خط مشی دسترسیدرباره ما
ثبت نامثبت نام
راهنماراهنما
فارسی
ورودورود
صفحه اصلیصفحه اصلی
جستجوی مدارک
تمام متن
منابع دیجیتالی
رکورد قبلیرکورد بعدی
Document Type : Latin Dissertation
Language of Document : English
Record Number : 149959
Doc. No : ET21751
Main Entry : Lap Poon Rupert Tang, B.S.,M.S.
Title Proper : Integrating Top-down and Bottom-up Approaches in Inductive Logic Programming: Applications in Natural Language Processing and Relational Data Mining
Note : This document is digital این مدرک بصورت الکترونیکی می باشد
Abstract : Inductive Logic Programming (ILP) is the intersection of Machine Learning andLogic Programming in which the learnerبs hypothesis space is the set of logic programs.There are two major ILP approaches: top-down and bottom-up. The formersearches the hypothesis space from general to specific while the latter the other wayround. Integrating both approaches has been demonstrated to be more eective.Integrated ILP systems were previously developed for two tasks: learning semanticparsers (Chillin), and mining relational data (Progol). Two new integratedILP systems for these tasks that overcome limitations of existing methods will bepresented.Cocktail is a new ILP algorithm for inducing semantic parsers. For thistask, two features of a parse state, functional structure and context, provide imviiportant information for disambiguation. A bottom-up approach is more suitablefor learning the former, while top-down is better for the latter. By allowing bothapproaches to induce program clauses and choosing the best combination of theirresults, Cocktail learns more eective parsers. Experimental results on learningnatural-language interfaces for two databases demonstrate that it learns moreaccurate parsers than Chillin, the previous best method for this task.Beth is a new integrated ILP algorithm for relational data mining. TheInverse Entailment approach to ILP, implemented in the Progol and Aleph systems,starts with the construction of a bottom clause, the most specific hypothesiscovering a seed example. When mining relational data with a large number of backgroundfacts, the bottom clause becomes intractably large, making learning veryinecient. A top-down approach heuristically guides the construction of clauseswithout building a bottom clause; however, it wastes time exploring clauses thatcover no positive examples. By using a top-down approach to heuristically guidethe construction of generalizations of a bottom clause, Beth combines the strengthof both approaches. Learning patterns for detecting potential terrorist activity is acurrent challenge problem for relational data mining. Experimental results on arti-ficial data for this task with over half a million facts show that Beth is significantlymore ecient at discovering such patterns than Aleph and m-Foil, two leadingILP systems.viii..tested for theQ1 PC1 bus cardBoth these projects mere sofixare des elopment efforts tonards contributing to dlfferentaspects of Roboucs and lZ1echatronics projects m the Controls and Roboucs Group..
Subject : Electericl tess
: برق
electronic file name : TL44933.pdf
Title and statement of responsibility and : Integrating Top-down and Bottom-up Approaches in Inductive Logic Programming: Applications in Natural Language Processing and Relational Data Mining [Thesis]
 
 
 
(در صورت عدم وضوح تصویر اینجا را کلیک نمایید)