Giordana, A., and Neri, F. 1994. Search internsive concept induction. Evolutionary Computation 3(4):375--416. Home/Search Document Not in Database Summary Related Articles Check |
This paper is cited in the following contexts:
Tractable Induction and Classification in First Order .. - Michèle.. (1997) (17 citations) (Correct)
.... to tackle complex domains such as organic chemistry (King, Srinivasan, Sternberg, 1995) or natural language (Mooney, 1996) And a number of FOL learners have been developed: FOIL (Quinlan, 1990) ML Smart (Bergadano Giordana, 1990) FOCL (Pazzani Kibler, 1992) KBG (Bisson, 1992) REGAL (Giordana Neri, 1994), PROGOL (Muggleton, 1995) RIBL (Emde Wettscherek, 1996) and REMO (Zucker Ganascia, 1996) to name a few. The specific difficulty of ILP is related to the matching step: If a clause involves three literals part and the description of a given device involves 40 parts, there are 40 3 possible ....
....induction and deduction are exponential in the size of FOL hypotheses. This limitation is dealt with in the literature by inducing short hypotheses, by means of search strategies (Quinlan, 1990; Bergadano Giordana, 1990; Pazzani Kibler, 1992; Muggleton, 1995) and or syntactic restrictions (Giordana Neri, 1994; Kietz Lubbe, 1994; Muggleton, 1995) This paper is interested in polynomial induction and use of FOL hypotheses, with no size restrictions. This is made possible by stochastic matching: one only considers some samples of matchings between examples and hypotheses instead of exhaustively ....
[Article contains additional citation context not shown here]
Giordana, A., and Neri, F. 1994. Search internsive concept induction. Evolutionary Computation 3(4):375--416.
Subsumption Stochastique, et Apprentissage Polynomial .. - Michèle.. (1997) (Correct)
....les roues d une voiture) cf [23] pour une discussion d etaill ee. La PLI apparait ainsi comme la voie royale de l apprentissage dans des domaines structur es tels la chimie organique [8] ou le langage naturel [13] Citons, parmi les nombreux syst emes de PLI existants : FOIL [17] KBG [1] REGAL [4], PROGOL [16] et RIBL [3] La difficult e a laquelle tout syst eme de PLI est confront e est le test de couverture entre une hypoth ese et un exemple. Pla cons nous dans le cas de langage de la logique des pr edicats sans symbole de fonctions autres que des constantes (DATALOG) Si une hypoth ese ....
....dit, l induction et la classification de formules en logique des pr edicats sont exponentielles en fonction de la taille des hypoth eses. Tous les syst emes de PLI existants forcent donc l induction d hypoth eses de petite taille, au moyen de biais de recherche [17, 16] et ou de biais de langage [4, 7, 16]. Cet article pr esente une heuristique permettant d induire et d utiliser des hypoth eses en Logique des pr edicats avec un cout polynomial, sans restriction a priori de l espace de recherche ou de la syntaxe des hypoth eses. L approche propos ee se fonde sur la notion de subsomption ....
[Article contains additional citation context not shown here]
Giordana, A., and Neri, F. Search internsive concept induction. Evolutionary Computation 3(4):375--416, 1994.
Online articles have much greater impact More about CiteSeer.IST Add search form to your site Submit documents Feedback
CiteSeer.IST - Copyright Penn State and NEC