lexicografie computationala feb., 2012 anca dinu university of bucharest
TRANSCRIPT
Lexicografie computationala
Feb., 2012
Anca Dinu
University of Bucharest
Intoducere
Lexicologia computationala este utilizarea calculatoarealor in studiul lexiconului (teoretic).
Lexicografia computationala inseamna crearea de machine readable dictionaries (MRD) (practic).
Se folosesc uneori ca sinonime.
Introducere
MRD sunt resurse esentiale pentru NLP (Summarization, question answering, inference, etc).
Importanta lor este si mai mare pentru limbile cu o morfologie bogata. Au o componenta generativa care construieste formele inflectionale pornind de la leme si reguli de formare.
Directii in CL
Adnotare de corpus (de obicei in XML): Markup Languages permit crearea de corpusuri adnotate standardizat din care apoi se pot extrage automat sau semi-automat (Corpus Pattern Analysis) date pentru crearea de lexicoane (structura argumentala, roluri tematice, etc.)
Directii in CL
Creare Lexical Knowledge Bases (LKBs). Contin aceleasi informatii ca un dictionar printat, avand in plus informatii sintactice, semantice si relationale (ontologii)
Ex: WordNet; FrameNet
WordNet
Nouns, verbs, adjectives and adverbs are grouped into sets of cognitive synonyms (synsets), each expressing a distinct concept. Synsets are interlinked by means of conceptual-semantic and lexical relations. The resulting network of meaningfully related words and concepts can be navigated with the browser.
Annotation: principii generale
Annotation schemata should focus on a single coherent theme:
Different linguistic phenomena should be annotated separately over the same corpus
Annotations must be consistent with each other:
Unification and merging of multiple annotation is necessary (standard XML)
Examples of Semantic Annotation
• Predicators and their named arguments:
[The man]agent painted [the wall]patient.
• Anaphors and their antecedents:
[The protein] inhibits growth in yeast. [It] blocks
production . . .
• Acronyms and their long forms:
[Platelet-derived growth factor] (known as [pdgf]) impacts . . .
• Semantic Typing of entities:
[The man]human fired [the gun]firearm.
Probleme cu LKB existente
Organizarea traditionala a lexicoanelor este statica, i.e. presupune ca intelesul unui cuvant pote fi definit exhaustiv printr-o enumerare a sensurilor (tip lista).
In consecinta, cand o problema de interpretare a limbajului natural se loveste de ambiguitate lexicala, sistemul incearca sa selecteze cea mai apropiata definitie din lista oferita de lexicon
Probleme cu LKB existente
2 dezavantaje:
Trebuie specificate a priori “toate” contextele posibile in care poate aparea un cuvant; in caz contrar, rezulta acoperire incompleta;
Nu se poate explica/prezice utilizarea creativa a cuvintelor in contexte noi
Solutie : Generative Lexicon (GL)
James Pustejovsky: 1995 (cartea Generative lexicon), 2001, 2005
De citit pt data viitoare articolul “Type Theory and Lexical Decomposition”
Language meaning is compositional. Compositionality is a desirable property of a
semantic model. Many linguistic phenomena appear non-
compositional. GL exploits richer representations and fixes
the holes in the compositionality model.
Assumptions for GL
Exemple de fapte lingvistice care par de natura non-compozitionala
• intensionality (think),
• binding (she),
• quantification (most),
• interrogatives (who),
• focus (only), and
• presuppositions (the king of France).
The meaning of a complex expression is determined by its structure and the meanings of its constituents.
Questions . . .
1. What is the nature of the structure?
2. What is the meaning of a constituent?
3. What counts as a constituent?
Compositionality
(1) a. Mary began [to eat her breakfast]. b. Mary began [eating her breakfast]. c. Mary began [her breakfast].
(2) a. Mary enjoyed her beer. b. John enjoys his coffee in the morning. c. Bill enjoyed the movie.
Challenges to Simple Compositionality
Challenges to Simple Compositionality
(3)a. The woman baked a potato in the oven. b. The woman baked a cake in the oven.
(4) a. John swept. b. John swept the floor. c. John swept the dirt into the corner. d. John swept the dirt off the sidewalk. e. John swept the floor clean. f. John swept the dirt into a pile.
Challenges to Simple Compositionality
shovel, rake, shave, weed.
(5) a. John whistled.b. John whistled at the dog.c. John whistled a tune.d. John whistled a warning.e. John whistled his appreciation.f. John whistled to the dog to come.
yell, snap, whisper.
(6) Externally Caused Events: break, etc. a. The vase broke. b. Mary broke the vase. c. The storm broke the window.
(7) Internally Caused Events (unacusatives): decay, bloom, etc.
a. The flowers bloomed early. b. *The gardener bloomed the flowers.
Challenges to Simple Compositionality
1. What is the nature of the function?
2. What does it apply to; i.e., what can be an argument?
1. John loves Mary.
2. love(Arg1,Arg2)
3. Apply love(Arg1,Arg2) to Mary
4. love(Arg1,Mary)
5. Apply love(Arg1,Mary) to John
6. love(John,Mary)
Compunere = aplicare de functii
Lambda Calcul
(a) e is a type.
(b) t is a type.
(c) If a and b are types, then a -> b is a type.
A simple type tree:
t
e e->t
Function Application: If α is of type a, and β is of
type a -> b, then β(α) is of type b.
Lambda calcul data viitoare