74 419 Artificial Intelligence 200506 Features Gaps Movement
- Slides: 20
74. 419 Artificial Intelligence 2005/06 Features, Gaps, Movement Questions and Passives
Features, Gaps, and Questions Features n Questions n Passives n Gaps n Movement n
Verb SUBCAT Feature Sample Rule: (VP) (V SUBCAT -np-vp: inf) (NP) (VP VFORM inf) Example: (VP) (want SUBCAT -np-vp: inf) (class) (to end VFORM inf)
Prep. Phrases - PFORM Feature Sample Rule: (VP) (V SUBCAT -np-pp: loc) (NP) (PP PFORM LOC) Example: (VP) (put SUBCAT -np-pp: loc) (the tire) (on the cap PFORM loc)
More Verb SUBCATs Sample Rule: (VP) (V SUBCAT -np-adjp) (NP) (ADJP) Example: (VP) (held SUBCAT -np-adjp) (the cop) (responsible)
Auxiliaries - Some Features Modal Verb form 'past' e. g. did buy person, number e. g. I, you, he, we, they did Complement Verb form 'base' e. g. did buy Verb form 'base' e. g. have paid Complement Verb form 'past participle' e. g. have seen
VP (AUX COMPFORM base) (VP VFORM base) can see have to match (unification) VP (AUX COMPFORM ? s) (VP VFORM ? s) general form VP (AUX COMPFORM ing) (VP VFORM ing) progressive is bringing VP (AUX COMPFORM pastprt) (VP VFORM pastprt) has brought perfect, with past participle
Questions
These characteristics give information about the type of the answer! Category (POS), here: PP-word (prepositional phrase) sentence / phrase: question Q, relative clause R type, here: time specification POS category, here: pronoun agreement, matches: 3 rd person singular or plural type of PP, here: location, motion spec. binary feature POSS: possessive QDET: used as determiner
Rules for NP (noun phrase) Features: POSS, WH PRO question word (possessive pronoun): who, whose (POSS +), which, what from rule 1 (POSS +): whose (book) Complement NP QDET question (determiner): which (book), what (file) PP with NP question: in which store, in whose pockets PP question: where, when inheritance of features
Parsing Question - Which dogs did he see? gap NP missing
Passives
Active and Passive - Example gap NP missing
Passives - Example The box was moved. VP[pastprt, +passgap] V[_np, pastprt] moved move is normally V[_np] np is missing, thus +passgap VP (AUX[be]) (VP[pastprt], +passgap) was moved VP[pastprt, +passgap] (AUX[be]) (VP[pastprt], +passgap) features are inherited downwards / passed upwards from main verb, VP (move, moved)
Movement
Additional References § James Allen: Natural Language Understanding. Benjamin Cummings, 1995
- Cis 419 upenn
- Cis 419 upenn
- Cis 519
- Cis 419
- Upenn cis 519
- Artificial intelligence
- Levels of language analysis
- Artificial intelligence
- 15-780 graduate artificial intelligence
- Artificial intelligence examples
- Roman yampolskiy
- Conclusion of artificial intelligence
- 15-780 graduate artificial intelligence
- Artificial intelligence
- Cs188 artificial intelligence
- Semantic network in ai
- 15-780 graduate artificial intelligence
- Rule based deduction system
- Cpsc 322: introduction to artificial intelligence
- 15-780 graduate artificial intelligence
- Artificial intelligence the next digital frontier