On Learning From Exercises

Natarajan, B.K.

doi:10.1016/b978-0-08-094829-4.50008-8

Cited by 9 publications

(2 citation statements)

References 5 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The full paper on Ulearnability includes a definition of polynomial U-learnability that allows the teacher to provide the learner with a background theory. A surprising aspect of this definition is that it not only captures the notion of inductive learning relative to background knowledge, as desired, but it also provides a generalisation of Natarajan's PAC-style formalisation of speed-up learning [33]. Although Natarajan's formalisation is quite appealing, few positive have been proven within it because it is also very demanding.…”

Section: Incorporating Background Knowledgementioning

confidence: 95%

Bayesian inductive logic programming

Muggleton

1994

Proceedings of the Seventh Annual Conference on Computational Learning Theory - COLT '94

View full text Add to dashboard Cite

Section: Incorporating Background Knowledgementioning

confidence: 95%

Bayesian inductive logic programming

Muggleton

1994

Proceedings of the Seventh Annual Conference on Computational Learning Theory - COLT '94

View full text Add to dashboard Cite

“…In many systems, learning hierarchically organized knowledge assumes that the structure of hierarchy or the order of the literals is known to the learner. Examples of such work include Marvin (Sammut & Banerji, 1986) and XLearn (Reddy & Tadepalli, 1997a), on the experimental side; learning from exercises by Natarajan (1989) and learning acyclic Horn sentences by Arimura (1997), on the theoretical side. In fact, Khardon shows that learning hierarchical strategies can be computationally hard when the structure of the hierarchy is not known (Khardon, 1996).…”

Section: Introductionmentioning

confidence: 99%

Machine Learning: Proceedings of the Fifteenth International Conference

Shavlik¹

1998

View full text Add to dashboard Cite

Public reporting burden for this collection of information is estimated to average 1 hour per response, including the time for reviewing instructions, searching existing data sources, gathering and maintaining the data needed, and completing and reviewing the collection of information. Send comments regarding this burden estimate or any other aspect of this collection of information, including suggestions for reducing this burden to Washington Headquarters Services, Directorate for Information Operations and Reports, 1215 Jefferson Davis Highway, Suite 1204, Arlington, VA 22202-4302, and to the Office of Management and Budget, Paperwork Reduction Project (0704-0188), Washington, DC 20503. AGENCY USE ONLY (Leave blank)2. REPORT DATE July 1998 REPORT TYPE AND DATES COVEREDFinal, 7/10/98-9/30/98 TITLE AND SUBTITLEMachine Learning: Proceedings of the Fifteenth International Conference AUTHOR(S)Jude Shavlik (editor) FUNDING NUMBERSN00014-98-1-0810 PERFORMING ORGANIZATION NAMES(S) AND ADDRESS(ES)Computer Sciences Department University of Wisconsin 1210 W. Dayton Street Madison, WI 53706 PERFORMING ORGANIZATION REPORT NUMBER144-HD17 SPONSORING /MONITORING AGENCY NAMES(S) AND ADDRESS(ES)Office All rights reserved. Printed in the United StatesNo part of this publication may be reproduced, stored in a retrieval system, or transmitted in any form or by any means^lectronic, mechanical, photocopying, recording, or otherwise without the prior written permission of the publisher. PrefaceThis volume contains the 66 technical papers presented at the Fifteenth International Conference on Machine Learning (ICML '98), held July 24r-27,1998, in Madison, Wisconsin U.S.A. These articles were selected based on the rigorous review and discussion of 215 submissions. All papers were presented orally as well as at an evening poster session.ICML '98 was one of ten AI-related conferences held in Madison during mid-summer 1998, in an ambitious, first-time experiment to see what kind of synergies would result from all this collocation. In particular, ICML '98 was held in the same building as, and concurrently with, the Computational Learning Theory (COLT) and Uncertainty in Artificial Intelligence (UAI) conferences. It also overlapped one day with the Inductive Logic Programming (ILP) conference.Registrants were allowed to attend, without additional costs, the technical sessions of the other conferences. COLT, ICML, and UAI each invited one of the plenary speakers and jointly invited the banquet speaker, plus there was a joint poster session and wrapup panel. The poster session contained about 150 papers, allowing conference attendees a chance to see or further discuss research presented during the four to five parallel tracks and fostering interaction among the various communities.I especially wish to thank:• the authors of all the papers for their technical contributions toward the advancement of machine learning.• Richard Sutton, who was the IMCL representative among the three joint invited speakers and spoke on "Reinforcement Learning: Ho...

show abstract

Learning first-order acyclic Horn programs from entailment

Reddy

Tadepalli

1998

Inductive Logic Programming

View full text Add to dashboard Cite

On Learning From Exercises

Cited by 9 publications

References 5 publications

Bayesian inductive logic programming

Bayesian inductive logic programming

Machine Learning: Proceedings of the Fifteenth International Conference

Learning first-order acyclic Horn programs from entailment

Contact Info

Product

Resources

About