SlideShare ist ein Scribd-Unternehmen logo
1 von 23
Downloaden Sie, um offline zu lesen
Glue Semantics for Proof Theorists

              Valeria de Paiva

       Nuance Communications, CA, USA


   Abstract Proof Theory – April, 2013
Introduction
                        Glue Semantics
                         Glue in Action


Introduction

      This talk is about the application of proof theoretic methods
      to the semantics of natural languages like English.
      Proof Theory had its beginnings as the poor cousin of Model
      Theory in Mathematical Logic.
      But it got a big boost from its use in Computer Science.
      Proof theory has applications in the design and specification
      of programming languages (type theories, compilers), in the
      foundations of security and as well as being essential to
      Artificial Intelligence and Automated Deduction.
      Proof Theory also has extensive applications in Computational
      Semantics of natural language. Here we concentrate on one
      application to the syntax-semantics interface: Glue
      Semantics
                                                                      2 / 23
Introduction
                          Glue Semantics
                           Glue in Action


Glue semantics?
   Glue semantics is a theory of the syntax-semantics interface of
   natural language that uses linear logic for meaning composition.
   Distinguish two separate logics in semantic interpretation
      1. Meaning logic: target logical representation
      2. Glue logic: logical specification of how chunks of meaning are
   assembled

   In principle, Glue uses any of several alternative grammar
   formalisms and any of the mainstream semantics.
   In practice, Glue started for LFG, with a vanilla Montague-style
   logic for meanings.
   Glue analyses have been proposed within HPSG, Context-free
   grammar, Categorial grammar, and TAG.
   Meaning languages in glue analyses include Discourse
   Representation Theory, First-order logic, and Natural Semantic
   Metalanguage(NSM).                                                    3 / 23
Introduction
                         Glue Semantics
                          Glue in Action


Linear Implication and (Multiplicative) Conjunction


   To assemble meanings we use (intuitionistic) multiplicative Linear
   Logic.

   Traditional implication: A, A → B           B
                            A, A → B           A∧B   Re-use A
   Linear implication:     A, A −◦ B           B
                           A, A −◦ B           A⊗B   Cannot re-use A

   Traditional conjunction: A ∧ B          A         Discard B
   Linear conjunction:      A⊗B            A         Cannot discard B




                                                                        4 / 23
Introduction
                          Glue Semantics
                           Glue in Action


The Linguistic Appeals of Linear Logic

   Resource usage: appealing idea for thinking about linguistic issues.

      1. How a string of words provides a sequence of resources that
   can be consumed to construct a syntactic analysis of a sentence.
   Lambek Calculus++
      2. How word meanings provide a collection of resources that
   can be used to construct the meaning of a sentence. (example)
      3. How linguistic context can make certain resources available,
   such as possible pronoun antecedents, that can be used to flesh
   out the interpretations of he, she or it.
   Only dealing with 2 above.
   To begin with it looks like the proof semantics we’re used to.


                                                                          5 / 23
Introduction
           Glue Semantics
            Glue in Action


Example:




                             6 / 23
Introduction
                        Glue Semantics
                         Glue in Action


Linguistic applications of linear logic

       Categorial and type-logical grammar (Moortgat,
       vanBenthem): including parsing categorial grammars (Morrill,
       Hepple) and compositional semantics of categorial grammars
       (Morrill, Carpenter)
       Resource-based reformulations of other grammatical theories
       Minimalism (Retore,Stabler)
       Lexical Functional Grammar (Saraswat,Muskens)
       Tree Adjoining Grammar (Abrusci)
       AI issues such as the frame problem (White) or planning
       (Dixon) with linguistic relevance
       ‘Glue semantics’ (a version of categorial semantics without an
       associated categorial grammar?) (Dalrymple, Lamping &
       Gupta))
                                                                        7 / 23
Introduction
                           Glue Semantics
                            Glue in Action


Identity Criteria for Proofs



   Two proofs of A, A → B         B:
                                     [A]1         A→B
                                                          →E
     A→B          A                           B
                      →E                           →, 1
           B                                 A→B               A
                                                                   →E
                                                          B
   These are not really distinct proofs:




                                                                        8 / 23
Introduction
                            Glue Semantics
                             Glue in Action


Lambda-Equivalence of Proof Terms


   Include proof terms in previous derivations:


                                              [x : A]1          f :A→B
                                                                            →E
    f :A→B            a:A                                f (x ) : B
                            →E                                          → I, 1
          f (a) : B                               λx .f (x ) : A → B                    a:A
                                                                 (λx .f (x ))(a) : B

   Note: f (a) = (λx .f (x ))(a)
   λ-equivalence of proof terms: semantic identity of derivations.



