Editing Cooperative Ai (section)

== <span style="color: #FFFFFF;">Remembering</span> ==
* '''Cooperation''' — Working jointly with others toward shared goals; distinct from coordination (working without conflict) and collaboration (active joint effort).
* '''Common knowledge''' — Information that everyone knows, everyone knows everyone knows, and so on; essential for coordination.
* '''Schelling point''' — A solution that people tend to converge on without communication, based on salience or convention.
* '''Social dilemma''' — A situation where individual rationality leads to collective suboptimality (e.g., Prisoner's Dilemma).
* '''Prisoner's Dilemma''' — A game where mutual defection is individually rational but mutual cooperation is collectively better.
* '''Folk theorem''' — In repeated games, cooperation can be sustained by strategies that punish defection (tit-for-tat).
* '''Mechanism design''' — Designing rules and incentives to achieve desired collective outcomes among self-interested agents.
* '''Hanabi (cooperative game)''' — A cooperative card game requiring communication and reasoning about partner beliefs; used as an AI benchmark.
* '''Other-play''' — A training method enabling agents to cooperate with partners they've never trained with.
* '''Ad hoc teamwork''' — The problem of an AI agent cooperating with unknown human or AI partners without prior coordination.
* '''Corrigibility''' — The property of an AI system that allows humans to correct, retrain, or shut it down cooperatively.
* '''Value alignment (cooperative)''' — Ensuring AI systems share or at least understand human values to enable genuine cooperation.
* '''Contract theory''' — Economic framework for designing agreements between parties with different information and interests.
* '''Team reward''' — In cooperative MARL, all agents share a single reward signal; maximizing it requires true cooperation.
</div>

<div style="background-color: #006400; color: #FFFFFF; padding: 20px; border-radius: 8px; margin-bottom: 15px;">