation Learning for Combinatorial Optimization." New Perspectives in Graph Machine Learning workshop at NeurIPS, 2025.

Yimeng Min^*, Yiwei Bai^*, Carla P. Gomes. "Unsupervised Learning for Solving the Travelling Salesman Problem." NeurIPS, 2023.

Yimeng Min^*, Frederik Wenkel^*, Michael Perlmutter, Guy Wolf. "Can Hybrid Geometric Scattering Networks Help Solve the Maximum Clique Problem?" NeurIPS, 2022.

Yimeng Min^*, Frederik Wenkel^*, Guy Wolf. "Scattering GCN: Overcoming Oversmoothness in Graph Convolutional Networks." NeurIPS, 2020.

^*Equal contribution