                                                                                       9 / 23
Introduction
                            Glue Semantics
                             Glue in Action


Curry-Howard Isomorphism (CHI)



   CHI = Pairing of proof rules with operations on proof terms
      But doesn’t work for all logics, or proof systems
   Defines interesting identity criteria for proofs
      Syntactically distinct derivations corresponding to same proof
   Intimate relation between logic and type-theory.
   Varied applications, e.g.
      — Proofs as programs
      — Semantic construction for natural language




                                                                       10 / 23
Introduction
                 Glue Semantics
                  Glue in Action


Example: Using LFG Grammar




                                   11 / 23
Introduction
                     Glue Semantics
                      Glue in Action


Cutting and Pasting 1...




                                       12 / 23
Introduction
                             Glue Semantics
                              Glue in Action


Example: Input to Semantic Interpretation



   Lexicon
    Word Meaning                      Glue
    John john                         ↑        where ↑= g
    Fred     fred                     ↑        where ↑= h
    saw      λy .λx . see(x , y )     ↑ .OBJ −◦ (↑ .SUBJ −◦ ↑)
                                           where ↑= f , f .OBJ = h, f .SUBJ = g
   Constituents g, h, f : semantic resources, consuming & producing
   meanings




                                                                             13 / 23
Introduction
                           Glue Semantics
                            Glue in Action


Lexical Premises: Their nature

                                        saw

               λy .λx . see(x , y )          :     h −◦ (g −◦ f )

                Meaning Term                       Glue Formula
                                                 (Propositional LL)

   Atomic propositions (f , g, h):
   • Correspond to syntactic constituents found in parsing
   • Denote resources used in semantic interpretation
     (Match production & consumption of constituent meanings)
   Meaning terms:
   • Expressions in some chosen meaning language
   • Language must support abstraction and application
   • . . . but otherwise relatively free choice
                                                                      14 / 23
Introduction
                         Glue Semantics
                          Glue in Action


The Form of Glue Derivations



                                  Γ        M:f
   where
   • Γ is set of lexical premises (instantiated by parse)
   • f is (LL atom corresponding to) sentential constituent
   • M is meaning term produced by derivation
   (Semantic) Ambiguity
       Often (many) alternative derivations Γ Mi : f
       each producing a different meaning term Mi for f
       Need to find all alternative derivations (efficiently!)



                                                              15 / 23
Introduction
                            Glue Semantics
                             Glue in Action


Alternative Derivations: Modifier Scope



   Consider phrase “alleged criminal from London”
    λx . criminal(x )            : f
    λP. alleged(P)               : f −◦ f
    λPλx . from(lon, x ) ∧ P(x ) : f −◦ f
   There are two normal derivations, resulting in:
   1. λx . from(lon, x ) ∧ alleged(criminal)(x ) : f
   2. alleged(λx . from(lon, x ) ∧ criminal(x )) : f




                                                       16 / 23
Introduction
                   Glue Semantics
                    Glue in Action


Two normal derivations




                                     17 / 23
Introduction
                                Glue Semantics
                                 Glue in Action


Skeleton-Modifier Derivations

      Modifier: any formula equivalent to φ −◦ φ
      Initial derivation separating modifiers from skeleton
                                                                g    g −◦ h −◦ f
                                                        h               h −◦ f        skeleton

                                                                    f


        g −◦ h −◦ f
        a −◦ (f −◦ f )
                                         ⇒                  a   a −◦ (f −◦ f )
        b −◦ (h −◦ f ) −◦ (h −◦ f )                                                   modifier
        g, h, a, b                                                  f −◦ f



                                                  b   b −◦ ((h −◦ f ) −◦ (h −◦ f ))
                                                                                      modifier
                                                       (h −◦ f ) −◦ (h −◦ f )

      Final derivation inserts modifiers
        — All scope ambiguities due to modifier insertion


                                                                                                 18 / 23
Introduction
                                            Glue Semantics
                                             Glue in Action


Quantifier Scope: Everyone saw something

                               everyone:                (g −◦ f ) −◦ f
   Premises                    saw:                     h −◦ (g −◦ f )
                               something:               (h −◦ f ) −◦ f
   Derivations:
                               ∃∀                                                            ∀∃
                                                               h −◦ (g −◦ f )    [h]

    h −◦ (g −◦ f )   [h]                                            g −◦ f             [g]

         g −◦ f            (g −◦ f ) −◦ f                                    f

                           f                                             h −◦ f              (h −◦ f ) −◦ f

                     h −◦ f                  (h −◦ f ) −◦ f                                  f

                                       f                                                g −◦ f                (g −◦ f ) −◦ f

