Decoupling Learning Rules from Representations

TL;DR

This paper introduces a method to partially decouple the choice of representations from learning rules in artificial intelligence, aiming to improve flexibility and consistency across different learning paradigms.

Contribution

It proposes a novel approach to decouple learning rules from representations, addressing a subtle coupling issue in neural network training.

Findings

Method applies to unsupervised, reinforcement, and supervised learning

Decoupling reduces unintended effects of representation choices

Enhances flexibility in designing AI systems

Abstract

In the artificial intelligence field, learning often corresponds to changing the parameters of a parameterized function. A learning rule is an algorithm or mathematical expression that specifies precisely how the parameters should be changed. When creating an artificial intelligence system, we must make two decisions: what representation should be used (i.e., what parameterized function should be used) and what learning rule should be used to search through the resulting set of representable functions. Using most learning rules, these two decisions are coupled in a subtle (and often unintentional) way. That is, using the same learning rule with two different representations that can represent the same sets of functions can result in two different outcomes. After arguing that this coupling is undesirable, particularly when using artificial neural networks, we present a method for partially decoupling these two decisions for a broad class of learning rules that span unsupervised learning, reinforcement learning, and supervised learning.

Full text

See pages 1-last of fixed.pdf