Satisficing Mentalizing: Bayesian Models of Theory of Mind Reasoning in Scenarios with Different Uncertainties
This addresses the need for efficient mental state inference in artificial systems like intelligent assistants, though it is incremental as it builds on existing Bayesian models.
The paper tackles the problem of efficient Theory of Mind reasoning under computational constraints by proposing a Switching approach that combines specialized Bayesian models, achieving more satisficing mentalizing compared to a Full Bayesian model based on human behavior data.
The ability to interpret the mental state of another agent based on its behavior, also called Theory of Mind (ToM), is crucial for humans in any kind of social interaction. Artificial systems, such as intelligent assistants, would also greatly benefit from such mentalizing capabilities. However, humans and systems alike are bound by limitations in their available computational resources. This raises the need for satisficing mentalizing, reconciling accuracy and efficiency in mental state inference that is good enough for a given situation. In this paper, we present different Bayesian models of ToM reasoning and evaluate them based on actual human behavior data that were generated under different kinds of uncertainties. We propose a Switching approach that combines specialized models, embodying simplifying presumptions, in order to achieve a more statisficing mentalizing compared to a Full Bayesian ToM model.