                                                                                                          f




                                                                                                                       19 / 23
Introduction
                               Glue Semantics
                                Glue in Action


With Meaning Terms



  saw : h −◦ (g −◦ f )      [y : h]
        saw (y ) : g −◦ f             [x : g]
                saw (y )(x ) : f
           λy .saw (y )(x ) : h −◦ f             everyone : (h −◦ f ) −◦ f
                            everyone(λy .saw (y )(x )) : f
                    λx .everyone(λy .saw (y )(x )) : g −◦ f                  something : (g −◦
                                        something(λx .everyone(λy .saw (y )(x ))) : f



                                                                                        20 / 23
Introduction
                        Glue Semantics
                         Glue in Action


Glue Sales Pitch


      Linguistically powerful & flexible approach
      Interesting analyses of scope, control (Asudeh), event-based
      semantics (Fry), intensional verbs (Dalrymple), context
      dependence, coordination.
      But many other phenomena still to do
      Grammar & semantics engineering
      Applicable to grammars besides LFG based ones
      Steep learning curve for writing lexical entries
      But turns out to allow plentiful re-use of “lingware”
      Can be implemented efficiently: Lev, also in NLTK open
      source github


                                                                     21 / 23
Introduction
                        Glue Semantics
                         Glue in Action


Conclusions


      For linguists: lots of language engineering to do, on a
      principled basis.
      For proof theorists: for this application cuts-with-axioms are
      not a negligible cut, they are the most important cuts ever.
      Counting how many there are and which derivations/proofs
      they give rise to, is solving the ambiguity of language problem!
      but you need a good grammar module..
      also the application sits "in-between" the proof-search and the
      proof-normalization paradigms...



                                                                         22 / 23
Introduction
                       Glue Semantics
                        Glue in Action


References



      PhD thesis of Asudeh and Lev (Stanford) and Kokkonidis
      (Oxford)
      Crouch and van Genabith (Linear Logic for Linguists)
      Online Bibliography
      http://users.ox.ac.uk/ lina1301/GlueBibliography.htm
      plus Google code
      http://nltk.googlecode.com/svn/trunk/doc/contrib/sem/gluesem.pd
      https://github.com/nltk/nltk/blob/master/examples/grammars/




                                                               23 / 23

Weitere ähnliche Inhalte

Was ist angesagt?

Prep语法笔记 未破解版
Prep语法笔记 未破解版Prep语法笔记 未破解版
Prep语法笔记 未破解版
pwyw000
 
Math63032modal
Math63032modalMath63032modal
Math63032modal
Hanibei
 
Jarrar: Introduction to logic and Logic Agents
Jarrar: Introduction to logic and Logic Agents Jarrar: Introduction to logic and Logic Agents
Jarrar: Introduction to logic and Logic Agents
Mustafa Jarrar
 
ACL読み会2014@PFI "Less Grammar, More Features"
ACL読み会2014@PFI "Less Grammar, More Features"ACL読み会2014@PFI "Less Grammar, More Features"
ACL読み会2014@PFI "Less Grammar, More Features"
nozyh
 

Was ist angesagt? (15)

Prep语法笔记 未破解版
Prep语法笔记 未破解版Prep语法笔记 未破解版
Prep语法笔记 未破解版
 
Jarrar: First Order Logic
Jarrar: First Order LogicJarrar: First Order Logic
Jarrar: First Order Logic
 
AI Lesson 12
AI Lesson 12AI Lesson 12
AI Lesson 12
 
Math63032modal
Math63032modalMath63032modal
Math63032modal
 
Nsu module 01-logic-final
Nsu module 01-logic-finalNsu module 01-logic-final
Nsu module 01-logic-final
 
Gwc 2016
Gwc 2016Gwc 2016
Gwc 2016
 
Jarrar: Description Logic
Jarrar: Description LogicJarrar: Description Logic
Jarrar: Description Logic
 
Poster - Bigorna, a toolkit for orthography migration challenges
Poster - Bigorna, a toolkit for orthography migration challengesPoster - Bigorna, a toolkit for orthography migration challenges
Poster - Bigorna, a toolkit for orthography migration challenges
 
AI Lesson 09
AI Lesson 09AI Lesson 09
AI Lesson 09
 
Integration in Finite Terms
Integration in Finite TermsIntegration in Finite Terms
Integration in Finite Terms
 
AI Lesson 13
AI Lesson 13AI Lesson 13
AI Lesson 13
 
Jarrar: Introduction to logic and Logic Agents
Jarrar: Introduction to logic and Logic Agents Jarrar: Introduction to logic and Logic Agents
Jarrar: Introduction to logic and Logic Agents
 
