Lazy Agent

The "lazy agent" problem in multi-agent reinforcement learning (MARL) describes scenarios where some agents within a cooperative system fail to contribute adequately to the shared goal, relying on teammates to carry the workload. Current research focuses on addressing this issue through improved credit assignment mechanisms, often leveraging causal inference techniques to better understand each agent's contribution to the overall reward. This work aims to enhance cooperation and efficiency in MARL systems, with applications ranging from distributed robotics and sensor networks to more general collaborative AI systems. Promising approaches include incorporating hypergraph convolutions to model agent interdependencies and employing novel aggregation strategies in federated learning to mitigate the impact of lazy clients.

Papers

March 26, 2024

Paths to Equilibrium in Games
Bora Yongacoglu, Gürdal Arslan, Lacra Pavel, Serdar Yüksel
Reinforcement Learning Multi Agent Reinforcement Learning Root to Leaf Path Equilibrium State Normal Form Game Lazy Agent

January 22, 2024

TurboSVM-FL: Boosting Federated Learning through SVM Aggregation for Lazy Clients
Mengdi Wang, Anna Bodonhelyi, Efe Bozkir, Enkelejda Kasneci
Collaborative Machine Learning Federated Image Classification Training Iteration Lazy Agent

November 5, 2023

Learning Independently from Causality in Multi-Agent Environments
Rafael Pina, Varuna De Silva, Corentin Artaud
LeArning Abstract Multi Agent Reinforcement Learning Causal Pattern Multi Agent Environment Individual Agent Lazy Agent Causality Based Perspective

June 20, 2023

Discovering Causality for Efficient Cooperation in Multi-Agent Environments
Rafael Pina, Varuna De Silva, Corentin Artaud
Causal Discovery Causal Effect Estimation Cooperative Multi Agent Reinforcement Learning Multi Agent Environment Event Causality Identification Causality Detection Lazy Agent

March 24, 2023

Causality Detection for Efficient Multi-Agent Reinforcement Learning
Rafael Pina, Varuna De Silva, Corentin Artaud
Multi Agent Reinforcement Learning Causal Effect Estimation Sub Optimal Policy Causality Detection Temporal Causal Lazy Agent

February 17, 2023

Loss of Distributed Coverage Using Lazy Agents Operating Under Discrete, Local, Event-Triggered Communication
Edward Vickery, Aditya A. Paranjape
Non Humanoid Robot Multi Agent System Interest Loss Discrete Environment Robot Communication Coverage Control Event Triggered Lazy Agent

December 9, 2021

Cooperative Multi-Agent Reinforcement Learning with Hypergraph Convolution
Yunpeng Bai, Chen Gong, Bin Zhang, Guoliang Fan, Xinwen Hou, Yu Liu
Multi Agent System Agent System Cooperative Multi Agent Reinforcement Learning Agent Benchmark Lazy Agent Hypergraph Convolution