DUOS meta-learns update rules for opponent shaping in simulated games
Source check: Source checked. Not independent scientific confirmation.
Finding: DUOS meta-trains an update rule that learns from an agent’s own experience without access to opponent parameters, policies, rewards or learning rules at deployment. In simulated general-sum games it ranked first in normalized mean score and self-play among tournament entrants; this is limited to the reported games.
Technology: Discovered Update Rule for Opponent Shaping (DUOS), a meta-learned update rule
Publication: Preprint · Test context: retrospective or benchmark. Catalog categories, not an evidence grade.
- AI contribution
- Discovered Update Rule for Opponent Shaping (DUOS), a meta-learned update rule
- Test and result
- The arXiv abstract reports first-place normalized mean score and self-play among tournament entrants; the evaluation is in simulated games.
- Limitation / still unknown
- The tournament and game simulations do not establish behavior with human opponents or real-world strategic settings.
Evidence sources and annotation context
Evidence context (descriptions follow the cited sources; test type is a catalog classification):
Evidence sources: https://arxiv.org/abs/2611.08335
Source checked · editorial details
The source page was checked against this entry. Independent confirmation is not assessed by this label.
Source type: paper. Paper: linked separately.
Review date: 22 Nov 2026. Review scope: Primary result-bearing material for the stated method, result and evaluation, with bibliographic version information..
No follow-up recorded; this is not a claim that no follow-up exists. Follow-up review scope and date are not recorded. No source-supported relationships recorded.
Editorial assessment: Impact 3/5 · Validation 2/5 · Novelty 3/5. Impact and novelty are editorial judgments; validation follows the annotated test type where one exists. Not an evidence grade. Scoring criteria.