AI Lesson 16
AI Lesson 16AI Lesson 16
AI Lesson 16
 
ACL読み会2014@PFI "Less Grammar, More Features"
ACL読み会2014@PFI "Less Grammar, More Features"ACL読み会2014@PFI "Less Grammar, More Features"
ACL読み会2014@PFI "Less Grammar, More Features"
 
AI Lesson 15
AI Lesson 15AI Lesson 15
AI Lesson 15
 

Andere mochten auch

Andere mochten auch (18)

CLiCS: Categorical Logic in Computer Science
CLiCS: Categorical Logic in Computer ScienceCLiCS: Categorical Logic in Computer Science
CLiCS: Categorical Logic in Computer Science
 
To Infinite Possibilities and Beyond...
To Infinite Possibilities and Beyond...To Infinite Possibilities and Beyond...
To Infinite Possibilities and Beyond...
 
Lean Logic for Lean Times: Entailment and Contradiction Revisited
Lean Logic for Lean Times: Entailment and Contradiction RevisitedLean Logic for Lean Times: Entailment and Contradiction Revisited
Lean Logic for Lean Times: Entailment and Contradiction Revisited
 
Contexts 4 quantification (CommonSense2013)
Contexts 4 quantification (CommonSense2013)Contexts 4 quantification (CommonSense2013)
Contexts 4 quantification (CommonSense2013)
 
Little engines of inference: contexts for quantification
Little engines of inference: contexts for quantificationLittle engines of inference: contexts for quantification
Little engines of inference: contexts for quantification
 
Who's afraid of Categorical models?
Who's afraid of Categorical models?Who's afraid of Categorical models?
Who's afraid of Categorical models?
 
Lexical Resources for Portuguese
Lexical Resources  for PortugueseLexical Resources  for Portuguese
Lexical Resources for Portuguese
 
Category Theory for All (NASSLLI 2012)
Category Theory for All (NASSLLI 2012)Category Theory for All (NASSLLI 2012)
Category Theory for All (NASSLLI 2012)
 
Intuitionistic Modal Logic: fifteen years later
Intuitionistic Modal Logic: fifteen years laterIntuitionistic Modal Logic: fifteen years later
Intuitionistic Modal Logic: fifteen years later
 
Constructive Modal Logics, Once Again
Constructive Modal Logics, Once AgainConstructive Modal Logics, Once Again
Constructive Modal Logics, Once Again
 
If, not when
If, not whenIf, not when
If, not when
 
Modal Type Theory
Modal Type TheoryModal Type Theory
Modal Type Theory
 
Seeing is Correcting:Linked Open Data for Portuguese
Seeing is Correcting:Linked Open Data for PortugueseSeeing is Correcting:Linked Open Data for Portuguese
Seeing is Correcting:Linked Open Data for Portuguese
 
Lean Logic for Lean Times: Varieties of Natural Logic
Lean Logic for Lean Times: Varieties of Natural LogicLean Logic for Lean Times: Varieties of Natural Logic
Lean Logic for Lean Times: Varieties of Natural Logic
 
Logics and Ontologies for Portuguese Understanding
Logics and Ontologies for Portuguese UnderstandingLogics and Ontologies for Portuguese Understanding
Logics and Ontologies for Portuguese Understanding
 
Portuguese Linguistic Tools: What, Why and How
Portuguese Linguistic Tools: What, Why and HowPortuguese Linguistic Tools: What, Why and How
Portuguese Linguistic Tools: What, Why and How
 
Smart Cities, Smart Citizens and Smart Decisions
Smart Cities, Smart Citizens and Smart DecisionsSmart Cities, Smart Citizens and Smart Decisions
Smart Cities, Smart Citizens and Smart Decisions
 
Garbage Collection en el JVM
Garbage Collection en el JVMGarbage Collection en el JVM
Garbage Collection en el JVM
 

Ähnlich wie Glue Semantics for Proof Theorists

Abstract Symbolic Automata: Mixed syntactic/semantic similarity analysis of e...
Abstract Symbolic Automata: Mixed syntactic/semantic similarity analysis of e...Abstract Symbolic Automata: Mixed syntactic/semantic similarity analysis of e...
Abstract Symbolic Automata: Mixed syntactic/semantic similarity analysis of e...
FACE
 
The LSA breaks downanalyzes what constitutes a good and bad a.docx
The LSA breaks downanalyzes what constitutes a good and bad a.docxThe LSA breaks downanalyzes what constitutes a good and bad a.docx
The LSA breaks downanalyzes what constitutes a good and bad a.docx
arnoldmeredith47041
 
