Observable Stochastic Game

Partially Observable Stochastic Games (POSGs) model multi-agent decision-making under uncertainty where agents have incomplete information about the environment and each other's actions. Current research focuses on developing efficient algorithms to find optimal or near-optimal strategies, often employing techniques like heuristic search value iteration (HSVI), linear programming, and deep reinforcement learning with neural network architectures (e.g., graph attention networks) to handle complex state spaces and observations. These advancements are crucial for addressing challenges in various domains, including AI safety, robust reinforcement learning, and multi-agent coordination in applications such as autonomous driving and air mobility, where incomplete information and strategic interactions are prevalent.

Papers

September 12, 2024

Games for AI Control: Models of Safety Evaluations of AI Deployment Protocols
Charlie Griffin, Louis Thomson, Buck Shlegeris, Alessandro Abate
Full Model Video Game Red Teaming Game AI Safety Evaluation Observable Stochastic Game AI Deployment Efficient Protocol AI Control

May 29, 2024

$\epsilon$-Optimally Solving Zero-Sum POSGs
Erwan Escudie, Matthia Sabatelli, Jilles Dibangoye
Value Iteration Observable Stochastic Game Optimal Value Function State Occupancy Matching Zero Sum

May 8, 2024

Imprecise Probabilities Meet Partial Observability: Game Semantics for Robust POMDPs
Eline M. Bovy, Marnix Suilen, Sebastian Junges, Nils Jansen
Markov Decision Process C POMDPs Observable Stochastic Game Imprecise Probability

April 16, 2024

HSVI-based Online Minimax Strategies for Partially Observable Stochastic Games with Neural Perception Mechanisms
Rui Yan, Gabriel Santos, Gethin Norman, David Parker, Marta Kwiatkowska
Optimal Strategy Observable Stochastic Game Informed Agent Perception Mechanism

February 13, 2024

Exploration by Optimization with Hybrid Regularizers: Logarithmic Regret with Adversarial Robustness in Partial Monitoring
Taira Tsuchiya, Shinji Ito, Junya Honda
Optimization Purpose Adversarial Robustness Regret Bound Logarithmic Regret Observable Stochastic Game Partial Monitoring

November 30, 2023

Optimal Attack and Defense for Reinforcement Learning
Jeremy McMahan, Young Wu, Xiaojin Zhu, Qiaomin Xie
Reinforcement Learning Stackelberg Game Adversarial Reinforcement Learning Defense Algorithm Observable Stochastic Game Optimal Attack

November 15, 2023

On Convex Optimal Value Functions For POSGs
Rafael F. Cunha, Jacopo Castellini, Johan Peralez, Jilles S. Dibangoye
Reinforcement Learning Markov Game Multi Agent Planning Observable Stochastic Game Optimal Value Function

October 17, 2023

Partially Observable Stochastic Games with Neural Perception Mechanisms
Rui Yan, Gabriel Santos, Gethin Norman, David Parker, Marta Kwiatkowska
Stochastic Game Observable Stochastic Game Multi Agent Sequential Decision Informed Agent Perception Mechanism

August 25, 2023

Collaborative Information Dissemination with Graph-based Multi-Agent Reinforcement Learning
Raffaele Galliera, Kristen Brent Venable, Matteo Bassani, Niranjan Suri
Multi Agent Reinforcement Learning Graph Reinforcement Learning Information Sharing Information Spread Observable Stochastic Game

August 16, 2023

Partially Observable Multi-Agent Reinforcement Learning with Information Sharing
Xiangyu Liu, Kaiqing Zhang
High Efficiency Multi Agent Reinforcement Learning Multi Agent Blessing Generation Information Sharing Observable Stochastic Game

April 25, 2023

Zero-shot Transfer Learning of Driving Policy via Socially Adversarial Traffic Flow
Dongkun Zhang, Jintao Xue, Yuxiang Cui, Yunkai Wang, Eryun Liu, Wei Jing, Junbo Chen, Rong Xiong, Yue Wang
Social Navigation Health Policy Recommendation Observable Stochastic Game Dense Traffic Adversarial Traffic

February 14, 2023

Graph Attention Multi-Agent Fleet Autonomy for Advanced Air Mobility
Malintha Fernando, Ransalu Senanayake, Heeyoul Choi, Martin Swany
Graph Attention Observable Stochastic Game Advanced Air Mobility Future Mobility Autonomous Mobility

October 26, 2022

HSVI can solve zero-sum Partially Observable Stochastic Games
Aurélien Delage, Olivier Buffet, Jilles S. Dibangoye, Abdallah Saffidine
Sequential Decision Making Problem Linear Programming C POMDPs Imperfect Information Game Observable Stochastic Game Zero Sum

October 19, 2022

MPOGames: Efficient Multimodal Partially Observable Dynamic Games
Oswin So, Paul Drews, Thomas Balch, Velin Dimitrov, Guy Rosman, Evangelos A. Theodorou
Nash Equilibrium Dynamic Game Imperfect Information Game Observable Stochastic Game Local Nash Equilibrium

July 29, 2022

Best-of-Both-Worlds Algorithms for Partial Monitoring
Taira Tsuchiya, Shinji Ito, Junya Honda
Regret Bound Optimal Action Adversarial Setting Observable Stochastic Game Best of Both World Algorithm Partial Monitoring

June 24, 2022

Toward multi-target self-organizing pursuit in a partially observable Markov game
Lijun Sun, Yu-Cheng Chang, Chao Lyu, Ye Shi, Yuhui Shi, Chin-Teng Lin
Observable Stochastic Game Coevolutionary Algorithm Pursuit Strategy

June 2, 2022

Sample-Efficient Reinforcement Learning of Partially Observable Markov Games
Qinghua Liu, Csaba Szepesvári, Chi Jin
Multi Agent Reinforcement Learning Partial Observability Sample Efficient Reinforcement Learning Approximate Nash Equilibrium Observable Stochastic Game Information Extensive Form Game

February 22, 2022

Observable Stochastic Game

Papers

Games for AI Control: Models of Safety Evaluations of AI Deployment Protocols

$\epsilon$-Optimally Solving Zero-Sum POSGs

Imprecise Probabilities Meet Partial Observability: Game Semantics for Robust POMDPs

HSVI-based Online Minimax Strategies for Partially Observable Stochastic Games with Neural Perception Mechanisms

Exploration by Optimization with Hybrid Regularizers: Logarithmic Regret with Adversarial Robustness in Partial Monitoring

Optimal Attack and Defense for Reinforcement Learning

On Convex Optimal Value Functions For POSGs

Partially Observable Stochastic Games with Neural Perception Mechanisms

Collaborative Information Dissemination with Graph-based Multi-Agent Reinforcement Learning

Partially Observable Multi-Agent Reinforcement Learning with Information Sharing

Zero-shot Transfer Learning of Driving Policy via Socially Adversarial Traffic Flow

Graph Attention Multi-Agent Fleet Autonomy for Advanced Air Mobility

HSVI can solve zero-sum Partially Observable Stochastic Games

MPOGames: Efficient Multimodal Partially Observable Dynamic Games

Best-of-Both-Worlds Algorithms for Partial Monitoring

Toward multi-target self-organizing pursuit in a partially observable Markov game

Sample-Efficient Reinforcement Learning of Partially Observable Markov Games

SIPOMDPLite-Net: Lightweight, Self-Interested Learning and Planning in POSGs with Sparse Interactions

Minimax Regret for Partial Monitoring: Infinite Outcomes and Rustichini's Regret