Intuitive physics learning in a deep-learning model inspired by developmental psychology

Piloto et al. introduce a deep-learning system which is able to learn basic rules of the physical world, such as object solidity and persistence. 'Intuitive physics' enables our pragmatic engagement with the physical world and forms a key component of 'common sense' aspects of thought. Current artificial intelligence systems pale in their understanding of intuitive physics, in comparison to even very young children. Here we address this gap between humans and machines by drawing on the field of developmental psychology. First, we introduce and open-source a machine-learning dataset designed to evaluate conceptual understanding of intuitive physics, adopting the violation-of-expectation (VoE) paradigm from developmental psychology. Second, we build a deep-learning system that learns intuitive physics directly from visual data, inspired by studies of visual cognition in children. We demonstrate that our model can learn a diverse set of physical concepts, which depends critically on object-level representations, consistent with findings from developmental psychology. We consider the implications of these results both for AI and for research on human cognition.

Automated summary

The authors introduce a deep-learning system that can learn the basic rules of the physical world, addressing the gap between current AI systems and human understanding of intuitive physics. They develop a machine-learning dataset to evaluate conceptual understanding and build a deep-learning system inspired by studies of visual cognition in children. Their model successfully learns a diverse set of physical concepts, consistent with findings from developmental psychology, with implications for AI and human cognition research.