[Emnlp] what is glo ve part ii - towards data science
[Emnlp] what is glo ve  part ii - towards data science[Emnlp] what is glo ve  part ii - towards data science
[Emnlp] what is glo ve part ii - towards data science
Nikhil Jaiswal
 

Ähnlich wie Glue Semantics for Proof Theorists (10)

形式言語理論への 測度論的アプローチ
形式言語理論への 測度論的アプローチ形式言語理論への 測度論的アプローチ
形式言語理論への 測度論的アプローチ
 
Constructive Adpositional Grammars, Formally
Constructive Adpositional Grammars, FormallyConstructive Adpositional Grammars, Formally
Constructive Adpositional Grammars, Formally
 
Abstract Symbolic Automata: Mixed syntactic/semantic similarity analysis of e...
Abstract Symbolic Automata: Mixed syntactic/semantic similarity analysis of e...Abstract Symbolic Automata: Mixed syntactic/semantic similarity analysis of e...
Abstract Symbolic Automata: Mixed syntactic/semantic similarity analysis of e...
 
The CLUES database: automated search for linguistic cognates
The CLUES database: automated search for linguistic cognatesThe CLUES database: automated search for linguistic cognates
The CLUES database: automated search for linguistic cognates
 
New word analogy corpus
New word analogy corpusNew word analogy corpus
New word analogy corpus
 
Clustering Based Approach Extracting Collocations
Clustering Based Approach Extracting CollocationsClustering Based Approach Extracting Collocations
Clustering Based Approach Extracting Collocations
 
Natural Language Generation from First-Order Expressions
Natural Language Generation from First-Order ExpressionsNatural Language Generation from First-Order Expressions
Natural Language Generation from First-Order Expressions
 
The LSA breaks downanalyzes what constitutes a good and bad a.docx
The LSA breaks downanalyzes what constitutes a good and bad a.docxThe LSA breaks downanalyzes what constitutes a good and bad a.docx
The LSA breaks downanalyzes what constitutes a good and bad a.docx
 
[Emnlp] what is glo ve part ii - towards data science
[Emnlp] what is glo ve  part ii - towards data science[Emnlp] what is glo ve  part ii - towards data science
[Emnlp] what is glo ve part ii - towards data science
 
Learning Word Subsumption Projections for the Russian Language
Learning Word Subsumption Projections for the Russian LanguageLearning Word Subsumption Projections for the Russian Language
Learning Word Subsumption Projections for the Russian Language
 

Mehr von Valeria de Paiva

Mehr von Valeria de Paiva (20)

Dialectica Comonoids
Dialectica ComonoidsDialectica Comonoids
Dialectica Comonoids
 
Dialectica Categorical Constructions
Dialectica Categorical ConstructionsDialectica Categorical Constructions
Dialectica Categorical Constructions
 
Logic & Representation 2021
Logic & Representation 2021Logic & Representation 2021
Logic & Representation 2021
 
Constructive Modal and Linear Logics
Constructive Modal and Linear LogicsConstructive Modal and Linear Logics
Constructive Modal and Linear Logics
 
Dialectica Categories Revisited
Dialectica Categories RevisitedDialectica Categories Revisited
Dialectica Categories Revisited
 
PLN para Tod@s
PLN para Tod@sPLN para Tod@s
PLN para Tod@s
 
Networked Mathematics: NLP tools for Better Science
Networked Mathematics: NLP tools for Better ScienceNetworked Mathematics: NLP tools for Better Science
Networked Mathematics: NLP tools for Better Science
 
Going Without: a modality and its role
Going Without: a modality and its roleGoing Without: a modality and its role
Going Without: a modality and its role
 
Problemas de Kolmogorov-Veloso
Problemas de Kolmogorov-VelosoProblemas de Kolmogorov-Veloso
Problemas de Kolmogorov-Veloso
 
Natural Language Inference: for Humans and Machines
Natural Language Inference: for Humans and MachinesNatural Language Inference: for Humans and Machines
Natural Language Inference: for Humans and Machines
 
Dialectica Petri Nets
Dialectica Petri NetsDialectica Petri Nets
Dialectica Petri Nets
 
The importance of Being Erneast: Open datasets in Portuguese
The importance of Being Erneast: Open datasets in PortugueseThe importance of Being Erneast: Open datasets in Portuguese
The importance of Being Erneast: Open datasets in Portuguese
 
Negation in the Ecumenical System
Negation in the Ecumenical SystemNegation in the Ecumenical System
Negation in the Ecumenical System
 
