---
title: "Unsupervised Learning by Backward Inhibition"
authors:
  - "Tomas Hrycej"
doi: ""
url: "https://ijcai.org/proceedings/1989/27"
pdf: "https://ijcai.org/proceedings/1989/0027.pdf"
venue: "Proceedings of the Eleventh International Joint Conference on Artificial Intelligence"
year: 1989
pages: "170-174"
track: "Parallel and Distributed Processing"
keywords:
---

# Unsupervised Learning by Backward Inhibition

Tomas Hrycej

*Proceedings of the Eleventh International Joint Conference on Artificial Intelligence*, pages 170-174.

## Abstract

Backward inhibition in a two-layer connectionist network can be used as an alternative to, or an enhancement of, the competitive model for unsupervised learning. Two feature discovery algorithms based on backward inhibition are presented. It is shown that they are superior to the competitive feature discovery algorithm in feature independence and controllable grain. Moreover, the representation in the feature layer is distributed, and a certain "classification hierarchy" is defined by the features discovered.