Inverse Attention Agents for Multi-Agent Systems

2024-10-29

AI Summary

Multi-agent systems suffer from poor adaptability to dynamic environments: conventional agents exhibit significant performance degradation when encountering unseen teammates or opponents. To address this, we propose agents endowed with an inverse attention mechanism—introducing the first algorithmic formulation of Theory of Mind (ToM) as a learnable attention module. This module explicitly models and infers others’ goals and intentions in real time, dynamically modulating attention weights over teammates and opponents. Our method integrates an end-to-end trainable inverse attention network into continuous-space multi-agent reinforcement learning frameworks, enabling zero-shot generalization across unfamiliar agents in both cooperative and competitive settings. Empirical evaluation on mixed cooperative-competitive tasks demonstrates substantially improved policy robustness. Human-subject experiments further confirm superior human-agent collaboration efficacy and behavioral fidelity compared to baseline models.

Technology Category

• Multiagent Systems: Adversarial Agents
• Machine Learning: Imitation Learning & Inverse Reinforcement Learning
• Cognitive Modeling & Cognitive Systems: Agent Architectures

Application Category

• Semantics and Knowledge: Data modeling to support human-machine intelligence, including LLMs agents, intelligent system behavior, explanations, and user-friendly interactions
• Responsible Web: Machine-in-the-loop, human agency and autonomy
• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems

Abstract

A major challenge for Multi-Agent Systems is enabling agents to adapt dynamically to diverse environments in which opponents and teammates may continually change. Agents trained using conventional methods tend to excel only within the confines of their training cohorts; their performance drops significantly when confronting unfamiliar agents. To address this shortcoming, we introduce Inverse Attention Agents that adopt concepts from the Theory of Mind (ToM) implemented algorithmically using an attention mechanism trained in an end-to-end manner. Crucial to determining the final actions of these agents, the weights in their attention model explicitly represent attention to different goals. We furthermore propose an inverse attention network that deduces the ToM of agents based on observations and prior actions. The network infers the attentional states of other agents, thereby refining the attention weights to adjust the agent's final action. We conduct experiments in a continuous environment, tackling demanding tasks encompassing cooperation, competition, and a blend of both. They demonstrate that the inverse attention network successfully infers the attention of other agents, and that this information improves agent performance. Additional human experiments show that, compared to baseline agent models, our inverse attention agents exhibit superior cooperation with humans and better emulate human behaviors.

Problem

• Enabling agents to adapt dynamically to diverse environments
• Improving performance when confronting unfamiliar agents
• Inferring attention of other agents to refine actions

Innovation

• Inverse Attention Agents using Theory of Mind
• End-to-end trained attention mechanism for adaptability
• Inverse network infers other agents' attention states