Constructive Modal and Linear Logics
Constructive Modal and Linear LogicsConstructive Modal and Linear Logics
Constructive Modal and Linear Logics
 
Semantics and Reasoning for NLP, AI and ACT
Semantics and Reasoning for NLP, AI and ACTSemantics and Reasoning for NLP, AI and ACT
Semantics and Reasoning for NLP, AI and ACT
 
NLCS 2013 opening slides
NLCS 2013 opening slidesNLCS 2013 opening slides
NLCS 2013 opening slides
 
Dialectica Comonads
Dialectica ComonadsDialectica Comonads
Dialectica Comonads
 
Categorical Explicit Substitutions
Categorical Explicit SubstitutionsCategorical Explicit Substitutions
Categorical Explicit Substitutions
 
Logic and Probabilistic Methods for Dialog
Logic and Probabilistic Methods for DialogLogic and Probabilistic Methods for Dialog
Logic and Probabilistic Methods for Dialog
 
Intuitive Semantics for Full Intuitionistic Linear Logic (2014)
Intuitive Semantics for Full Intuitionistic Linear Logic (2014)Intuitive Semantics for Full Intuitionistic Linear Logic (2014)
Intuitive Semantics for Full Intuitionistic Linear Logic (2014)
 

Kürzlich hochgeladen

Kürzlich hochgeladen (20)

Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024Finology Group – Insurtech Innovation Award 2024
Finology Group – Insurtech Innovation Award 2024
 
Handwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed textsHandwritten Text Recognition for manuscripts and early printed texts
Handwritten Text Recognition for manuscripts and early printed texts
 
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
Mastering MySQL Database Architecture: Deep Dive into MySQL Shell and MySQL R...
 
Evaluating the top large language models.pdf
Evaluating the top large language models.pdfEvaluating the top large language models.pdf
Evaluating the top large language models.pdf
 
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
Bajaj Allianz Life Insurance Company - Insurer Innovation Award 2024
 
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
Raspberry Pi 5: Challenges and Solutions in Bringing up an OpenGL/Vulkan Driv...
 
GenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdfGenAI Risks & Security Meetup 01052024.pdf
GenAI Risks & Security Meetup 01052024.pdf
 
Boost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdfBoost Fertility New Invention Ups Success Rates.pdf
Boost Fertility New Invention Ups Success Rates.pdf
 
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men08448380779 Call Girls In Greater Kailash - I Women Seeking Men
08448380779 Call Girls In Greater Kailash - I Women Seeking Men
 
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...Workshop - Best of Both Worlds_ Combine  KG and Vector search for  enhanced R...
Workshop - Best of Both Worlds_ Combine KG and Vector search for enhanced R...
 
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
Strategies for Unlocking Knowledge Management in Microsoft 365 in the Copilot...
 
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law DevelopmentsTrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
TrustArc Webinar - Stay Ahead of US State Data Privacy Law Developments
 
presentation ICT roal in 21st century education
presentation ICT roal in 21st century educationpresentation ICT roal in 21st century education
presentation ICT roal in 21st century education
 
The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024The 7 Things I Know About Cyber Security After 25 Years | April 2024
The 7 Things I Know About Cyber Security After 25 Years | April 2024
 
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
08448380779 Call Girls In Diplomatic Enclave Women Seeking Men
 
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdfUnderstanding Discord NSFW Servers A Guide for Responsible Users.pdf
Understanding Discord NSFW Servers A Guide for Responsible Users.pdf
 
2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...2024: Domino Containers - The Next Step. News from the Domino Container commu...
2024: Domino Containers - The Next Step. News from the Domino Container commu...
 
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemkeProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
ProductAnonymous-April2024-WinProductDiscovery-MelissaKlemke
 
Boost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivityBoost PC performance: How more available memory can improve productivity
Boost PC performance: How more available memory can improve productivity
 
🐬 The future of MySQL is Postgres 🐘
🐬  The future of MySQL is Postgres   🐘🐬  The future of MySQL is Postgres   🐘
🐬 The future of MySQL is Postgres 🐘
 

