---
title: "Concept Learning by Structured Examples - An Algebraic Approach"
authors:
  - "Fritz Wysotzki"
  - "Werner Kolbe"
  - "Joachim Selbig"
doi: ""
url: "https://ijcai.org/proceedings/1981/32"
pdf: "https://ijcai.org/proceedings/1981/0032.pdf"
venue: "Proceedings of the Seventh International Joint Conference on Artificial Intelligence"
year: 1981
pages: "153-157"
track: "Learning"
keywords:
---

# Concept Learning by Structured Examples - An Algebraic Approach

Fritz Wysotzki, Werner Kolbe, Joachim Selbig

*Proceedings of the Seventh International Joint Conference on Artificial Intelligence*, pages 153-157.

## Abstract

A system learning concepffi from training samples consisting of structured objec ts is described. It is based on descriptions invariant under 1somorph1sm. In order to get a unified mathematical formalism recent graph theoret1c results are used. The structures are transformed into feature vectors and after that a concept learn1ng algor ithm developing decision trees is applied which is an extension of algorithms foun d in psychological experiments. It corresponds to a general-to-spec1f1c depth-f1rst sea rch with reexamination of past events. The generalization ability is demonstrated by means of the blocks world example and it is shown that the algorithm can successfully handle pract1cal problems w1th samples of abou t one hundred of relatively complicated structures 1n a reasonable t1me. . Additionally, the problem of representation and learning context dependent concep~ 1s disc ussed in the paper.