Pereira, Fernando C. N Prolog and natural-language analysis 1987.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....[3, 16, 11] as well as corresponding theoretical formalisms [9, 15] started to appear. Concerning the use of deductive databases there exist so far very few approaches (e.g. 26] which is quite surprising if one considers the long tradition of logic programming in natural language processing [19, 6]. One of the main obstacles to the successful practicable use of natural language interfaces is the often necessary high amount of manual knowledge engineering. Manual knowledge engineering is a timeconsuming and tedious process that may require extensive efforts by highly experienced experts in ....

Pereira, F. C. N., Shieber, S. M.: Prolog and Natural-language Analysis. CSLI, Stanford (1987)

....findall(Goal, top down interpret(Goal) SolutionList) generalize all solutions(SolutionList,GeneralizedSolutionsForGoal) FIGURE 9 Generalizing all solutions for goal. Figure 10 provides the definition of top down interpret 1, a top down interpreter taken from Pereira and Shieber (1987, pp. 160f. 7 This interpreter top down interpret(true) top down interpret(Goal) clause( Goal : Body) top down interpret(Body) top down interpret( Body1, Body2) top down interpret(Body1) top down interpret(Body2) FIGURE 10 A simple top down interpreter. 5 If some kind of ....

....to massive nondeterminism, this information can be used to automatically make get goal 1 select the relevant goals. 6 Notice that in case there exists only one solution to a goal, the effect of performing constraint propagation on that goal is similar to its partial evaluation, see, for example, Pereira and Shieber 1987. 7 The predicate is renamed here for expository reasons. The nonunit and unit clauses representing an hpsg theory are represented as clause( Head : Body ) and clause( Head : true ) respectively. ix x Walt Detmar Meurers and Guido Minnen falls prey to nontermination. For example, ....

Pereira, Fernando, and Stuart Shieber. 1987. Prolog and Natural Language Analysis. CSLI Lecture Notes, No. 10. Chicago, USA: Chicago University Press.

....has been for a long time one of the programming paradigms of choice for the development of NLP systems. Logic programming languages (e.g. Prolog) offer features such as backtracking, unification, and symbolic data representation which greatly facilitate the development of NLP applications (Pereira 1987). In recent years the use of declarative languages, such as Prolog and Lisp, for NLP applications has declined. This is mostly due to the increasing computational requirements of these applications, which were not satisfied by the older slow implementations of declarative languages. Even though ....

Pereira, F. and Shieber, S.M. 1987. Prolog and NaturalLanguage Analysis. Cambridge University Press.

....problems though. First, as pointed out by Pereira and Shieber, because the two DCG rules for verb phrases with ditransitive verbs are identical when Gap is instantiated to nogap, simple sentences with ditransitive verbs will have two identical parses, depending on which rule is used in the parse [26]. In addition, the system of switches is too limited for grammars intended to handle multiple extractions from nested structures. This problem is actually present in GPSG itself, since there is only a single level of slashing. Therefore gap threading parsers (as these are called) are typically ....

....rule for ditransitive verb phrases. The di#erence list of gaps is divided between the two object noun phrases; if there is just one gap in the list it can be used in only one of the two noun phrases. This technique for implementing GPSG parsers has been developed extensively by Pereira and Shieber [26] and others. Unfortunately, while using a di#erence list to manage gap distribution solves the problems mentioned, the solution is still unsatisfactory. In particular, a problem common to both implementations is that they introduce a good deal of complexity in constructing the grammar. Insuring ....

Fernando C. N. Pereira and Stuart M. Shieber. Prolog and Natural-Language Analysis, volume 10. CLSI, Stanford, CA, 1987.

.... assume the existence of a noun phrase (a gap) and then attempt to parse a sentence [24] Intuitionistic contexts do not naturally enforce the constraint that the assumed gap must be used while parsing the relative clause and that the gap cannot appear in certain positions ( island constraints [27]) ffl Intuitionistic contexts can be used to manage a data base. While adding facts, querying facts, and performing hypothetical reasoning ( if I pass CS121 will I graduate ) are easy to model using intuitionistic contexts, updating and retracting facts are not straightforward [3, 9, 19] ffl A ....

....rel( whom, bob, married, ann] are not. As this parser rules out subject extraction, sentences that require such extractions must be handled with additional specialized grammer rules. Several similar types of island constraints occur in natural language parsing problems [27]. The use of ed formulas may aid in handling these constraints as well. 6 Related Work There are many ways in which linear logic can be fruitfully exploited to address aspects of logic programming. Girard modeled the difference between the classical, external logic of Horn clauses and the ....

