Active Predictive Coding: A Unified Neural Framework for Learning Hierarchical World Models for Perception and Planning
It addresses foundational issues in AI for perception and planning, offering a unified solution to multiple open problems, though it appears incremental as it builds on existing predictive coding and hierarchical methods.
The paper tackles the problems of learning compositional representations for equivariant vision and solving large-scale planning by composing action sequences, proposing an active predictive coding framework that learns hierarchical world models, and demonstrates viability on vision datasets and a planning problem.
Predictive coding has emerged as a prominent model of how the brain learns through predictions, anticipating the importance accorded to predictive learning in recent AI architectures such as transformers. Here we propose a new framework for predictive coding called active predictive coding which can learn hierarchical world models and solve two radically different open problems in AI: (1) how do we learn compositional representations, e.g., part-whole hierarchies, for equivariant vision? and (2) how do we solve large-scale planning problems, which are hard for traditional reinforcement learning, by composing complex action sequences from primitive policies? Our approach exploits hypernetworks, self-supervised learning and reinforcement learning to learn hierarchical world models that combine task-invariant state transition networks and task-dependent policy networks at multiple abstraction levels. We demonstrate the viability of our approach on a variety of vision datasets (MNIST, FashionMNIST, Omniglot) as well as on a scalable hierarchical planning problem. Our results represent, to our knowledge, the first demonstration of a unified solution to the part-whole learning problem posed by Hinton, the nested reference frames problem posed by Hawkins, and the integrated state-action hierarchy learning problem in reinforcement learning.