High Energy Physics in the Age of AI and Formalization

by Dr Jinzheng Li(李金政） (IBS-CTPU-PTC)

Wednesday 21 Oct 2026, 13:30 → 14:30 Asia/Shanghai

Tsung-Dao Lee Institute/N6F-N601 - Meeting Room (Tsung-Dao Lee Institute)

Abstract:

How much of high-energy physics research can AI do today, and how can we trust its results? From exploring models beyond the Standard Model to deriving their phenomenological consequences, AI-assisted reasoning can conceal subtle errors in assumptions, conventions, and results inherited from the literature. This talk explores how mathematical formalization can make these dependencies explicit and theoretical arguments machine-checkable. Using Lean, I will introduce the principles and limits of formal verification and discuss how AI can assist proof construction and benefit from formal feedback. I will also introduce Physlib, a formalization library for physics, and discuss the gaps between its current coverage and the needs of frontier high-energy physics research.

Speaker Bio:

Jinzheng Li conducted his doctoral research at Northeastern University under the supervision of Professor Pran Nath and will join IBS-CTPU-PTC as a Senior Researcher in November 2026. His research focuses on particle cosmology, including dark matter, hidden-sector thermal histories, first-order phase transitions and stochastic gravitational waves, with publications in Physical Review D and JHEP. His recent work also explores AI-assisted formal reasoning. He has contributed to the development of MerLean and MerLean-Prover, investigating how large language models and Lean can support autoformalization, theorem proving and machine-checked research in mathematics and theoretical physics. He currently serves as a reviewer for Physlib.

 

Alternative online link: https://meeting.tencent.com/dm/5hbAxKUo6DSZ
ID: 427612188