Fernando C. N. Pereira and Stuart M. Shieber. Prolog and Natural-Language Analysis, volume 10. CLSI, Stanford, CA, 1987.

....rules where possible. When fi reduction amounts to simply instantiating variables according to the structure of the terms, explicit fi reduction can be replaced by appropriately structuring the argument positions where the logical forms are returned. This technique is called partial execution (Pereira 1987). A higher order DCG rule is partially executed by recursively applying all functions to their arguments; if a function variable has not been instantiated but is applied to some argument, all of its occurrences in the rule are instantiated to an abstraction of the form X Y, where X and Y are new ....

Pereira, F. C. N. (1987) "Prolog and Natural-Language Analysis," CSLI.

....Baltimore, MD 21228 5398. Voice: 410 455 3099. Fax: 410 4553969. Email: mayfield cs.umbc.edu. 2 1 Introduction Memo functions and memoization are well known concepts in AI programming. They have been discussed since the Sixties and are often used as examples in introductory programming texts [16, 13, 14]. However, the automation of memoization as a practical software engineering tool for AI systems has never received a detailed treatment. This paper describes how automatic memoization can be made viable on a large scale. It points out advantages and uses of automatic memoization not previously ....

F. C. N. Pereira and S. M. Shieber. Prolog and Natural-Language Analysis. csli, Stanford, CA, 1987.

....expression. The unification of Arg into the lambda feature of Fun will at the same time take care of the embedded operation application of to its argument. We use the well known technique of partially evaluating the expressions in the functor s list with regard to functional application (Pereira and Shieber, 1987). The examples in Section 4.3 below illustrate how this works. In the implementation, no ff conversion takes place as usual in unification based simulations of the calculus. In addition, we have to pass up both the quantifier stores and the anchors of functor and argument. 5 For an n place ....

Pereira, F. and S. Shieber (1987). Prolog and Natural Language Analysis. Number 10 in CSLI Lecture Notes.

....Acknowledgements. This work was supported in part by the Advanced Research Projects Agency . 1 Introduction Memo functions and memoization are well known concepts in AI programming. They have been around since the sixties and are often used as examples in introductory programming texts [15, 12, 13]. However, the automation of memoization as a practical software engineering tool for AI systems has never received a detailed treatment. This paper describes how automatic memoization can be made viable on a large scale. It points out advantages and uses of automatic memoization not previously ....

F. C. N. Pereira and S. M. Shieber. Prolog and Natural-Language Analysis. csli, Stanford, CA, 1987.

....has been for a long time one of the programming paradigms of choice for the development of NLP systems. Logic programming languages (e.g. Prolog) offer features such as backtracking, unification, and symbolic data representation which greatly facilitate the development of NLP applications (Pereira 1987). In recent years the use of declarative languages, such as Prolog and Lisp, for NLP applications has declined. This is mostly due to the increasing computational requirements of these applications, which were not satisfied by the older slow implementations of declarative languages. Even though ....

Pereira, F. and Shieber, S.M. 1987. Prolog and Natural-Language Analysis. CSLI Lecture Notes, number 10. Cambridge University Press.

....a verb phrase. Examples of common grammar types based on these formalisms are described later on in this chapter. 2.3. 1 Phrase Structure Grammar The most common method of representing the structural constituency is the Phrase structure tree with its corresponding Phrase Structure Grammar (PSG) [8]. A PSG is simply an ordinary context free grammar, with the non terminals as rules and the terminals covering the lexical elements in the language. A grammar for a small language can look like this; Grammar Lexicon S NP VP Det a VP V P about VP V NP NP ovens VP VP PP NP Olle NP ....

Fernando C.N. Pereira, Stuart M. Shieber (1987). Prolog and Natural-Language Analysis. Center for the Study of Language and Information, Leland Stanford Junior University. ISBN 0-937073-18-0

....Problem Consider the problem of learning a grammar for English from examples of parsed sentences. In principle, this can be done by learning a Prolog parsing program. In order to test this idea we generated examples for the following grammar, which is a small modification of ones described by Pereira and Shieber (1987). s(H2,P0,P) np(H1,P0,P1) vp(H2,P1,P) number(H1,X) number(H2,X) np(H,P0,P) det(HD,P0,P1) n(H,P1, P) number(HD,X) number(H,X) np(H,P0,P) det(HD,P0,P1) n(H,P1, P2) rel(H2,P2,P) number(H,X) number(H2,X) number(HD,X) np(H,P0,P) pn(H,P0,P) vp(H,P0,P) ....

