ACL RD-TEC 1.0 Summarization of W95-0103
Paper Title:
PREPOSITIONAL PHRASE ATTACHMENT THROUGH A BACKED-OFF MODEL
PREPOSITIONAL PHRASE ATTACHMENT THROUGH A BACKED-OFF MODEL
Authors: Michael Collins and James Brooks
Primarily assigned technology terms:
- algorithm
- greedy search
- language modeling
- likelihood estimation
- maximum entropy
- maximum entropy model
- maximum likelihood
- maximum likelihood estimation
- modeling
- morphological analyser
- morphological analysis
- n-gram language modeling
- processing
- recognition
- search
- smoothing
- speech recognition
- statistical approaches
- supervised training
- training method
- treebanking
- unsupervised method
- unsupervised training
- verb phrases
Other assigned terms:
- ambiguity
- analyser
- attachment ambiguity
- case
- conditional probabilities
- conditional probability
- corpora
- data sets
- development set
- entropy
- error rate
- estimation
- events
- frequency cut-off
- head noun
- human performance
- language models
- likelihood
- measure
- method
- n-gram
- n-grams
- nouns
- phrase
- phrase attachment
- pp-attachment
- preposition
- prepositional phrase
- prepositional phrase attachment
- prepositions
- probabilities
- probability
- semantic
- semantic class
- semantic classes
- sentence
- sentences
- sparse data
- sparse data problem
- stem
- symbol
- test data
- test set
- text
- training
- training and test data
- training data
- training material
- transformation
- treebank
- verb
- verb attachment
- verb phrase
- vocabulary
- wall street journal corpus
- word
- word-class information
- wordnet
- words
- wsj corpus
