Active Predictive Coding Networks: A Neural Solution to the Problem of Learning Reference Frames and Part-Whole Hierarchies

Abstract

We introduce Active Predictive Coding Networks (APCNs), a new class of neural networks that solve a major problem posed by Hinton and others in the fields of artificial intelligence and brain modeling: how can neural networks learn intrinsic reference frames for objects and parse visual scenes into part-whole hierarchies by dynamically allocating nodes in a parse tree? APCNs address this problem by using a novel combination of ideas: (1) hypernetworks are used for dynamically generating recurrent neural networks that predict parts and their locations within intrinsic reference frames conditioned on higher objectlevel embedding vectors, and (2) reinforcement learning is used in conjunction with backpropagation for end-to-end learning of model parameters. The APCN architecture lends itself naturally to multilevel hierarchical learning and is closely related to predictive coding models of cortical function. Using the MNIST, Fashion-MNIST and Omniglot datasets, we demonstrate that APCNs can (a) learn to parse images into part-whole hierarchies, (b) learn compositional representations, and (c) transfer their knowledge to unseen classes of objects. With their ability to dynamically generate parse trees with part locations for objects, APCNs offer a new framework for explainable AI that leverages advances in deep learning while retaining interpretability and compositionality.

Dynamic Predictive Coding: A Model of Hierarchical Sequence Learning and Prediction in the Neocortex

“…The DPC model can be augmented to incorporate actions, e.g., by augmenting the higher-level state with a higher-level action input to predict the dynamics of lower-level sensory consequences of the action. Such a model also opens the door to using probabilistic inference for hierarchical planning [18,[51][52][53], where actions are selected to minimize the sensory prediction errors with respect to preferred goal states (see also [54]). Such a model could also potentially be used to understand the neural basis of navigation and planning [10,[55][56][57] as an emergent property of prediction error minimization.…”

A Sensory-Motor Theory of the Neocortex based on Active Predictive Coding

“…The lower-level state neurons (at level i) maintain an estimate of the lower-level state s i t,τ where τ denotes a time step at the lower level within the higher level time interval given by t. In the two-level network, this lower-level state makes a prediction of the input via a "decoder" network D. In the simplest case where D is a linear matrix U , this lowest level of APC is equivalent to the generative model using in sparse coding (I = U s where s is sparse [49]). More generally, D can be a 1-layer RELU network [48] or a multi-layer decoder [11,12].…”

“…In the simplest case where D is a linear matrix U , this lowest level of APC is equivalent to the generative model using in sparse coding ( I = Us where s is sparse [49]). More generally, D can be a 1-layer RELU network [48] or a multi-layer decoder [11, 12].…”

“…The APC model is well-suited to the sensory-motor nature of human vision, given its integrated state and action networks. We briefly review here our recent simulation study of the APC model applied to the active vision and part-whole learning problem [10, 11].…”

“…Feedback from higher cortical areas modulates the dynamics of both networks, leading to representations that operate at multiple levels of sensory and motor abstraction, as observed in cortical hierarchies implicated in perception and action [8,9]. The author's team has previously presented the computational elements of APC to AI audiences [10][11][12]. Here we explore the application of APC to understanding cortical function.…”