AI Paper Atlas

Research map / Multi-agent systems and games / Games and equilibria

Conservative Offline Policy Adaptation in Multi-Agent Games

Chengjie Wu, Pingzhong Tang, Jun Yang, Yujing Hu, Tangjie Lv, Changjie Fan, Chongjie Zhang

NeurIPS 2023 · Multi-agent systems and games › Games and equilibria · approach: Learning in Markov games · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add conservative dataset-constrained self-play instead of BC-First proxyBC-First · Maze case 1test reward: -4 -> 1Improved
remove exploitation (Self-Play, i.e. CSP without dataset) (ablation)CSP · Maze case 2test reward: 16 -> 1Hurt
use CSP instead of pure dataset-followingDataset-Only · Maze case 1test reward: -64 -> 1Improved
use CSP instead of BC-FirstBC-First · MPE predator-prey, adversaries, 10Repisode reward: 21.3 -> 34.6Improved
increase dataset quality 10R to 100SBC-First · MPE good agentepisode reward: -29.8 -> -18.5Improved
increase dataset quality 10R to 100SCSP · MPE good agentepisode reward: -16.7 -> -17.1No clear effect
use MAIL imitation instead of CSPCSP · MPE adversaries, 10Repisode reward: 34.6 -> 4.3Hurt
use CSP instead of BC-FirstBC-First · MuJoCo YouShallNotPassHumans target 2winning rate: 0.46 -> 0.65Improved
use CSP instead of Self-PlaySelf-Play · MuJoCo YouShallNotPassHumans target 3episode reward: 318 -> 1203Improved
use CSP instead of BC-FirstBC-First · Google Football 3vs1 defender target 1winning rate: 0.64 -> 0.9Improved
use CSP instead of BC-FirstBC-First · Google Football 3vs1 attacker target 1winning rate: 0.83 -> 0.81No clear effect
use BC-First (non-conservative) instead of Self-Play on MPE adversaries 100SSelf-Play · MPE adversaries, 100Sepisode reward: 29.7 -> 27.8Hurt

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy