Sample-Efficient Policy Space Response Oracles with Joint Experience Best Response Conference: Proc. of the 25th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2026); May 25 – 29, 2026; Paphos, Cyprus
Ariyan Bighashdel, Thiago D. Simão, Frans A. Oliehoek · 2026
Read Sample-Efficient Policy Space Response Oracles with Joint Experience Best Response Conference: Proc. of the 25th International Conference on Autonomous Agents and Multiagent Systems (AAMAS 2026); May 25 – 29, 2026; Paphos, Cyprus by Ariyan Bighashdel, Thiago D. Simão, Frans A. Oliehoek (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-2602-06599.txt.
Description
Ariyan Bighashdel, Thiago D. Simão, and Frans A. Oliehoek work on computer agents that learn strategies by playing against growing sets of rival strategies. Their setting is a framework that slowly expands a pool of strategies and trains best replies to them. Training a separate reply for each agent is costly, so the authors look for a cheaper path.
Their method collects game records once under the current mix of strategies, then reuses that shared data to compute a best reply for every agent at the same time. Because learning from stored data can drift off course, they add three fixes, including a variant that adds extra exploration during data collection. In benchmark tests the explored version gave the best balance of accuracy and cost, while a hybrid version nearly matched the original method for a fraction of the data.
Catalog record
| CREATOR | Ariyan Bighashdel, Thiago D. Simão, Frans A. Oliehoek |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 53 KB |
| VERSION | 1.0 |
| LICENSE | http://creativecommons.org/licenses/by/4.0/ |
| RIGHTS | CC BY |
| SOURCE | arXiv |
Source: https://arxiv.org/abs/2602-06599