Pereira, F., & Shieber, S. (1987). Prolog and naturallanguage analysis. Stanford : Center for the Study of Language and Information.

....resources for Dutch. Since students nowadays have access to powerful hardware and both tools and data can be distributed easily over the internet, there are no real practical obstacles. Text books which are concerned primarily with computational semantics and natural language interfaces, such as Pereira and Shieber (1987) and Blackburn and Bos (1998) tend to introduce a toy domain, such as a geography database or an excerpt of a movie script, as application area. In trying to develop exercises which are closer to real applications, we have explored the possibilities of using webaccessible databases as back end ....

....of less than 20 sentences each) and constructed by hand. This reflects the fact that this exercise is primarily concerned with the implementation of a sophisticated linguistic analysis. 4 Natural Language Interfaces Practical courses in natural language interfaces or computational semantics (Pereira and Shieber 1987, Blackburn and Bos 1998) have used a toy database, such as a geographical database or an excerpt of a movie script, as application domain. The growing amount of information available on the internet provides opportunities for accessing much larger databases (such as public transport timetables or ....

Pereira, F. C. and Shieber, S. M.(1987), Prolog and Natural Language Analysis, Center for the Study of Language and Information Stanford.

....of propositions with parses that assign grammatical types of constituency, such as noun phrase (np) or sentence (sent) to strings in natural language. This was, in fact, one of the guiding principles behind the creation of Prolog, which was originally designed for natural language processing [21]. To find whether John likes Mary is a grammatical sentence in English in a natural deduction style, we can try to find a proof that the sentence is in L(G) where G is the putative LCG for English, with a distinguished category, sent, corresponding to sentence hood: u 1 np u 2 sentnnp=np u ....

Pereira, F.C.N. and Shieber, S.M., 1987. Prolog and Natural-Language Analysis. CSLI Lecture Notes, Stanford. 11

....may apply twice (steps 3 4) first to scan the and then to scan duck. 4 Note here that although duck is ambiguous between N and V, the top down prediction only allows it to be parsed as a V. 3 Chart parsing was introduced by Kaplan (1973) and Kay (1973) Introductory treatments are provided in Pereira and Shieber (1987), Gazdar and Mellish (1989) and Ross (1989) The discussion of Earley s algorithm here will be necessarily brief. For details of the algorithm, the reader is referred to Earley (1970) For a logical reconstruction of this parsing algorithm, see Pereira and Warren (1983) For augmentation of ....

....in putting active states on the top and inactive states on the bottom. 7 A parser can be obtained from this recognizer either by building up a tree structure with each scan completion step or by incorporating features and building up the tree structure as a complex valued feature as described in Pereira and Shieber (1987). The latter approach is more general since it allows a semantic analysis of the sentence to be built up in the same manner as the syntactic structure. Some Open Problems in Head driven Generation 169 no guidance. It makes more sense to follow the strategy of generating heads first and then ....

[Article contains additional citation context not shown here]

Pereira, F. and S. Shieber. 1987. Prolog and Natural Language Analysis. CSLI Lecture Notes No. 10. Chicago: Chicago University Press.

....as well as eliminating it, in which case LHIP will suffer from the same termination problems as an ordinary DCG formalism interpreted in this way. And as with an ordinary DCG formalism, it is possible to apply different parsing methods to LHIP in order to circumvent these problems (see e.g. Pereira and Shieber, 1987). A related issue concerns the interpretation of embedded Prolog code. Reordering of RHS clauses will result in code which precedes a head within a LHIP rule being evaluated after it; judicious freezing of goals and avoidance of unsafe cuts are therefore required. LHIP provides a number of ways ....

Pereira, F.C.N. and S.M. Shieber (1987) Prolog and Natural Language Analysis, CSLI Lecture Notes No. 10, Stanford University.

....fP (v)gi ) h; fQgigi and semantic composition is defined in terms of a function application principle. In these systems scope ambiguity is accounted for in terms 4 In fact often on both in the sense that a calculus style reformulation is actually implemented (or often rather approximated c.f. [Pereira and Shieber, 1987]) in terms of unification in a logical programming language. of the order of application of partial DRS resulting in sets of fully specified representations. 5 Apart from their historical interest, what seems remarkable about these systems is the notion of a partial DRS which in some form or ....

.... like [ Johnson and Klein, 1986 ] the approach is based on the threading technique but in [ Johnson and Kay, 1990 ] this is hidden in the definition of semantic constructor operators on the one hand and in a Prolog based abstraction approximation of semantic composition in the style of [ Pereira and Shieber, 1987 ] In addition, the paper provides an abstraction barrier between syntactic and semantic representations in terms of three sets of semantic constructor definitions resulting in Montague grammar, Situation Semantics and Discourse Representation Theory style representations depending on which ....

