An AI-Assisted Formalization of the Poincaré Conjecture

📅 2026-10-06
📈 Citations: 0
✨ Influential: 0
📄 PDF
🤖 AI Summary
This study addresses the lack of formalization infrastructure and low collaborative efficiency in the geometric analysis formalization of the Poincaré conjecture by proposing a milestone-decomposed parallel agent workflow. Leveraging Lean 4 as the verification framework, this approach integrates mathematical blueprint planning with multi-agent parallel reasoning to enable efficient human-AI collaborative formalization of complex mathematical proofs. The project successfully completes the AI-assisted formal verification of the Poincaré conjecture and establishes a reusable formalization infrastructure that significantly reduces the verification costs for subsequent related theorems. Ultimately, this work provides a scalable new paradigm for large-scale mathematical formalization.
📝 Abstract
We present an AI-assisted Lean 4 formalization of the Poincaré conjecture. The project began with limited reusable formal infrastructure for the geometric analysis behind the proof. To organize this work, we combined a proof blueprint prepared by mathematicians with explicit milestone statements. These milestones enabled parallel agent work and gave mathematicians clear points to locate blockers and provide effective mathematical guidance. Our analysis identifies the human interventions and organizational choices behind this workflow. The project provides a starting point toward reusable infrastructure for future formalization projects; such infrastructure, once developed, could eventually reduce the cost of verifying mathematical results in geometric analysis.
Problem

Research questions and friction points this paper is trying to address.

Poincaré conjecture
formalization
geometric analysis
Lean 4
reusable infrastructure
Innovation

Methods, ideas, or system contributions that make the work stand out.

AI-assisted formalization
Lean 4
Poincaré conjecture
proof blueprint
parallel agents
Z
Zhiyuan Zhang
School of Mathematical Sciences, Peking University
A
Axel Delaval
Beijing International Center for Mathematical Research, Peking University
L
Leheng Chen
Beijing International Center for Mathematical Research, Peking University
J
Jinxuan Chen
Beijing International Center for Mathematical Research, Peking University
Jie Xu
Jie Xu
Professor of Systems Engineering and Operations Research, George Mason University
digital twin simulationAI/MLtransportation electrificationhealthcare analyticsdata centers
Y
Yuxuan Liao
School of Mathematical Sciences, Beijing Normal University
J
Jiedong Jiang
School of Mathematical Sciences, Peking University
Chunlei Liu
Chunlei Liu
School of Mathematical Sciences, Capital Normal University
B
Bin Dong
Beijing International Center for Mathematical Research and the New Cornerstone Science Laboratory, Peking University