Finding online neural update rules by learning to remember

TL;DR

This paper explores learning online local update rules for neural networks by training neural networks to remember past experiences, providing insights into learning mechanisms and potential applications in meta-learning and memory systems.

Contribution

It introduces a novel approach of learning neural update rules through memory-based objectives, differing from traditional back-propagation methods.

Findings

Analysis reveals how different network types perform under the memory objective.

Short-term back-propagation effectively trains the update rules.

Insights into what constitutes effective learning rules are provided.

Abstract

We investigate learning of the online local update rules for neural activations (bodies) and weights (synapses) from scratch. We represent the states of each weight and activation by small vectors, and parameterize their updates using (meta-) neural networks. Different neuron types are represented by different embedding vectors which allows the same two functions to be used for all neurons. Instead of training directly for the objective using evolution or long term back-propagation, as is commonly done in similar systems, we motivate and study a different objective: That of remembering past snippets of experience. We explain how this objective relates to standard back-propagation training and other forms of learning. We train for this objective using short term back-propagation and analyze the performance as a function of both the different network types and the difficulty of the problem. We find that this analysis gives interesting insights onto what constitutes a learning rule. We also discuss how such system could form a natural substrate for addressing topics such as episodic memories, meta-learning and auxiliary objectives.