Glue Semantics for Proof Theorists

  • 1. Glue Semantics for Proof Theorists Valeria de Paiva Nuance Communications, CA, USA Abstract Proof Theory – April, 2013
  • 2. Introduction Glue Semantics Glue in Action Introduction This talk is about the application of proof theoretic methods to the semantics of natural languages like English. Proof Theory had its beginnings as the poor cousin of Model Theory in Mathematical Logic. But it got a big boost from its use in Computer Science. Proof theory has applications in the design and specification of programming languages (type theories, compilers), in the foundations of security and as well as being essential to Artificial Intelligence and Automated Deduction. Proof Theory also has extensive applications in Computational Semantics of natural language. Here we concentrate on one application to the syntax-semantics interface: Glue Semantics 2 / 23
  • 3. Introduction Glue Semantics Glue in Action Glue semantics? Glue semantics is a theory of the syntax-semantics interface of natural language that uses linear logic for meaning composition. Distinguish two separate logics in semantic interpretation 1. Meaning logic: target logical representation 2. Glue logic: logical specification of how chunks of meaning are assembled In principle, Glue uses any of several alternative grammar formalisms and any of the mainstream semantics. In practice, Glue started for LFG, with a vanilla Montague-style logic for meanings. Glue analyses have been proposed within HPSG, Context-free grammar, Categorial grammar, and TAG. Meaning languages in glue analyses include Discourse Representation Theory, First-order logic, and Natural Semantic Metalanguage(NSM). 3 / 23
  • 4. Introduction Glue Semantics Glue in Action Linear Implication and (Multiplicative) Conjunction To assemble meanings we use (intuitionistic) multiplicative Linear Logic. Traditional implication: A, A → B B A, A → B A∧B Re-use A Linear implication: A, A −◦ B B A, A −◦ B A⊗B Cannot re-use A Traditional conjunction: A ∧ B A Discard B Linear conjunction: A⊗B A Cannot discard B 4 / 23
  • 5. Introduction Glue Semantics Glue in Action The Linguistic Appeals of Linear Logic Resource usage: appealing idea for thinking about linguistic issues. 1. How a string of words provides a sequence of resources that can be consumed to construct a syntactic analysis of a sentence. Lambek Calculus++ 2. How word meanings provide a collection of resources that can be used to construct the meaning of a sentence. (example) 3. How linguistic context can make certain resources available, such as possible pronoun antecedents, that can be used to flesh out the interpretations of he, she or it. Only dealing with 2 above. To begin with it looks like the proof semantics we’re used to. 5 / 23
  • 6. Introduction Glue Semantics Glue in Action Example: 6 / 23
  • 7. Introduction Glue Semantics Glue in Action Linguistic applications of linear logic Categorial and type-logical grammar (Moortgat, vanBenthem): including parsing categorial grammars (Morrill, Hepple) and compositional semantics of categorial grammars (Morrill, Carpenter) Resource-based reformulations of other grammatical theories Minimalism (Retore,Stabler) Lexical Functional Grammar (Saraswat,Muskens) Tree Adjoining Grammar (Abrusci) AI issues such as the frame problem (White) or planning (Dixon) with linguistic relevance ‘Glue semantics’ (a version of categorial semantics without an associated categorial grammar?) (Dalrymple, Lamping & Gupta)) 7 / 23
  • 8. Introduction Glue Semantics Glue in Action Identity Criteria for Proofs Two proofs of A, A → B B: [A]1 A→B →E A→B A B →E →, 1 B A→B A →E B These are not really distinct proofs: 8 / 23
  • 9. Introduction Glue Semantics Glue in Action Lambda-Equivalence of Proof Terms Include proof terms in previous derivations: [x : A]1 f :A→B →E f :A→B a:A f (x ) : B →E → I, 1 f (a) : B λx .f (x ) : A → B a:A (λx .f (x ))(a) : B Note: f (a) = (λx .f (x ))(a) λ-equivalence of proof terms: semantic identity of derivations. 9 / 23
  • 10. Introduction Glue Semantics Glue in Action Curry-Howard Isomorphism (CHI) CHI = Pairing of proof rules with operations on proof terms But doesn’t work for all logics, or proof systems Defines interesting identity criteria for proofs Syntactically distinct derivations corresponding to same proof Intimate relation between logic and type-theory. Varied applications, e.g. — Proofs as programs — Semantic construction for natural language 10 / 23
