Improving Multi-Agent Cooperation using Theory of Mind

07/30/2020
by   Terence X. Lim, et al.
0

Recent advances in Artificial Intelligence have produced agents that can beat human world champions at games like Go, Starcraft, and Dota2. However, most of these models do not seem to play in a human-like manner: People infer others' intentions from their behaviour, and use these inferences in scheming and strategizing. Here, using a Bayesian Theory of Mind (ToM) approach, we investigated how much an explicit representation of others' intentions improves performance in a cooperative game. We compared the performance of humans playing with optimal-planning agents with and without ToM, in a cooperative game where players have to flexibly cooperate to achieve joint goals. We find that teams with ToM agents significantly outperform non-ToM agents when collaborating with all types of partners: non-ToM, ToM, as well as human players, and that the benefit of ToM increases the more ToM agents there are. These findings have implications for designing better cooperative agents.

READ FULL TEXT

page 3

page 4

page 5

research
09/16/2020

Theory of Mind with Guilt Aversion Facilitates Cooperative Reinforcement Learning

Guilt aversion induces experience of a utility loss in people if they be...
research
05/20/2021

Human-agent coordination in a group formation game

Coordination and cooperation between humans and autonomous agents in coo...
research
03/08/2021

Esports Agents with a Theory of Mind: Towards Better Engagement, Education, and Engineering

The role of AI in esports is shifting from leveraging games as a testbed...
research
02/13/2020

Sequential Cooperative Bayesian Inference

Cooperation is often implicitly assumed when learning from other agents....
research
09/30/2022

Combining Theory of Mind and Abduction for Cooperation under Imperfect Information

In this paper, we formalise and implement an agent model for cooperation...
research
02/01/2019

The Hanabi Challenge: A New Frontier for AI Research

From the early days of computing, games have been important testbeds for...
research
11/24/2022

Melting Pot 2.0

Multi-agent artificial intelligence research promises a path to develop ...

Please sign up or login with your details

Forgot password? Click here to reset