Summary
A research team from the machine learning community presents a detailed case study on using an AI research system to improve bounds on the Grothendieck constant KG, which measures the gap between combinatorial optimization problems and their continuous relaxations. The exact value of KG remains unknown; the authors report tightening the best known bounds to 6π/11 ≤ KG ≤ π/(2log(1+√2)) - 10^-4. Notably, these improvements were produced by an AI system whose insights were judged novel by domain experts. The paper discusses lessons learned from AI-assisted mathematics: the system's strengths and weaknesses, and how to structure problems to create ideal conditions for AI to generate breakthrough insights. Authored by Alan Li, Rahul Saha, Anton Xue, Swarat Chaudhuri, Adam Klivans, Pravesh K Kothari, and Raghu Meka, the work is available on arXiv (2608.11195).
Overview
AI agents are increasingly used in mathematics research, but how to use them effectively is often unclear. This paper presents an extensive case study of how AI was used to improve bounds on the Grothendieck constant KG, a quantity that captures the hardness gap between combinatorial problems and their continuous (semidefinite) relaxations.
While the precise value of KG is not known, the authors recently tightened the best known bounds to:
6π/11 ≤ KG ≤ π/(2log(1+√2)) - 10^-4
Crucially, these improvements were achieved by an AI research system capable of arriving at insights that domain experts considered novel.
Key Contributions
- A documented case study of a long-horizon AI research effort on a genuine open problem in mathematics.
- New best-known bounds on the Grothendieck constant KG.
- A detailed discussion of using AI for mathematical research, covering:
- The strengths and weaknesses of AI research systems in this setting.
- Practical experience on creating ideal conditions for AI to produce breakthrough insights.
Links
- arXiv: 2608.11195
Authors: Alan Li, Rahul Saha, Anton Xue, Swarat Chaudhuri, Adam Klivans, Pravesh K Kothari, Raghu Meka
---
*Auto-collected on 2026-08-13*
This page is an English static mirror generated for search and AI citation.
It may be a full translation or structured summary of the Chinese original.
Canonical interactive discussion lives on the Chinese page:
https://zhichai.net/topic/178633403