TY - GEN

T1 - Learning coherent concepts

AU - Garg, Ashutosh

AU - Roth, Dan

N1 - Publisher Copyright:
© Springer-Verlag Berlin Heidelberg 2001.
Copyright:
Copyright 2020 Elsevier B.V., All rights reserved.

PY - 2001

Y1 - 2001

N2 - This paper develops a theory for learning scenarios where multiple learners co-exist but there are mutual coherency constraints on their outcomes. This is natural in cognitive learning situations, where “natural” constraints are imposed on the outcomes of classifiers so that a valid sentence, image or any other domain representation is produced. We formalize these learning situations, after a model suggested in [11] and study generalization abilities of learning algorithms under these conditions in several frameworks. We show that the mere existence of coherency constraints, even without the learner’s awareness of them, deems the learning problem easier than predicted by general theories and explains the ability to generalize well from a fairly small number of examples. In particular, it is shown that within this model one can develop an understanding to several realistic learning situations such as highly biased training sets and low dimensional data that is embedded in high dimensional instance spaces.

AB - This paper develops a theory for learning scenarios where multiple learners co-exist but there are mutual coherency constraints on their outcomes. This is natural in cognitive learning situations, where “natural” constraints are imposed on the outcomes of classifiers so that a valid sentence, image or any other domain representation is produced. We formalize these learning situations, after a model suggested in [11] and study generalization abilities of learning algorithms under these conditions in several frameworks. We show that the mere existence of coherency constraints, even without the learner’s awareness of them, deems the learning problem easier than predicted by general theories and explains the ability to generalize well from a fairly small number of examples. In particular, it is shown that within this model one can develop an understanding to several realistic learning situations such as highly biased training sets and low dimensional data that is embedded in high dimensional instance spaces.

UR - http://www.scopus.com/inward/record.url?scp=84948157744&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84948157744&partnerID=8YFLogxK

U2 - 10.1007/3-540-45583-3_12

DO - 10.1007/3-540-45583-3_12

M3 - Conference contribution

AN - SCOPUS:84948157744

SN - 3540428755

SN - 9783540428756

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 135

EP - 150

BT - Algorithmic Learning Theory - 12th International Conference, ALT 2001, Proceedings

A2 - Abe, Naoki

A2 - Khardon, Roni

A2 - Zeugmann, Thomas

PB - Springer-Verlag Berlin Heidelberg

T2 - 12th Annual Conference on Algorithmic Learning Theory, ALT 2001

Y2 - 25 November 2001 through 28 November 2001

ER -