Agent Probing Interaction Policies

2019/11/21 by Siddharth Ghiya, Ghiya, Siddharth, Oluwafemi Azeez +3
Computer Science · #Artificial Intelligence (cs.AI) #Artificial Intelligence in Games #Data Stream Mining Techniques #FOS: Computer and information sciences #Machine Learning (cs.LG) #Multiagent Systems (cs.MA) #Reinforcement Learning in Robotics

paper · pdf · doi:10.48550/arxiv.1911.09535

openalex publication_date 2019/11/21 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28

Abstract

Reinforcement learning in a multi agent system is difficult because these systems are inherently non-stationary in nature. In such a case, identifying the type of the opposite agent is crucial and can help us address this non-stationary environment. We have investigated if we can employ some probing policies which help us better identify the type of the other agent in the environment. We've made a simplifying assumption that the other agent has a stationary policy that our probing policy is trying to approximate. Our work extends Environmental Probing Interaction Policy framework to handle multi agent environments.

Citations

• Long Short-Term Memory
• Learning Multiagent Communication with Backpropagation
• Learning to reinforcement learn
• Model-Agnostic Meta-Learning for Fast Adaptation of Deep Networks
• Multi-Agent Actor-Critic for Mixed Cooperative-Competitive Environments
• Continuous Adaptation via Meta-Learning in Nonstationary and Competitive\n Environments
• Probabilistic Recursive Reasoning for Multi-Agent Reinforcement Learning

Related

• Environment Probing Interaction Policies
• Learning to Model Opponent Learning
• Metric Policy Representations for Opponent Modeling
• Learning through Probing: a decentralized reinforcement learning architecture for social dilemmas
• Quantifying Agent Interaction in Multi-agent Reinforcement Learning for Cost-efficient Generalization