Emergence: Overcoming Privileged Information Bias in Asymmetric Embodied Agents via Active Querying

TL;DR

This paper addresses the challenge of asymmetric information in embodied AI agents, introducing an active querying method to improve collaboration success rates by reducing grounding errors.

Contribution

It proposes a novel active querying framework within an asymmetric reasoning setting, demonstrating its effectiveness over standard instruction protocols in embodied environments.

Findings

Leader perceives target in 35% of episodes

Team succeeds in 17% of episodes, with 50% failure due to grounding errors

Active querying doubles clarification requests compared to push-based methods

Abstract

Large Language Models (LLMs) act as powerful reasoning engines but struggle with "symbol grounding" in embodied environments, particularly when information is asymmetrically distributed. We investigate the Privileged Information Bias (or "Curse of Knowledge"), where a knowledgeable "Leader" agent fails to guide a sensor-limited "Follower" due to a lack of Theory of Mind. To quantify this phenomenon, we propose a novel Asymmetric Assistive Reasoning framework within AI2-THOR. Our experiments reveal a significant "Success Gap": while the Leader successfully perceives the target in 35.0% of episodes, the collaborative team succeeds only 17.0% of the time, implying that nearly 50% of feasible plans fail solely due to communicative grounding errors. We demonstrate that a "Pull-based" protocol (active querying) is significantly more robust than standard "Push-based" instruction, with successful episodes featuring 2x the frequency of clarification requests. This research isolates the mechanism of active uncertainty reduction as a prerequisite for safe human-AI and robot-robot collaboration.