Implicitly Aligning Humans and Autonomous Agents through Shared Task Abstractions

TL;DR

This paper introduces HA$^2$, a hierarchical reinforcement learning framework that enhances zero-shot coordination between autonomous agents and humans by mimicking human-like shared task abstractions, leading to improved adaptability and performance.

Contribution

The paper proposes HA$^2$, a novel hierarchical RL approach that models shared task abstractions to improve autonomous agents' zero-shot coordination with humans and unseen agents.

Findings

HA$^2$ outperforms existing methods in Overcooked environment

HA$^2$ shows improved resilience to environmental shifts

HA$^2$ achieves statistically significant gains in coordination metrics

Abstract

In collaborative tasks, autonomous agents fall short of humans in their capability to quickly adapt to new and unfamiliar teammates. We posit that a limiting factor for zero-shot coordination is the lack of shared task abstractions, a mechanism humans rely on to implicitly align with teammates. To address this gap, we introduce HA$^2$: Hierarchical Ad Hoc Agents, a framework leveraging hierarchical reinforcement learning to mimic the structured approach humans use in collaboration. We evaluate HA$^2$ in the Overcooked environment, demonstrating statistically significant improvement over existing baselines when paired with both unseen agents and humans, providing better resilience to environmental shifts, and outperforming all state-of-the-art methods.