AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

K-level Reasoning for Zero-Shot Coordination in Hanabi

Brandon Cui, Hengyuan Hu, Luis Pineda, Jakob Nicolaus Foerster

NeurIPS 2021 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Human-AI coordination agents · 8 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
train KLR hierarchy levels synchronously rather than sequentially (ablation)KLR · Hanabi cross-playscore: 17.17\pm 0.98 -> 22.14\pm 0.12Improved
train KLR hierarchy levels synchronously rather than sequentially (ablation)KLR · Hanabi with Color Botscore: 14.80\pm 1.77 -> 17.21 \pm 0.86Improved
train KLR hierarchy levels synchronously rather than sequentially (ablation)KLR · Hanabi with Rank Botscore: 12.36\pm1.44 -> 15.90\pm0.30Improved
train KLR hierarchy levels synchronously rather than sequentially (ablation)KLR · Hanabi with Clone Botscore: 13.03 \pm 1.91 -> 15.86 \pm 0.23Improved
co-train a best response to the KLR hierarchy (ablation)Synchronous KLR · Hanabi cross-playscore: 22.14\pm 0.12 -> 23.29 \pm 0.05Improved
co-train a best response to the KLR hierarchy (ablation)Synchronous KLR · Hanabi with Rank Botscore: 15.90\pm0.30 -> 17.01 \pm 0.45Improved
co-train a best response to the KLR hierarchy (ablation)Synchronous KLR · Hanabi with Clone Botscore: 15.86 \pm 0.23 -> 16.59\pm 0.16Improved
replace KLR's preceding-level partner with a Poisson mixture of lower levels (ablation)Synchronous KLR · Hanabi cross-playscore: 22.14\pm 0.12 -> 21.42\pm\textbf{0.10}Hurt
replace KLR's preceding-level partner with a Poisson mixture of lower levels (ablation)Synchronous KLR · Hanabi with Clone Botscore: 15.86 \pm 0.23 -> 14.59 \pm 0.25Hurt
evaluate against independently trained level-4 partners rather than the training partnerSequential KLR (level 5) · Hanabi level-4 partner cross-playpercentage of games bombing out: 0.4 \pm 0.2 -> 19.8 \pm 18.0Hurt
evaluate against independently trained level-4 partners rather than the training partnerSynchronous KLR (level 5) · Hanabi level-4 partner cross-playpercentage of games bombing out: 0.4 \pm 0.2 -> 0.6 \pm 0.3No clear effect

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