Learning "Partner-Aware" Collaborators in Multi-Party Collaboration
This addresses the challenge of effective collaboration between AI agents and humans in multi-turn settings, though it is incremental as it builds on existing alignment and interruptability literature.
The paper tackles the problem of LLM agents ignoring interventions in multi-party collaboration, proposing a partner-aware learning algorithm that improves common-ground convergence and solution diversity in collaborative tasks.
Large Language Models (LLMs) are increasingly bring deployed in agentic settings where they act as collaborators with humans. Therefore, it is increasingly important to be able to evaluate their abilities to collaborate effectively in multi-turn, multi-party tasks. In this paper, we build on the AI alignment and safe interruptability literature to offer novel theoretical insights on collaborative behavior between LLM-driven collaborator agents and an intervention agent. Our goal is to learn an ideal partner-aware collaborator that increases the group's common-ground (CG)-alignment on task-relevant propositions-by intelligently collecting information provided in interventions by a partner agent.We show how LLM agents trained using standard RLHF and related approaches are naturally inclined to ignore possibly well-meaning interventions, which makes increasing group common ground non-trivial in this setting. We employ a two-player Modified-Action MDP to examine this suboptimal behavior of standard AI agents, and propose Interruptible Collaborative Roleplayer (ICR)-a novel partner-aware learning algorithm to train CG-optimal collaborators. Experiments on multiple collaborative task environments show that ICR, on average, is more capable of promoting successful CG convergence and exploring more diverse solutions in such tasks.