74 419 Artificial Intelligence 200506 From Syntax to

74. 419 Artificial Intelligence 2005/06 From Syntax to Semantics

From Syntax to Semantics § Grammatical Extensions § Sentence Structures § Noun Phrase - Modifications § Verb Phrase - Subcategorization § Feature Structures § -expressions

Grammar – Sentence Level Constructs n declarative S NP VP “This flight leaves at 9 am. ” n imperative S VP “Book this flight for me. ” n yes-no-question S Aux NP VP “Does this flight leave at 9 am? ” n wh-question S Wh-NP Aux NP VP “When does this flight leave Winnipeg? ”

Grammar – Noun Phrase Modification 1 head = the central noun of the NP (+ modifiers) n modifiers before the head noun (prenominal) n determiner the, a, this, some, . . . n predeterminer all the flights n cardinal numbers, ordinal numbers one flight, the first flight, . . . n quantifiers much, little n adjectives a first-class flight, a long flight n adjective phrase the least expensive flight NP (Det) (Card) (Ord) (Quant) (AP) Nominal

Grammar – Noun Phrase Modification 2 n modifiers after the head noun (post-nominal) n prepositional phrase PP all flights from Chicago Nominal PP (PP) n non-finite clause, gerundive postmodifers all flights arriving after 7 pm Nominal Gerund. VP Gerund. V NP | Gerund. V PP |. . . n relative clause a flight that serves breakfast Nominal Rel. Clause (who | that) VP

Grammar – Verb Subcategorization VP = Verb + other constituents. Different verbs accept or need different constituents → Verb Subcategorization; captured in verb frames. n sentential complement VP Verb inf-sentence I want to fly from Boston to Chicago. n NP complement VP Verb NP I want this flight. n no complement VP Verb I sleep. n more forms VP Verb PP PP I fly from Boston to Chicago.

Grammar – Feature Structures 1 Feature Structures n describe additional syntactic-semantic information, like category, person, number, e. g. goes <verb, 3 rd, singular> n specify feature structure constraints (agreements) as part of the grammar rules n during parsing, check agreements of feature structures (unification) e. g. or S NP VP <NP number>=<VP number> S NP VP <NP agreement>=<VP agreement>

Grammar – Feature Structures 2 Sub-categories specify attached phrases, e. g. NP modifiers or Verb complements like NP “. . . the man who chased the cat out of the house. . . ” central noun + sub-categories + agreements “. . . the man chased the barking dog who bit him. . . ” central verb + sub-categories + agreements Agreements are passed on / inherited within phrases, e. g. agreement of VP derived from Head-Verb of VP, through special Unification functions <VP agreement> determined by <Verb agreement> <NP agreement> determined by <Nom agreement>

Semantics Distinguish between n n surface structure (syntactic structure) and deep structure (semantic structure) of sentences. Different forms of Semantic Representation n logic based n ontology based / semantic language / interlingua n Case Frame structures n DL and similar KR languages n linguistics based Ontologies

Semantics - Lambda Calculus 1 Logic representations often involve Lambda-Calculus: n represent central phrases (verb) as -expressions n -expression is like a function, which can be applied to terms n insert semantic representation of complement or modifier phrases etc. in place of variables x, y: loves (x, y) FOPL sentence x y loves (x, y) -expression, function x y loves (x, y) (John) y loves (John, y)

Semantics - Lambda Calculus 2 Transform sentence into lambda-expression: “AI Caramba is close to ICSI. ” specific: close-to (AI Caramba, ICSI) general: x, y: close-to (x, y) x=AI Caramba y=ICSI Lambda Conversion: x y: close-to (x, y) (AI Caramba) Lambda Reduction: y: close-to (AI Caramba, y) close-to (AI Caramba, ICSI)

Semantics - Lambda Calculus 3 Lambda Expressions can be constructed from central (VP) expression, inserting semantic representations for complement (NP, PP) phrases: event subject-NP object-NP Verb serves { x y e IS-A (e, Serving) Server (e, y) Served (e, x)} represents general semantics for the verb 'serve Fill in appropriate expressions for x, y, for example 'meat' for y derived from Noun in NP as complement to Verb.

Inter. Lingua (IL) approach n n n An Ontology, a language-independent classification of objects, event, relations A Semantic Lexicon, which connects lexical items to nodes (concepts) in the ontology An analyzer that constructs IL representations and selects (an? ) appropriate one

Deriving basic semantic dependency (a toy example) Input: John makes tools Syntactic Analysis: cat verb tense present subject root john cat noun-proper object root tool cat noun number plural

Relevant parts of the (appropriate senses of the) lexicon entries for John and tool John-n 1 syn-struc root john cat noun-proper sem-struc human name gender male tool-n 1 syn-struc root cat sem-struc tool n tool john

Semantics Semantic Representation through: § Case Frame structures § DL and similar KR languages § linguistics based Ontologies General: Map surface structure to semantic structure n Derive phrases as sub-structures n Find concepts for central phrases (VP, NP) n Assign phrases to appropriate roles around central concepts.

Additional References Jurafsky, D. & J. H. Martin, Speech and Language Processing, Prentice-Hall, 2000. (Chapters 9 and 10)