Bing Yan
Portrait of Bing Yan

Bing Yan

Ph.D. candidate in Computer Science, NYU
Ph.D. in Chemistry, MIT
Visiting Researcher, Meta FAIR

CV Google Scholar GitHub X

About

My research integrates specialized AI models with experiments to understand and control the mechanisms of chemical and electrochemical reactions. I build models that distinguish competing mechanistic hypotheses and design the experiments that most sharply resolve them, closing the loop between experiment and computation to predict, and ultimately control, reaction selectivity.

At NYU, I am advised by Prof. Kyunghyun Cho; at Meta, I am mentored by Dr. Ben Miller and Dr. Ricky Chen; and at MIT, I was advised by Prof. Yogesh Surendranath during my PhD and by Prof. Yuriy Roman during my postdoc.

Research Interests

  • Mechanism discrimination: integrating AI models with experimental data to tell competing reaction mechanisms apart.
  • Experiment design in the loop: choosing the next measurements that most sharply resolve mechanistic and kinetic uncertainty.
  • Understanding and controlling selectivity in chemical and electrochemical reactions.
  • Machine learning for chemistry: generative models (diffusion and flow matching) and large language models for reasoning over molecules and reactions.

Education

  • 2022 - present Ph.D. in Computer Science, New York University
    Advisor Prof. Kyunghyun Cho
  • 2014 - 2019 Ph.D. in Chemistry, Massachusetts Institute of Technology
    Advisor Prof. Yogesh Surendranath
  • 2010 - 2014 B.S. in Chemistry, Peking University
    Graduated with Highest Honor; advisor Prof. Song Gao

Awards & Honors

  • 2026 Poster Award, Gordon Research Conference on AI for Materials, Energy, and Chemical Sciences
  • 2022 - 2027 Henry M. MacCracken Fellowship, NYU
  • 2021 Rowland Fellowship, finalist
  • 2018 Moore Fellowship
  • 2014 Graduated with Highest Honor, Peking University (10 laureates every 2 years)

News

  • Feb 6, 2026

    Received a Poster Award at the Gordon Research Conference on AI for Materials, Energy, and Chemical Sciences for EVA-Flow: A Unified Environment-Aware Generative Model for 3D Molecular Conformations.

  • Aug 8, 2025

    Our Consistency paper is out in Digital Discovery. We evaluate how consistently LLMs answer when the same molecule is given as a SMILES string versus an IUPAC name, and find low consistency even after mapped training and KL regularization, suggesting that LLMs may capture surface-level patterns rather than underlying chemistry.

  • Jul 26, 2024

    Our CatScore paper is out in Digital Discovery. We introduce CatScore, a learning-based metric for evaluating asymmetric catalysis in organic chemistry that enables efficient, effective assessment of diverse catalyst-design models at both the instance and system levels.

Selected Publications

See all publications →