  • 11. Introduction Glue Semantics Glue in Action Example: Using LFG Grammar 11 / 23
  • 12. Introduction Glue Semantics Glue in Action Cutting and Pasting 1... 12 / 23
  • 13. Introduction Glue Semantics Glue in Action Example: Input to Semantic Interpretation Lexicon Word Meaning Glue John john ↑ where ↑= g Fred fred ↑ where ↑= h saw λy .λx . see(x , y ) ↑ .OBJ −◦ (↑ .SUBJ −◦ ↑) where ↑= f , f .OBJ = h, f .SUBJ = g Constituents g, h, f : semantic resources, consuming & producing meanings 13 / 23
  • 14. Introduction Glue Semantics Glue in Action Lexical Premises: Their nature saw λy .λx . see(x , y ) : h −◦ (g −◦ f ) Meaning Term Glue Formula (Propositional LL) Atomic propositions (f , g, h): • Correspond to syntactic constituents found in parsing • Denote resources used in semantic interpretation (Match production & consumption of constituent meanings) Meaning terms: • Expressions in some chosen meaning language • Language must support abstraction and application • . . . but otherwise relatively free choice 14 / 23
  • 15. Introduction Glue Semantics Glue in Action The Form of Glue Derivations Γ M:f where • Γ is set of lexical premises (instantiated by parse) • f is (LL atom corresponding to) sentential constituent • M is meaning term produced by derivation (Semantic) Ambiguity Often (many) alternative derivations Γ Mi : f each producing a different meaning term Mi for f Need to find all alternative derivations (efficiently!) 15 / 23
  • 16. Introduction Glue Semantics Glue in Action Alternative Derivations: Modifier Scope Consider phrase “alleged criminal from London” λx . criminal(x ) : f λP. alleged(P) : f −◦ f λPλx . from(lon, x ) ∧ P(x ) : f −◦ f There are two normal derivations, resulting in: 1. λx . from(lon, x ) ∧ alleged(criminal)(x ) : f 2. alleged(λx . from(lon, x ) ∧ criminal(x )) : f 16 / 23
  • 17. Introduction Glue Semantics Glue in Action Two normal derivations 17 / 23
  • 18. Introduction Glue Semantics Glue in Action Skeleton-Modifier Derivations Modifier: any formula equivalent to φ −◦ φ Initial derivation separating modifiers from skeleton g g −◦ h −◦ f h h −◦ f skeleton f g −◦ h −◦ f a −◦ (f −◦ f ) ⇒ a a −◦ (f −◦ f ) b −◦ (h −◦ f ) −◦ (h −◦ f ) modifier g, h, a, b f −◦ f b b −◦ ((h −◦ f ) −◦ (h −◦ f )) modifier (h −◦ f ) −◦ (h −◦ f ) Final derivation inserts modifiers — All scope ambiguities due to modifier insertion 18 / 23
  • 19. Introduction Glue Semantics Glue in Action Quantifier Scope: Everyone saw something everyone: (g −◦ f ) −◦ f Premises saw: h −◦ (g −◦ f ) something: (h −◦ f ) −◦ f Derivations: ∃∀ ∀∃ h −◦ (g −◦ f ) [h] h −◦ (g −◦ f ) [h] g −◦ f [g] g −◦ f (g −◦ f ) −◦ f f f h −◦ f (h −◦ f ) −◦ f h −◦ f (h −◦ f ) −◦ f f f g −◦ f (g −◦ f ) −◦ f f 19 / 23
  • 20. Introduction Glue Semantics Glue in Action With Meaning Terms saw : h −◦ (g −◦ f ) [y : h] saw (y ) : g −◦ f [x : g] saw (y )(x ) : f λy .saw (y )(x ) : h −◦ f everyone : (h −◦ f ) −◦ f everyone(λy .saw (y )(x )) : f λx .everyone(λy .saw (y )(x )) : g −◦ f something : (g −◦ something(λx .everyone(λy .saw (y )(x ))) : f 20 / 23
  • 21. Introduction Glue Semantics Glue in Action Glue Sales Pitch Linguistically powerful & flexible approach Interesting analyses of scope, control (Asudeh), event-based semantics (Fry), intensional verbs (Dalrymple), context dependence, coordination. But many other phenomena still to do Grammar & semantics engineering Applicable to grammars besides LFG based ones Steep learning curve for writing lexical entries But turns out to allow plentiful re-use of “lingware” Can be implemented efficiently: Lev, also in NLTK open source github 21 / 23
  • 22. Introduction Glue Semantics Glue in Action Conclusions For linguists: lots of language engineering to do, on a principled basis. For proof theorists: for this application cuts-with-axioms are not a negligible cut, they are the most important cuts ever. Counting how many there are and which derivations/proofs they give rise to, is solving the ambiguity of language problem! but you need a good grammar module.. also the application sits "in-between" the proof-search and the proof-normalization paradigms... 22 / 23
  • 23. Introduction Glue Semantics Glue in Action References PhD thesis of Asudeh and Lev (Stanford) and Kokkonidis (Oxford) Crouch and van Genabith (Linear Logic for Linguists) Online Bibliography http://users.ox.ac.uk/ lina1301/GlueBibliography.htm plus Google code http://nltk.googlecode.com/svn/trunk/doc/contrib/sem/gluesem.pd https://github.com/nltk/nltk/blob/master/examples/grammars/ 23 / 23