Machine Theory of Mind

Abstract

Theory of mind (ToM; Premack & Woodruff, 1978) broadly refers to humans' ability to represent the mental states of others, including their desires, beliefs, and intentions. We propose to train a machine to build such models too. We design a Theory of Mind neural network -- a ToMnet -- which uses meta-learning to build models of the agents it encounters, from observations of their behaviour alone. Through this process, it acquires a strong prior model for agents' behaviour, as well as the ability to bootstrap to richer predictions about agents' characteristics and mental states using only a small number of behavioural observations. We apply the ToMnet to agents behaving in simple gridworld environments, showing that it learns to model random, algorithmic, and deep reinforcement learning agents from varied populations, and that it passes classic ToM tasks such as the "Sally-Anne" test (Wimmer & Perner, 1983; Baron-Cohen et al., 1985) of recognising that others can hold false beliefs about the world. We argue that this system -- which autonomously learns how to model other agents in its world -- is an important step forward for developing multi-agent AI systems, for building intermediating technology for machine-human interaction, and for advancing the progress on interpretable AI.

arXiv comment21 pages, 15 figures

References in corpus (8)

• Stochastic Backpropagation and Approximate Inference in Deep Generative Models
• Autonomous Agents Modelling Other Agents: A Comprehensive Survey and Open Problems
• Cooperative Inverse Reinforcement Learning
• Generative Adversarial Imitation Learning
• Deep Variational Information Bottleneck
• One-Shot Imitation Learning
• Meta-Learning and Universality: Deep Representations and Gradient Descent can Approximate any Learning Algorithm
• Building Machines that Learn and Think for Themselves: Commentary on Lake et al., Behavioral and Brain Sciences, 2017

Cited by in corpus (32)

• A Survey and Critique of Multiagent Deep Reinforcement Learning
• The Next Decade in AI: Four Steps Towards Robust Artificial Intelligence
• Levels of explainable artificial intelligence for human-aligned conversational explanations
• Watch-And-Help: A Challenge for Social Perception and Human-AI Collaboration
• On the Critical Role of Conventions in Adaptive Human-AI Collaboration
• On the Feasibility of Learning, Rather than Assuming, Human Biases for Reward Inference
• Learning Latent Representations to Influence Multi-Agent Interaction
• F2A2: Flexible Fully-decentralized Approximate Actor-critic for Cooperative Multi-agent Reinforcement Learning
• M${}^3$RL: Mind-aware Multi-agent Management Reinforcement Learning
• Resonating Minds -- Emergent Collaboration Through Hierarchical Active Inference
• Modeling Theory of Mind in Multi-Agent Games Using Adaptive Feedback Control
• Neural Recursive Belief States in Multi-Agent Reinforcement Learning
• Learning to Communicate Implicitly By Actions
• Causal Analysis of Agent Behavior for AI Safety
• A Regularized Opponent Model with Maximum Entropy Objective
• PsiPhi-Learning: Reinforcement Learning with Demonstrations using Successor Features and Inverse Temporal Difference Learning
• Meta-Learning of Structured Task Distributions in Humans and Machines
• Model-based Reinforcement Learning for Decentralized Multiagent Rendezvous
• Modelling Bounded Rationality in Multi-Agent Interactions by Generalized Recursive Reasoning
• Agent cognition through micro-simulations: Adaptive and tunable intelligence with NetLogo LevelSpace
• Distributed Policy Iteration for Scalable Approximation of Cooperative Multi-Agent Policies
• Motion Reasoning for Goal-Based Imitation Learning
• Interactive AI with a Theory of Mind
• Theory of Mind with Guilt Aversion Facilitates Cooperative Reinforcement Learning
• Amanuensis: The Programmer's Apprentice
• Interactive Agent Modeling by Learning to Probe
• Evaluating Theory of Mind in Question Answering
• Guilty Artificial Minds
• Learning Triadic Belief Dynamics in Nonverbal Communication from Videos
• On the potential for open-endedness in neural networks
• Theory of Machine Networks: A Case Study
• On Memory Mechanism in Multi-Agent Reinforcement Learning