Guarded Constraints in Natural Language Processing

Slides:



Advertisements
Similar presentations
Feature-based Grammar Ling 571 Deep Techniques for NLP February 2, 2001.
Advertisements

 Christel Kemke 2007/08 COMP 4060 Natural Language Processing Feature Structures and Unification.
Feature Structures and Parsing Unification Grammars Algorithms for NLP 18 November 2014.
Copyright © Cengage Learning. All rights reserved.
Statistical NLP: Lecture 3
ISBN Chapter 3 Describing Syntax and Semantics.
LTAG Semantics on the Derivation Tree Presented by Maria I. Tchalakova.
Natural Language Processing - Feature Structures - Feature Structures and Unification.
NLP and Speech Course Review. Morphological Analyzer Lexicon Part-of-Speech (POS) Tagging Grammar Rules Parser thethe – determiner Det NP → Det.
NLP and Speech 2004 Feature Structures Feature Structures and Unification.
1 Introduction to Computability Theory Lecture15: Reductions Prof. Amos Israeli.
1 Introduction to Computability Theory Lecture12: Reductions Prof. Amos Israeli.
Features & Unification Ling 571 Deep Processing Techniques for NLP January 26, 2011.
Features & Unification Ling 571 Deep Processing Techniques for NLP January 31, 2011.
Constraint Logic Programming Ryan Kinworthy. Overview Introduction Logic Programming LP as a constraint programming language Constraint Logic Programming.
Relational Data Mining in Finance Haonan Zhang CFWin /04/2003.
Sag et al., Chapter 4 Complex Feature Values 10/7/04 Michael Mulyar.
 Christel Kemke 2007/08 COMP 4060 Natural Language Processing Feature Structures and Unification.
Features and Unification
 2003 CSLI Publications Ling 566 Oct 16, 2007 How the Grammar Works.
Semi-Automatic Learning of Transfer Rules for Machine Translation of Low-Density Languages Katharina Probst April 5, 2002.
C SC 620 Advanced Topics in Natural Language Processing 3/9 Lecture 14.
Describing Syntax and Semantics
1 A Chart Parser for Analyzing Modern Standard Arabic Sentence Eman Othman Computer Science Dept., Institute of Statistical Studies and Research (ISSR),
11 CS 388: Natural Language Processing: Syntactic Parsing Raymond J. Mooney University of Texas at Austin.
LING 388: Language and Computers Sandiway Fong Lecture 17.
Empirical Methods in Information Extraction Claire Cardie Appeared in AI Magazine, 18:4, Summarized by Seong-Bae Park.
Tree-adjoining grammar (TAG) is a grammar formalism defined by Aravind Joshi and introduced in Tree-adjoining grammars are somewhat similar to context-free.
Intro to Lexing & Parsing CS 153. Two pieces conceptually: – Recognizing syntactically valid phrases. – Extracting semantic content from the syntax. E.g.,
Lesson 3 CDT301 – Compiler Theory, Spring 2011 Teacher: Linus Källberg.
2007CLINT-LIN-FEATSTR1 Computational Linguistics for Linguists Feature Structures.
1 Introduction to Computational Linguistics Eleni Miltsakaki AUTH Fall 2005-Lecture 4.
Head-driven Phrase Structure Grammar (HPSG)
CSA2050 Introduction to Computational Linguistics Lecture 1 Overview.
What you have learned and how you can use it : Grammars and Lexicons Parts I-III.
CSA2050 Introduction to Computational Linguistics Parsing I.
Section 11.3 Features structures in the Grammar ─ Jin Wang.
Supertagging CMSC Natural Language Processing January 31, 2006.
 2003 CSLI Publications Ling 566 Oct 17, 2011 How the Grammar Works.
Computer Security: Chapter 5 Operating Systems Security.
10/31/00 1 Introduction to Cognitive Science Linguistics Component Topic: Formal Grammars: Generating and Parsing Lecturer: Dr Bodomo.
Universiteit Twente Juggling Word Graphs A method for modeling the meaning of sentences using extended knowledge graphs.
Default Inheritance in Constraint-based Frameworks Christof Rumpf Heinrich-Heine-Universität Düsseldorf June 24, duesseldorf.de/~rumpf/talks/DefaultInheritance2.pdf.
WP4 Models and Contents Quality Assessment
Logic.
Approaches to Machine Translation
Names and Attributes Names are a key programming language feature
Lecture – VIII Monojit Choudhury RS, CSE, IIT Kharagpur
Statistical NLP: Lecture 3
Basic Parsing with Context Free Grammars Chapter 13
Introduction Defining the Problem as a State Space Search.
Yenka Portfolio Level for this topic: Student Name : My Levels
CS 388: Natural Language Processing: Syntactic Parsing
Instructor: Nick Cercone CSEB -
Natural Language - General
CSc4730/6730 Scientific Visualization
Lecture 7: Introduction to Parsing (Syntax Analysis)
CHAPTER 2 Context-Free Languages
Introduction to Parsing
Approaches to Machine Translation
CS246: Information Retrieval
Ling 566 Oct 14, 2008 How the Grammar Works.
Chapter 3 Describing Syntax and Semantics.
CSA2050 Introduction to Computational Linguistics
Compilers Principles, Techniques, & Tools Taught by Jing Zhang
Structure of a Lexicon Debasri Chakrabarti 13-May-19.
Lecture 7: Definite Clause Grammars
Representations & Reasoning Systems (RRS) (2.2)
C++ Object Oriented 1.
Presentation transcript:

