Deep Learning without Weight Transport

TL;DR

This paper introduces biologically plausible mechanisms enabling deep learning without weight transport, achieving performance close to standard backpropagation on complex visual tasks.

Contribution

It proposes two novel mechanisms, a weight mirror and a modified learning algorithm, that allow accurate feedback weight learning without weight transport in large neural networks.

Findings

Outperforms feedback alignment and sign-symmetry methods on ImageNet

Nearly matches backpropagation performance on visual recognition tasks

Demonstrates biological plausibility in deep learning algorithms

Abstract

Current algorithms for deep learning probably cannot run in the brain because they rely on weight transport, where forward-path neurons transmit their synaptic weights to a feedback path, in a way that is likely impossible biologically. An algorithm called feedback alignment achieves deep learning without weight transport by using random feedback weights, but it performs poorly on hard visual-recognition tasks. Here we describe two mechanisms - a neural circuit called a weight mirror and a modification of an algorithm proposed by Kolen and Pollack in 1994 - both of which let the feedback path learn appropriate synaptic weights quickly and accurately even in large networks, without weight transport or complex wiring.Tested on the ImageNet visual-recognition task, these mechanisms outperform both feedback alignment and the newer sign-symmetry method, and nearly match backprop, the standard algorithm of deep learning, which uses weight transport.