Reinforcement Learning On Human Decision Models For Uniquely Collaborative AI Teammates

Abstract

In 2021 the Johns Hopkins University Applied Physics Laboratory held an internal challenge to develop artificially intelligent (AI) agents that could excel at the collaborative card game Hanabi. Agents were evaluated on their ability to play with human players whom the agents had never previously encountered. This study details the development of the agent that won the challenge by achieving a human-play average score of 16.5, outperforming the current state-of-the-art for human-bot Hanabi scores. The winning agent's development consisted of observing and accurately modeling the author's decision making in Hanabi, then training with a behavioral clone of the author. Notably, the agent discovered a human-complementary play style by first mimicking human decision making, then exploring variations to the human-like strategy that led to higher simulated human-bot scores. This work examines in detail the design and implementation of this human compatible Hanabi teammate, as well as the exis

Stats

• citations0
• S2 citations—
• github stars0
• HF likes0
• heat score0.00
• arxiv keykantack2021reinforcement

Related papers

• Evaluation Of Human-ai Teams For Learned And Rule-based Agents In Hanabi (2021)0.00
• In Pursuit Of Predictive Models Of Human Preferences Toward AI Teammates (2025)0.00
• Collaborating With Humans Without Human Data (2021)0.00
• Simplified Action Decoder For Deep Multi-agent Reinforcement Learning (2019)4.03
• Human-ai Coordination Via Human-regularized Search And Learning (2022)0.00
• Enhancing Human Experience In Human-agent Collaboration: A Human-centered Modeling Approach Based On Positive Human Gain (2024)0.00
• Theory Of Mind For Deep Reinforcement Learning In Hanabi (2021)0.00
• Evaluating The Rainbow DQN Agent In Hanabi With Unseen Partners (2020)0.00