English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

Self-play is a technique for machine learning in multi-agent systems where alearning algorithm learns by interacting with copies of itself. Self-play isuseful for generating large quantities of data for learning, but has thedrawback that the agents the learner will face post-training may havedramatically different behavior than the learner came to expect by interactingwith itself. For the special case of two-player constant-sum games, self-playthat reaches Nash equilibrium is guaranteed to produce strategies that performwell against any post-training opponent; however, no such guarantee exists formultiplayer games. We show that in games that approximately decompose into aset of two-player constant-sum games (called constant-sum polymatrix games)where global $\epsilon$-Nash equilibria are boundedly far from Nash equilibriain each subgame (called subgame stability), any no-external-regret algorithmthat learns by self-play will produce a strategy with bounded vulnerability.For the first time, our results identify a structural property of multiplayergames that enable performance guarantees for the strategies produced by a broadclass of self-play algorithms. We demonstrate our findings through experimentson Leduc poker.

Guarantees for Self-Play in Multiplayer Games via Polymatrix Decomposability