Decoupling Learning Rules from Representations

by   Philip S. Thomas, et al.

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.


page 1

page 2

page 3

page 4


Proceedings of the Tenth Conference on Uncertainty in Artificial Intelligence (1994)

This is the Proceedings of the Tenth Conference on Uncertainty in Artifi...

How to Amend a Constitution? Model, Axioms, and Supermajority Rules

A self-governed community or society must have rules by which group deci...

Hacia los Comités de Ética en Inteligencia Artificial

The goal of Artificial Intelligence based systems is to take decisions t...

Manipulation-Proof Machine Learning

An increasing number of decisions are guided by machine learning algorit...

Modus ponens and modus tollens for the compositional rule of inference with aggregation functions

The compositional rule of inference (CRI) proposed by Zadeh has been wid...

Comments on Sejnowski's "The unreasonable effectiveness of deep learning in artificial intelligence" [arXiv:2002.04806]

Terry Sejnowski's 2020 paper [arXiv:2002.04806] is entitled "The unreaso...

Backprop as Functor: A compositional perspective on supervised learning

A supervised learning algorithm searches over a set of functions A → B p...