Guarded Constraints in Natural Language Processing Kathryn L. Baker May 3, 2002 Committee: Bob Carpenter, Chair Teruko Mitamura, Co-Chair Chris Manning Carl Pollard Rich Thomason

Overview Introduction and Thesis Statement Linguistic Theories Guarded Descriptions and Guarded Rules Implementation and Evaluation Conclusions 7/8/2019

Feature Structures Feature-value pairs encode information Framework for linguistic theories (FUG, LFG, HPSG…) Widely used in NLP (frame grammars, interlinguas…) Can use unification to combine them. 7/8/2019

Example: “The dog runs.” 7/8/2019

German determiners have features nom acc dat gen masc der den dem des fem die neut das plur den is masculine accusative or dative plural. den Mann is masculine accusative den Männern is dative plural 7/8/2019

Constraints on German Determiners Constraints narrow down set of answers, and hence the search space, for a problem. e.g. X < 5, X > 2, then X = {3,4} Computational linguists can use constraints to narrow down an analysis for some piece of language. Noun constrains a determiner in German. Specifically, we can describe the information we get, such as “case=accusative” 7/8/2019

Underspecification When more than one answer is still possible, a feature is considered underspecified. The verb constrains sheep in these cases. The sheep bleats. The sheep graze. The number feature is considered underspecified. Types can also be underspecified, in a typed system. 7/8/2019

Typed Feature Structures This feature structure is underspecified for number. This feature structure has number specified. 7/8/2019

Implementing Feature-based Linguistic Theories Linguists can use underspecification to great advantage. The benefit is generality. Sometimes, an underspecified portion of a feature structure is shared. Our goal in implementation is to maintain the elegance of the original theories, and parse with them. 7/8/2019

Specific Problems for NLP Increased local ambiguity if lexicon contains a priori an entry for every possibility. e.g. extracted complements for a sentence that is not gapped; every combination of binder + bound arguments. Process may not terminate if arguments are underspecified. e.g. applying a lexical rule over an uninstantiated list of complements Constraints may be prematurely satisfied. Vacuously true in absence of contradiction 7/8/2019

Solution: Guarded Constraints Wait until we have more information, then process the underspecified part of the feature structure. For example, describe what we want to know from the verb, in the case of the sheep bleats, before processing sheep Waiting is also called delaying. Constraints that are delayed are called guarded constraints. 7/8/2019

Thesis Statement Delays for processing a linguistic theory can be generalized with a specification for guarded constraints as descriptions of typed feature structures. My work enables the successful parsing of modern grammars as written, even with highly lexicalized constraints. 7/8/2019

Contributions of the Thesis Guarded descriptions added to the theory of typed feature structures. Extension to theory of Carpenter (1992) Demonstrate theory in a logic programming system. Extended ALE system of Carpenter & Penn (1994) Identification of various linguistic phenomena which can be implemented using guarding. 7/8/2019

Linguistic Theories Linguistic Theories Introduction and Thesis Statement Linguistic Theories Guarded Descriptions and Guarded Rules Implementation and Evaluation Conclusions 7/8/2019

Cross-language Phenomena as a Basis for Guarding Argument Raising in German, French and Italian (Johnson 1986, Hinrichs & Nakazawa 1989, Monachesi 1993, Abeille & Godard 1994, Baker 1999) Romance Clitics (Miller 1992, Monachesi 1993, Sag & Godard 1993, Abeille et al. 1995) Morphology German Passive (Kathol 1994, Pollard 1994) Causative Chicheŵa (Alsina 1992) Japanese (Manning et al. 1999) 7/8/2019

