Abstract
In this paper, we formalise and implement an agent model for cooperation under imperfect information. It is based on Theory of Mind (the cognitive ability to understand the mental state of others) and abductive reasoning (the inference paradigm that computes explanations from observations). The combination of these two techniques allows agents to derive the motives behind the actions of their peers, and incorporate this knowledge into their own decision-making. We have implemented this model in a totally domain-independent fashion and successfully tested it for the cooperative card game Hanabi.
Author supplied keywords
Cite
CITATION STYLE
Montes, N., Osman, N., & Sierra, C. (2022). Combining Theory of Mind and Abduction for Cooperation Under Imperfect Information. In Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics) (Vol. 13442 LNAI, pp. 294–311). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-20614-6_17
Register to see more suggestions
Mendeley helps you to discover research relevant for your work.
