Unsupervised Learning through Prediction in a Model of Cortex

TL;DR

This paper introduces PJOIN, a primitive for cortical computation that enables complex learning and pattern matching tasks, aligning with observed brain phenomena like memory prediction and hierarchical information flow.

Contribution

It extends Valiant's model with PJOIN, demonstrating its implementation and capability to perform cognitively relevant learning and prediction tasks.

Findings

PJOIN can be implemented within Valiant's cortical model.

PJOIN enables complex pattern matching and learning tasks.

The approach models brain phenomena such as memory prediction and hierarchical traffic.

Abstract

We propose a primitive called PJOIN, for "predictive join," which combines and extends the operations JOIN and LINK, which Valiant proposed as the basis of a computational theory of cortex. We show that PJOIN can be implemented in Valiant's model. We also show that, using PJOIN, certain reasonably complex learning and pattern matching tasks can be performed, in a way that involves phenomena which have been observed in cognition and the brain, namely memory-based prediction and downward traffic in the cortical hierarchy.