How Auto-Encoders Could Provide Credit Assignment in Deep Networks via Target Propagation

TL;DR

This paper introduces a method using auto-encoders for target propagation to improve credit assignment in deep networks, especially with non-linearities and discrete units, reducing reliance on derivatives.

Contribution

It proposes a novel layer-local training signal via reconstruction, generalizing gradients for deep networks with discrete units, and offers theoretical motivations and conjectures for biological plausibility.

Findings

Reconstruction acts as a local training signal for deep layers.

Target propagation can handle discrete and non-linear transformations.

Auto-encoders improve representation learning by flattening manifolds.

Abstract

We propose to exploit {\em reconstruction} as a layer-local training signal for deep learning. Reconstructions can be propagated in a form of target propagation playing a role similar to back-propagation but helping to reduce the reliance on derivatives in order to perform credit assignment across many levels of possibly strong non-linearities (which is difficult for back-propagation). A regularized auto-encoder tends produce a reconstruction that is a more likely version of its input, i.e., a small move in the direction of higher likelihood. By generalizing gradients, target propagation may also allow to train deep networks with discrete hidden units. If the auto-encoder takes both a representation of input and target (or of any side information) in input, then its reconstruction of input representation provides a target towards a representation that is more likely, conditioned on all the side information. A deep auto-encoder decoding path generalizes gradient propagation in a learned way that can could thus handle not just infinitesimal changes but larger, discrete changes, hopefully allowing credit assignment through a long chain of non-linear operations. In addition to each layer being a good auto-encoder, the encoder also learns to please the upper layers by transforming the data into a space where it is easier to model by them, flattening manifolds and disentangling factors. The motivations and theoretical justifications for this approach are laid down in this paper, along with conjectures that will have to be verified either mathematically or experimentally, including a hypothesis stating that such auto-encoder mediated target propagation could play in brains the role of credit assignment through many non-linear, noisy and discrete transformations.