Catalogo Articoli (Spogli Riviste)

OPAC HELP

Titolo:
Improved lower bounds for learning from noisy examples: An information-theoretic approach
Autore:
Gentile, C; Helmbold, DP;
Indirizzi:
Univ Milan, DSI, I-20135 Milan, Italy Univ Milan Milan Italy I-20135Univ Milan, DSI, I-20135 Milan, Italy Univ Calif Santa Cruz, Dept Comp Sci, Santa Cruz, CA 95064 USA Univ Calif Santa Cruz Santa Cruz CA USA 95064 i, Santa Cruz, CA 95064 USA
Titolo Testata:
INFORMATION AND COMPUTATION
fascicolo: 2, volume: 166, anno: 2001,
pagine: 133 - 155
SICI:
0890-5401(20010501)166:2<133:ILBFLF>2.0.ZU;2-F
Fonte:
ISI
Lingua:
ENG
Soggetto:
COMPLEXITY; QUERIES; COUNTEREXAMPLES; DISTRIBUTIONS; LEARNABILITY; ERRORS; NUMBER;
Keywords:
PAC learning; entropy; mutual information; lower bounds; noisy examples;
Tipo documento:
Article
Natura:
Periodico
Settore Disciplinare:
Engineering, Computing & Technology
Citazioni:
38
Recensione:
Indirizzi per estratti:
Indirizzo: Gentile, C Univ Milan, DSI, Via Comelico 39, I-20135 Milan, Italy Univ Milan Via Comelico 39 Milan Italy I-20135 35 Milan, Italy
Citazione:
C. Gentile e D.P. Helmbold, "Improved lower bounds for learning from noisy examples: An information-theoretic approach", INF COMPUT, 166(2), 2001, pp. 133-155

Abstract

This paper presents a general information-theoretic approach for obtaininglower bounds on the number of examples required for Probably ApproximatelyCorrect (PAC) learning in the presence of noise. This approach deals directly with the fundamental information quantities, avoiding a Bayesian analysis. The technique is applied to several different models, illustrating its generality and power. The resulting bounds add logarithmic factors to (or improve the constants in) previously known lower bounds. (C) 2001 Academic Press.

ASDD Area Sistemi Dipartimentali e Documentali, Università di Bologna, Catalogo delle riviste ed altri periodici
Documento generato il 31/03/20 alle ore 15:18:43