Backpropagation at the Infinitesimal Inference Limit of Energy-Based Models: Unifying Predictive Coding, Equilibrium Propagation, and Contrastive Hebbian Learning

TL;DR

This paper develops a unified theoretical framework showing how energy-based models can approximate backpropagation, linking various biologically plausible algorithms like predictive coding, equilibrium propagation, and contrastive Hebbian learning.

Contribution

It provides a comprehensive theory connecting energy-based models with backpropagation, unifying existing algorithms and enabling the derivation of new BP-approximating methods.

Findings

Unified theory of EBMs approximating BP

Linking predictive coding, equilibrium propagation, and contrastive Hebbian learning

Derivation of new BP-approximating algorithms

Abstract

How the brain performs credit assignment is a fundamental unsolved problem in neuroscience. Many `biologically plausible' algorithms have been proposed, which compute gradients that approximate those computed by backpropagation (BP), and which operate in ways that more closely satisfy the constraints imposed by neural circuitry. Many such algorithms utilize the framework of energy-based models (EBMs), in which all free variables in the model are optimized to minimize a global energy function. However, in the literature, these algorithms exist in isolation and no unified theory exists linking them together. Here, we provide a comprehensive theory of the conditions under which EBMs can approximate BP, which lets us unify many of the BP approximation results in the literature (namely, predictive coding, equilibrium propagation, and contrastive Hebbian learning) and demonstrate that their approximation to BP arises from a simple and general mathematical property of EBMs at free-phase equilibrium. This property can then be exploited in different ways with different energy functions, and these specific choices yield a family of BP-approximating algorithms, which both includes the known results in the literature and can be used to derive new ones.