Reinforcing the World’s Edge: A Continual Learning Problem in the Multi-Agent-World Boundary
Dane Malenfant · 2026
Read Reinforcing the World’s Edge: A Continual Learning Problem in the Multi-Agent-World Boundary by Dane Malenfant (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-2603-06813.txt.
Description
Malenfant connects decentralized multi-agent reinforcement learning to ongoing learning: peer updates change each focal agent's rewards and dynamics even when the joint game stays fixed. A stable core captures abstract patterns shared by many successful runs, while survival tracks a pattern's coverage, not an unchanged frontier. A conditioning limit caps coverage loss in terms of run-law drift and reference success share. Combined with peer-policy movement, that limit proves survival under bounded peer updates with positive coverage margin. An effective-conflict rule yields a first-exit law.
Catalog record
| CREATOR | Dane Malenfant |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 18712 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2603.06813