ACL RD-TEC 1.0 Summarization of W97-1014
Paper Title:
WORD TRIGGERS AND THE EM ALGORITHM
WORD TRIGGERS AND THE EM ALGORITHM
Authors: Christoph Tillmann and Hermann Ney
Primarily assigned technology terms:
- algorithm
- computational linguistics
- computational natural language learning
- em algorithm
- em training
- language learning
- language model training
- language modeling
- learning
- leaving-one-out technique
- likelihood estimation
- linear interpolation
- maximum likelihood
- maximum likelihood estimation
- maximum-entropy
- maximum-likelihood
- measuring
- model training
- modeling
- natural language learning
- reestimation
- text dictation
Other assigned terms:
- approach
- association for computational linguistics
- backing-off model
- baseline model
- bigram
- cache
- case
- conditional probability
- convergence
- corpora
- corpus size
- derivation
- estimation
- evaluations
- experimental results
- grammars
- interpolation
- knowledge
- language model
- likelihood
- linear combination
- linguistics
- log-likelihood
- long-distance dependencies
- measures
- method
- natural language
- paragraph
- perplexity
- probability
- relative frequency
- sentences
- symbol
- technique
- test corpus
- text
- text corpus
- training
- training corpora
- training corpus
- trigram
- trigram model
- unigram
- vocabulary
- vocabulary size
- wall street journal corpus
- word
- word pair
- words
