CLASSIC Learning

Description logics , also called terminological logics , are commonly used in knowledge-based systems to describe objects and their relationships. We investigate the learnability of a typical description logic, CLASSIC, and show that CLASSIC sentences are learnable in polynomial time in the exact learning model using equivalence queries and membership queries (which are in essence, “subsumption queries”). We show that membership queries alone are insufficient for polynomial time learning of CLASSIC sentences. Combined with earlier negative results of Cohen and Hirsh showing that, given standard complexity theoretic assumptions, equivalence queries alone are insufficient (or random examples alone in the PAC setting are insufficient), this shows that both sources of information are necessary for efficient learning in that neither type alone is sufficient. In addition, we show that a modification of the algorithm deals robustly with persistent malicious two-sided classification noise in the membership queries with the probability of a misclassification bounded below 1/2.

[1]  Dana Angluin,et al.  When won't membership queries help? , 1991, STOC '91.

[2]  Ronald L. Rivest,et al.  Inference of finite automata using homing sequences , 1989, STOC '89.

[3]  Sunit K. Gala,et al.  Classification as a query processing technique in the CANDIDE semantic data model , 1989, [1989] Proceedings. Fifth International Conference on Data Engineering.

[4]  Dana Angluin,et al.  Learning Regular Sets from Queries and Counterexamples , 1987, Inf. Comput..

[5]  Premkumar T. Devanbu,et al.  LaSSIE—a knowledge-based software information system , 1991, ICSE '90.

[6]  W. Hoeffding.  Probability Inequalities for sums of Bounded Random Variables , 1963 .

[7]  Scott E. Decatur.  Statistical queries and faulty PAC oracles , 1993, COLT '93.

[8]  Daniel G. Bobrow,et al.  On Overview of KRL, a Knowledge Representation Language , 1976, Cogn. Sci..

[9]  William W. Cohen.  Pac-Learning a Restricted Class of Recursive Logic Programs , 1993, AAAI.

[10]  Peter F. Patel-Schneider,et al.  A Semantics and Complete Algorithm for Subsumption in the CLASSIC Description Logic , 1993, J. Artif. Intell. Res..

[11]  Saso Dzeroski,et al.  PAC-learnability of determinate logic programs , 1992, COLT '92.

[12]  N. Littlestone.  Learning Quickly When Irrelevant Attributes Abound: A New Linear-Threshold Algorithm , 1987, 28th Annual Symposium on Foundations of Computer Science (sfcs 1987).

[13]  Dana Angluin,et al.  Computational learning theory: survey and selected bibliography , 1992, STOC '92.

[14]  Dana Angluin,et al.  Learning with malicious membership queries and exceptions (extended abstract) , 1994, COLT '94.

[15]  Peter Auer,et al.  On-line learning of rectangles in noisy environments , 1993, COLT '93.

[16]  Leslie G. Valiant,et al.  Cryptographic limitations on learning Boolean formulae and finite automata , 1994, JACM.

[17]  Avrim Blum,et al.  Learning switching concepts , 1992, COLT '92.

[18]  Dana Angluin,et al.  Learning with hints , 1988, COLT '88.

[19]  Jr. Charles David Page.  Anti-unification in constraint logics: foundations and applications to learnability in first-order logic, to speed-up learning, and to deduction , 1993 .

[20]  Ronitt Rubinfeld,et al.  Learning fallible finite state automata , 1993, COLT '93.

[21]  David Haussler,et al.  Predicting {0,1}-functions on randomly drawn points , 1988, COLT '88.

[22]  D. Angluin.  Queries and Concept Learning , 1988 .

[23]  William W. Cohen,et al.  Learning the Classic Description Logic: Theoretical and Experimental Results , 1994, KR.

[24]  Leslie G. Valiant,et al.  A theory of the learnable , 1984, STOC '84.

[25]  Alexander Borgida,et al.  Computing Least Common Subsumers in Description Logics , 1992, AAAI.

[26]  William W. Cohen.  Cryptographic Limitations on Learning One-Clause Logic Programs , 1993, AAAI.

[27]  William W. Cohen,et al.  Learnability of description logics , 1992, COLT '92.

[28]  Michael Frazier,et al.  Learning From Entailment: An Application to Propositional Horn Sentences , 1993, ICML.

[29]  Leonard Pitt,et al.  Prediction-Preserving Reducibility , 1990, J. Comput. Syst. Sci..

[30]  György Turán,et al.  Learning with queries but incomplete information (extended abstract) , 1994, COLT '94.

[31]  Michael Frazier,et al.  Learning from a consistently ignorant teacher , 1994, COLT '94.

[32]  Peter F. Patel-Schneider,et al.  A Four-Valued Semantics for Terminological Logics , 1989, Artif. Intell..

[33]  Hector J. Levesque,et al.  Krypton: A Functional Approach to Knowledge Representation , 1983, Computer.

[34]  George Shackelford,et al.  Learning k-DNF with noise in the attributes , 1988, Annual Conference Computational Learning Theory.

[35]  D. Angluin,et al.  Randomly fallible teachers: Learning monotone DNF with an incomplete membership oracle , 1991, Machine Learning.

[36]  David Haussler,et al.  Learnability and the Vapnik-Chervonenkis dimension , 1989, JACM.

[37]  Ming Li,et al.  Learning in the Presence of Malicious Errors , 1993, SIAM J. Comput..

[38]  Chidanand Apté,et al.  Organizing Knowledge in a Complex Financial Domain , 1987, IEEE Expert.

[39]  Thomas G. Dietterich,et al.  Learning and Inductive Inference , 1982 .

[40]  Deborah L. McGuinness,et al.  CLASSIC: a structural data model for objects , 1989, SIGMOD '89.

[41]  Zhixiang Chen,et al.  On-line learning of rectangles , 1992, COLT '92.