1CMSC 723: Intro toComputational LinguisticsNovember 24, 2004Lecture 12: Lexical SemanticsBonnie Dorr Christof MonzMeaning So far, we have focused on the structure oflanguage, not on what things mean We have seen that words have differentmeaning, depending on the context in whichthey are used Every day language tasks that require somesemantic processing: Answering an essay question on an exam Deciding what to order at a restaurant by reading amenu Realizing you’ve been insultedMeaning (continued) meaning representations arerepresentations that link linguistic formsto knowledge of the world We are going to cover: What is the meaning of a word How can we represent the meaning What formalisms can be used• Meaning representation languagesWhat Can Serve as a MeaningRepresentation? Anything that serves the core practicalpurposes of a program that is doingsemantic processing What is a Meaning RepresentationLanguage? What is Semantic Analysis?2Requirements for MeaningRepresentation Verifiability Unambiguous Representation Canonical Form Inference ExpressivenessVerifiability System can match input representationagainst representations in knowledgebase. If it finds a match, it can returnYes; Otherwise No. Does Maharani serve vegetarian food?Serves(Maharani,vegetarian food)Unambiguous Representation Single linguistic input can have differentmeaning representations Each representation unambiguouslycharacterizes one meaning. Example: small cars and motorcycles areallowed car(x) & small(x) & motorcycle(y) & small(y) &allowed(x) & allowed(y) car(x) & small(x) & motorcycle(y) & allowed(x) &allowed(y)Ambiguity and Vagueness An expression is ambiguous if, in a givencontext, it can be disambiguated to have aspecific meaning, from a number of discrete,possible meanings. E.g., bank (financialinstitution) vs bank (river bank) An expression is vague, if it refers to a range ofa scalar variable, such that, even in a specificcontext, it’s hard to specify the range entirely.E.g., he’s tall, it’s warm, etc.3Representing Similar Concepts Distinct inputs could have the same meaning Does Maharani have vegetarian dishes? Do they have vegetarian food at Maharani? Are vegetarian dishes served at Maharani? Does Maharani serve vegetarian fare? Alternatives Four different semantic representations Store all possible meaning representations in KBCanonical Form Solution: Inputs that mean same thinghave same meaning representation Is this easy? No! Vegetarian dishes, vegetarian food,vegetarian fare Have, serve What to do?How to Produce aCanonical Form Systematic Meaning Representations can bederived from thesaurus food ___ dish ___|____one overlapping meaning sense fare ___| We can systematically relate syntacticconstructions [S [NP Maharani] serves [NP vegetariandishes]] [S [NP vegetarian dishes] are served at [NPMaharani]]Inference Consider a more complex request Can vegetarians eat at Maharani? Vs: Does Maharani serve vegetarian food? Why do these result in the same answer? Inference: Draw conclusions about truthof propositions not explicitly stored in KB serve(Maharani,VegetarianFood) =>CanEat(Vegetarians,AtMaharani)4Non-Yes/No Questions Example: I'd like to find a restaurantwhere I can get vegetarian food.serve(x,VegetarianFood) Matching succeeds only if variable xcan be replaced by known object inKB.Meaning Structure of Language Human Languages Display a basic predicate-argument structure Make use of variables Make use of quantifiers Display a partially compositional semanticsCompositionality The compositionality principle is animportant principle in formal semantics: The meaning of an expression is a strictfunction of the meanings of its parts It allows to build meaningrepresentations incrementally Standard predicate logic does not adhereto this principle (donkey sentences)Predicate-Argument Structure Represent concepts and relationships among them Some words act like arguments and some words actlike predicates: Nouns as concepts or arguments: red(ball) Adj, Adv, Verbs as predicates: red(ball) Subcategorization (argument) frames specify number,position, and syntactic category of arguments Examples: NP give NP2 NP1 NP give NP1 to NP2 give(x,y,z)5Semantic (thematic) Roles Semantic Roles: Participants in an event Agent: George hit Bill. Bill was hit by George Patient: George hit Bill. Bill was hit by George Semantic (Selectional) Restrictions: Constrain thetypes of arguments verbs take George assassinated the senator *The spider assassinated the fly Verb subcategorization: Allows linking arguments insurface structure with their semantic roles Prepositions are like verbs Under(ItalianRestaurant,$15)First Order Predicate Calculus(FOPC) FOPC provides sound computationalbasis for verifiability, inference,expressiveness Supports determination of truth Supports compositionality of meaning Supports question-answering (via variables) Supports inferenceFOPC Syntax Terms Constants: Maharani Functions: LocationOf(Maharani) Variables: x in LocationOf(x) Predicates: Relations that hold among objects Serves(Maharani,VegetarianFood) Logical Connectives: Permit compositionality ofmeaning I only have $5 and I don’t have a lot of time Have(I,$5) Have(I,LotofTime)! "¬FOPC Semantics Sentences in FOPC can be assignedtruth values True or False6Variables and Quantifiers Existential (∃): There exists A restaurant that serves Mexican food near UMD(∃x) Restaurant(x) Serves(x,MexicalFood)Near(LocationOf(x),LocationOf(UMD)) Universal (∀): For all All vegetarian restaurants serve vegetarian food(∀x) VegetarianRestaurant(x) ->Serves(x,VegetarianFood)! "! "FOPC Examples John gave Mary a book Previously: Give(John,Mary,book) Better:(∃x) Giving(x) Giver(John,x) Givee(Mary,x) Given(book,x) Full Definition of Give:(∃w,x,y,z) Giving(x) Giver(w,x) Givee(z,x) Given(y,x)! "! "! "! "! "! "Why use Variables? Multiple sentences containing “eat” I ate. I ate a turkey sandwich. I ate a turkey sandwich at my desk. I ate at my desk. I ate lunch. I ate a turkey sandwich for lunch I ate a turkey sandwich for lunch at my desk. Seven different
View Full Document