Cross-linguistic Phenomena Complement Extraction (Pollard & Sag 1994) Quantifier Raising (Pollard & Yoo 1997) Ordering Constraints (Erbach et al. 1995, Penn 1999) 7/8/2019

A Common Thread: Argument Sharing Two heads have a common argument or arguments. Causative: causee linked to both causative and embedded predicate Verb and auxiliary share syntactic arguments in argument raising Quantifier store of verb phrase derived from quantifiers of its arguments During processing, one head might need to gain information from the other. 7/8/2019

Example: Argument Raising by Auxiliary Wird Sandy Kim sehen? Will Sandy Kim see “Will Sandy see Kim?” Arguments of the verb sehen are the subject Sandy and the object Kim. Arguments of the auxiliary verb wird are Sandy, Kim, and the verb sehen. 7/8/2019

An auxiliary’s arguments are underspecified Wird Sandy Kim sehen? 7/8/2019

Lexicalization of Constraints Concurrent trend towards lexicalization of constraints, such as binding theory, quantifier scoping, and complement extraction. e.g. for a verb, “Apply a binding constraint to my argument structure, however that is filled in.” 7/8/2019

Argument Raising + Complement Extraction Kim wird Sandy sehen. Kim will Sandy see “Sandy will see Kim.” The object Kim has been fronted (extracted). A lexical rule for complement extraction says, “If I have a lexical entry with an object on the complements list of a verb, then I also have an entry with that object in the fronted position.” 7/8/2019

Lexically specified Complement Extraction (topicalization, or fronting) (Pollard & Sag, 1994) Kim wird Sandy sehen. 7/8/2019

Raising by Auxiliary + Complement Extraction = Processing Problems We know nothing about the verb. Thus, it is difficult to extract arguments from the auxiliary’s complements list. If we know the verb is sehen, then we know there is a direct object. Solution: Wait until we know what the verbs arguments are to process this lexically specified constraint. 7/8/2019

Guarded Descriptions and Guarded Rules Introduction and Thesis Statement Linguistic Theories Guarded Descriptions and Guarded Rules Implementation and Evaluation Conclusions 7/8/2019

Feature Structure Descriptions “third person singular” The description is: I used feature structure descriptions to guard a portion of a feature structure. 7/8/2019

Typed Feature Structure Descriptions Descriptions are a shorthand for picking out a particular feature structure. (Rounds & Kasper 1986). Descriptions are relative to a particular type. 7/8/2019

Guarded Descriptions: Extending the Description Language If the feature structure satisfies φ, then add the information in the description ψ. Else, if it doesn’t, add the description υ. Example: verb are NUMBER:plural  true; PERSON:second 7/8/2019

Satisfaction Conditions for Guarded Descriptions If a feature structure satisfies a description φ, then it also satisfies ψ; else it satisfies υ. 7/8/2019

Negation Negation is defined as failure if every extension of a feature structure satisfies a description, else true. Definition follows Moshier (1988). 7/8/2019

Inequations An instance of negation. Compile the description that contains an inequation as in instance of guarding (transparent to the user). 7/8/2019

Example: Binding Theory For non-reflexives, if the nominal objects are co-indexed, then fail, else true. Johni likes himselfi *Johni likes Johni. Johni likes Johnj. (Subj,loc:cont:((npro;ppro),index:(SubjInd,=\=Ind)))| bind(Subj,(Arg,loc:cont:index:Ind)) (=\=Ind) compiled as SubjInd==Indfail;true 7/8/2019

Guarded Rules (Constraint Logic Programming -- CLP) A constraint can be associated with a rule. This affects the order of goal resolution (constraint must be satisfied first). I will use descriptions of feature structures as the guards on rules which have feature structure arguments. 7/8/2019

Guarded Rules Equivalent to using guarded descriptions. With negation defined, we can guard rules using the regular description language. 7/8/2019

Argument Raising with Guards Kim wird Sandy sehen. Kim will Sandy see. “Kim will see Sandy.” Put a guard on argument raising. Each member of the verb’s complement’s list must satisfy at least the type subst (substantive) before the arguments can be raised by auxiliary. A substantive is noun, verb, etc. 7/8/2019