[Article contains additional citation context not shown here]

Pereira, F.C.N. and Shieber, S.M. 1987. Prolog and Natural Language Analysis, volume 10 of CSLI Lecture Notes. CSLI, Stanford. Distributed by University of Chicago Press.

....with semantic rules composed of operator expressions. Actually, the picture is slightly more complicated: in addition to syntactic phrase structure rules we also have Montague style semantic annotations for those rules. The annotations are given in the manner of the semantic composition rules in [ Pereira and Shieber, 1987 ] where abstraction and fi conversion are approximated 4 in terms of Prolog variables and unification. Types are implicit and intensionality is not accounted for. The semantic annotations are mainly responsible to link the semantic argument positions of e.g. verbs with the variables or ....

Pereira, F.C.N. and Shieber, S.M. 1987. Prolog and Natural Language Analysis, volume 10 of CSLI Lecture Notes. CSLI, Stanford. Distributed by University of Chicago Press.

....expression. The unification of Arg into the lambda feature of Fun will at the same time take care of the embedded operation application of to its argument. We use the well known technique of partially evaluating the expressions in the functor s list with regard to functional application (Pereira and Shieber, 1987). The examples in Section 4.3 below illustrate how this works. In the implementation, no ff conversion takes place as usual in unification based simulations of the calculus. In addition, we have to pass up both the quantifier stores and the anchors of functor and argument. 5 For an n place ....

Pereira, F. and S. Shieber (1987). Prolog and Natural Language Analysis. Number 10 in CSLI Lecture Notes. Standford: CSLI.

....##### ,###### #### ###### ####### ######### #### ### ###### ## ###### ##### #### ### , reversible) ######## ## ####### ,## ##### .###### ###### #### ######### ##### ## #### ####### ###### ##### . Shieber, 1988; Strzalkowski, 1994) ## ##### ###### ,##### ####### ####### #### ######## ###### #### ,(Pereira and Shieber, 1987) Definite Clause Grammar (DCG) ###### . Typed Feature Structures TFS) ##### #### ####### #### ## ######## ########## ###### ##### ######### ##### ### ########### , unification) ####### #### ## ##### ### #### ########## ## ####### . Carpenter, 1992b) ###### ## ######### ###### ## ##### ....

....### ALE #### .#### ##### ##### ## ######### ### ###### ### #### #### ### , Earley, 1970) # (Younger, 1967) ## ########## ####### #### #### ##### ##### ###### ##### ### ##### ####### ### ########## . Earley # (CYK) Cocke Younger Kasami ######## #### ##### ##### #### ##### ##### ## #### #### #### (Pereira and Shieber, 1987, Section 6.6) ###### ## ### ##### ### ##### ### ###### ### ,#### ##### ##### ######## ###### ### ### ###### .###### ##### ##### ,###### ###### ###### ## ##### #### ###### #### ## ##### ,### ## .#### ##### ##### .###### ## ##### ###### ### #### ####### #### ##### ,Reiter ## ##### .###### #### ....

[Article contains additional citation context not shown here]

Pereira, Fernando C.N. and Stuart M. Shieber. 1987. Prolog and Natural-Language Analysis. Number 10 in CSLI Lecture Notes. CSLI, Stanford University, Stanford, CA.

....Schemata The application of deductive transformation rules, e.g. the unfold fold rules, requires user intervention and thus makes it difficult to automate transformations. Strategies have been proposed to overcome practical problems of individual unfold fold rules and to automate transformations [Pereira Shieber 87; Lakhotia Sterling 88; Nielson Nielson 90; Proietti Pettorossi 90] However, these strategies do not lead easily to strategies for transformations consisting of several applications of unfold fold rules. We suggest a new approach that is based on transformation schemata and promises a ....

F. C. N. Pereira, S. M. Shieber, Prolog and Natural-Language Analysis, CSLI, Lectures Notes 10, 1987

No context found.

Pereira, Fernando C. N Prolog and natural-language analysis 1987.

No context found.

F.C.N. Pereira and S.M. Shieber. Prolog and Natural Language Analysis. CSLI Publications, 1987.

No context found.

Fernando C.N. Pereira and Stuart M. Shieber. 1987. Prolog and Natural-Language Analysis. CSLI.

No context found.

Fernando C.N. Pereira and Stuart M. Shieber. 1987. Prolog and Natural-Language Analysis.

First 50 documents  Next 50

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