Intrinsic Physical Concepts Discovery with Object-Centric Predictive Models
Authors:
- Qu TangQTQu Tang
- Zhen LeiZLZhen Lei
- ZhaoXiang ZhangZZZhaoXiang Zhang
Preprint
Abstract
The ability to discover abstract physical concepts and understand how they work in the world through observing lies at the core of human intelligence. The acquisition of this ability is based on compositionally perceiving the environment in terms of objects and relations in an unsupervised manner. Recent approaches learn object-centric representations and capture visually observable concepts of objects, e.g., shape, size, and location. In this paper, we take a step forward and try to discover and represent intrinsic physical concepts such as mass and charge. We introduce the \uppercase{phy}sical \uppercase{c}oncepts \uppercase{i}nference \uppercase{ne}twork (PHYCINE), a system that infers physical concepts in different abstract levels without supervision. The key insights underlining PHYCINE are two-fold, commonsense knowledge emerges with prediction, and physical concepts of different abstract levels should be reasoned in a bottom-up fashion. Empirical evaluation demonstrates that variables inferred by our system work in accordance with the properties of the corresponding physical concepts. We also show that object representations containing the discovered physical concepts variables could help achieve better performance in causal reasoning tasks, i.e., ComPhy.Keywords
All Related Versions
This publication has 3 references indexed in Scilit:
- Compositional Physical Reasoning of Objects and Events From VideosIEEE Transactions on Pattern Analysis and Machine Intelligence, 2025
- Primary Visual Cortex Represents the Difference Between Past and PresentCerebral Cortex, 2013
- Motion-Dependent Representation of Space in Area MT+Neuron, 2013