Argument Raising (cont.) Guard on the COMPS list of the subcategorized verb. Comps:loc:cat:head:subst | aux_raising(Aux,Comps,Subj) 7/8/2019

Add Complement Extraction Sandy sehen wird Kim. Sandy see will Kim “Kim will see Sandy.” Know the semantic relation of the embedded verb before arguments can be extracted from an auxiliary. Prop:nucleus:relation | lexical_rule(aux(comps:PVPComps)  aux(slash:(PVP,cont:Prop),comps:PVPComps) 7/8/2019

Suspending Goals 7/8/2019

Implementation and Evaluation Introduction and Thesis Statement Linguistic Theories Guarded Descriptions and Guarded Rules Implementation and Evaluation Conclusions 7/8/2019

Logic Grammar Systems Early systems ALE (Carpenter & Penn 1994) FUG (Kay 1983) PATR-II (Shieber et al. 1985) LOGIN (Ait-Kaci and Nasr 1986) ALE (Carpenter & Penn 1994) ALEP (Erbach et al. 1995) CUF (Dorre & Dorna 1993) ConTroll (Gotz et al. 1997) FUF (Elhadad & Robin 1992) 7/8/2019

Implementation Extension to ALE (Carpenter & Penn 1994) Grammars: Allow guards on the type of a typed feature structure. Grammars: Japanese causative and binding theory (Manning et al. 1999). Inequations are compiled as waits on disproving isomorphism. German partial verb phrase fronting (Baker 1999) Lexical quantifier scoping (Pollard & Yoo 1997) 7/8/2019

Other Possible Strategies Brute Force or naïve approach (multiply out possibilities, esp. in the lexicon) Accommodation (setting limits not in the original theory; edit program or grammar) e.g. limit the number of a verb’s complements. Hand-threading the grammar e.g. put lexical constraints in the phrase-structure rules. 7/8/2019

Results of Grammar Comparison Guarding may reduce lexicon size substantially, especially with respect to the brute force approach. Accommodation strategy may prune solutions. Hand threaded grammar performs fastest, but is knowledge-intensive. Other: Slowness of guarding due to blocking and unblocking of guards (up to 10x slower than the “ideal” case). Implementation could be improved. 7/8/2019

Raising by Auxiliary+PVP Fronting (parse times, in msec) Sentence Naïve Unguarded Guarded Thread Sandy sieht Kim. Sandy wird Kim sehen können. Sandy sehen wird Kim. Sandy wird Kim sehen. Sandy sehen können wird Kim. Wird Kim gehen können? Wird Kim sehen können? Wird Kim gehen? wird Kim können (VP) Wird Kim Sandy sehen? Wird Kim Sandy sehen können? 70 5590 1070 1240 3450 2620 4260 540 1860 1220 5850 60 4990 470 810 2010 2120 3990 1460 940 1230 8240 3620 350 710 1110 2210 3270 430 580 1570 8900 840 200 280 390 330 610 120 160 320 880 7/8/2019

Advantages of the Approach Theoretical relevance Faithful rendering of both the linguistic theory and the theory I develop for guarded feature structures. Enables complex theories to be implemented as written. Algorithm is parser-independent In a study of quantifier raising, the guards are unblocked in exactly the same order, with three different parsers. 7/8/2019

Conclusions Conclusions Introduction and Thesis Statement Linguistic Theories Guarded Descriptions and Guarded Rules Implementation and Evaluation Conclusions 7/8/2019

Conclusions First generalization of delays over feature structures. Flexible; guards can be applied to lexical rules, grammar rules, type constraints, etc. Guarding is relevant for a wide range of linguistic phenomena and also across languages. 7/8/2019

Future Work Speedier implementation so that guards are not unblocked every time the type variable changes. Transfer the approach to non-typed formalisms (e.g. LFG). 7/8/2019

Extra slides begin here 7/8/2019

Example (CLP) A program to append two lists together will go into an unbounded search if the first or third argument is a variable. Wait until either of these is nonvariable. (nonvar(Xs);nonvar(Zs)) | append([X|Xs],Ys,[X|Zs])  append(Xs,Ys,Zs) 7/8/2019

Guarding on Typed Feature Structures In implementing delays, we put a guard on the type of the information we are looking for. We never delay any goals that could have been solved with a less specific type. In asking whether a feature structure F satisfies a feature-value pair Feat:Desc, a wait is placed on the type of F if it does not already satisfy a type for which Feat is appropriate. Recursively, a wait is placed on the feature structure rooted at Feat with respect to Desc. 7/8/2019