Relational inductive bias for physical construction in humans and machines

Abstract

While current deep learning systems excel at tasks such as object classification, language processing, and gameplay, few can construct or modify a complex system such as a tower of blocks. We hypothesize that what these systems lack is a "relational inductive bias": a capacity for reasoning about inter-object relations and making choices over a structured description of a scene. To test this hypothesis, we focus on a task that involves gluing pairs of blocks together to stabilize a tower, and quantify how well humans perform. We then introduce a deep reinforcement learning agent which uses object- and relation-centric scene and policy representations and apply it to the task. Our results show that these structured representations allow the agent to outperform both humans and more naive approaches, suggesting that relational inductive bias is an important component in solving structured reasoning problems and for building more intelligent, flexible machines.

arXiv commentIn Proceedings of the Annual Meeting of the Cognitive Science Society (CogSci 2018)

Cited by in corpus (11)

• Graph Neural Networks: A Review of Methods and Applications
• Hamiltonian Neural Networks
• Rapid trial-and-error learning with simulation supports flexible tool use and physical reasoning
• GAG: Global Attributed Graph Neural Network for Streaming Session-based Recommendation
• On the Binding Problem in Artificial Neural Networks
• Graph Element Networks: adaptive, structured computation and memory
• Brick-by-Brick: Combinatorial Construction with Deep Reinforcement Learning
• Modeling Attention Flow on Graphs
• Feed-Forward Neural Networks Need Inductive Bias to Learn Equality Relations
• Factors for the Generalisation of Identity Relations by Neural Networks
• Structured agents for physical construction