Discovering mathematical concepts through a multi-agent system
Daattavya Aggarwal, Oisin Kim, Carl Henrik Ek, Challenger Mishra · 2026
Read Discovering mathematical concepts through a multi-agent system by Daattavya Aggarwal, Oisin Kim, Carl Henrik Ek, Challenger Mishra (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2603-04528.txt.
Description
Aggarwal and colleagues build a multi-agent system that finds math concepts. Concepts emerge from tests, proof attempts, and counter cases, and their system poses conjectures, tries to prove them, and adapts from feedback and shifting data. Inspired by Euler's polyhedron conjecture and an open literature challenge, they benchmark autonomous recovery of homology from polyhedral data plus linear algebra knowledge. The system completes that learning problem. Ablation tests test the full dynamic against controls, supporting the claim that combining local processes yields well aligned notions of math interestingness.
Catalog record
| CREATOR | Daattavya Aggarwal, Oisin Kim, Carl Henrik Ek, Challenger Mishra |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 66035 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.04528