Food

Abductive Equivalential Translation and its application to Natural Language Database Interfacing

Description
Abductive Equivalential Translation and its application to Natural Language Database Interfacing
Categories
Published
of 163
All materials on our website are shared by users. If you have any questions about copyright issues, please report us to resolve them. We are always happy to assist you.
Related Documents
Share
Transcript
  arXiv:cmp-lg/9405024v1 26 May 1994  Abductive Equivalential Translationand its application toNatural Language Database Interfacing Dissertation submitted in partial fulfilment of the requirements for the degree of  Doctor of PhilosophyRoyal Institute of Technology, Stockholm Manny Rayner SRI International, Cambridge September 1993  Abstract The thesis describes a logical formalization of natural-language database in-terfacing. We assume the existence of a “natural language engine” capableof mediating between surface linguistic string and their representations as“literal” logical forms: the focus of interest will be the question of relating“literal” logical forms to representations in terms of primitives meaningfulto the underlying database engine. We begin by describing the nature of theproblem, and show how a variety of interface functionalities can be consid-ered as instances of a type of formal inference task which we call “AbductiveEquivalential Translation” (AET); functionalities which can be reduced tothis form include answering questions, responding to commands, reasoningabout the completeness of answers, answering meta-questions of type “Doyou know...”, and generating assertions and questions. In each case, a “lin-guistic domain theory” (LDT) Γ and an input formula  F   are given, and thegoal is to construct a formula with certain properties which is equivalent to F  , given Γ and a set of permitted assumptions. If the LDT is of a certainspecified type, whose formulas are either conditional equivalences or Horn-clauses, we show that the AET problem can be reduced to a goal-directedinference method. We present an abstract description of this method, andsketch its realization in Prolog. The relationship between AET and severalproblems previously discussed in the literature is discussed. In particular,we show how AET can provide a simple and elegant solution to the so-called“Doctor on Board” problem, and in effect allows a “relativization” of theClosed World Assumption. The ideas in the thesis have all been implementedconcretely within the SRI CLARE project, using a real projects and pay-ments database. The LDT for the example database is described in detail,and examples of the types of functionality that can be achieved within theexample domain are presented. KEYWORDS : Natural language processing, natural language inter-faces, databases, logic programming, equivalences, Closed World Assump-tion, “Doctor on Board” problemi  ii  Acknowledgements Nearly every PhD thesis contains a lengthy acknowledgements section, andthis one is no exception. I would like to start by admitting my debt to theother members of the CLARE project: Hiyan Alshawi, Dave Carter, DickCrouch, Steve Pulman and Arnold Smith. Without the unique contributionsthey each made to the project, the work I report here could quite simplynot have been carried out. I would particularly like to thank Hiyan Alshawi.Many of the key ideas grew out of discussions with Hiyan, and I am sure thatsome of them are really more his than mine. I would also like to thank theCLARE sponsors: BP Research, British Aerospace, British Telecom, SRIInternational, the UK Defence Research Agency, and the UK Departmentof Trade and Industry.I would never have gotten around to writing the thesis if I hadn’t beenpatiently persuaded to do so by Carl-Gustaf Jansson, my supervisor, Elisa-beth Kron, my wife, and Lotta Almgren and Alison Golding, our sometime au pairs  . I hope they consider their efforts rewarded. For similar reasons,I must also thank my own thesis students, Bj¨orn Gamb¨ack, Barney Pelland Christer Samuelsson, for mercilessly teasing me during periods when Ithought I wanted to be doing quite different things. Jerry Hobbs, ¨OstenDahl, Seif Haridi and Sture H¨agglund agreed to act as my opponent and ex-aminers respectively despite the heavy demands made on their time. Manypeople encouraged me with kind remarks about previous drafts of this work,and among them I would particularly like to mention Barbara Grosz andKaren Sparck-Jones.Finally I would like to thank you, the reader. Please read on: there area few good bits, I think, in the middle of all the boring stuff. And I needyour discerning mind to tell me which are which.Manny RaynerCambridge, September 1993iii
Search
Similar documents
View more...
Tags
Related Search
We Need Your Support
Thank you for visiting our website and your interest in our free products and services. We are nonprofit website to share and download documents. To the running of this website, we need your help to support us.

Thanks to everyone for your continued support.

No, Thanks