Feedback alignment in deep convolutional networks

TL;DR

This paper explores biologically plausible learning algorithms for deep convolutional networks, showing that a modified feedback alignment method with sign agreement can perform comparably to backpropagation.

Contribution

It introduces techniques to improve the scalability of feedback alignment in deep networks and demonstrates that sign-based weight symmetry suffices for effective learning.

Findings

Modified feedback alignment achieves competitive performance with backpropagation.

Sign agreement in weights is sufficient for effective learning in deep networks.

Mechanisms promoting weight alignment are crucial for biologically plausible learning.

Abstract

Ongoing studies have identified similarities between neural representations in biological networks and in deep artificial neural networks. This has led to renewed interest in developing analogies between the backpropagation learning algorithm used to train artificial networks and the synaptic plasticity rules operative in the brain. These efforts are challenged by biologically implausible features of backpropagation, one of which is a reliance on symmetric forward and backward synaptic weights. A number of methods have been proposed that do not rely on weight symmetry but, thus far, these have failed to scale to deep convolutional networks and complex data. We identify principal obstacles to the scalability of such algorithms and introduce several techniques to mitigate them. We demonstrate that a modification of the feedback alignment method that enforces a weaker form of weight symmetry, one that requires agreement of weight sign but not magnitude, can achieve performance competitive with backpropagation. Our results complement those of Bartunov et al. (2018) and Xiao et al. (2018b) and suggest that mechanisms that promote alignment of feedforward and feedback weights are critical for learning in deep networks.