ACL RD-TEC 1.0 Summarization of W05-0821
Paper Title:
IMPROVED LANGUAGE MODELING FOR STATISTICAL MACHINE TRANSLATION
IMPROVED LANGUAGE MODELING FOR STATISTICAL MACHINE TRANSLATION
Authors: Katrin Kirchhoff and Mei Yang
Primarily assigned technology terms:
- automatic speech recognition
- backoff technique
- boosting
- decoder
- decoding
- generalized parallel backoff
- genetic algorithms
- giza
- guided search
- kneser-ney smoothing
- language model adaptation
- language modeling
- list rescoring
- machine translation
- machine translation system
- machine translation systems
- maximum-entropy
- maximum-likelihood
- model adaptation
- model rescoring
- modeling
- noisy channel model
- normalization
- optimization
- optimization procedure
- phrase-based decoding
- porter stemmer
- preprocessing
- probability estimation
- processing
- recognition
- rescoring
- search
- smoothing
- smt system
- speech recognition
- statistical language modeling
- statistical machine translation
- statistical machine translation system
- stemmer
- tagger
- tagging
- translation system
- translation systems
- two-pass decoding
- weight optimization
- word translation
Other assigned terms:
- approach
- backoff
- baseline model
- bias
- bleu
- bleu score
- bleu scores
- case
- development set
- discriminative model
- error rate
- estimation
- factored language model
- feature
- feature information
- feature vectors
- french
- hypotheses
- hypothesis
- information sources
- interpolation
- knowledge
- language model
- language model score
- language models
- language pair
- language pairs
- linguistic
- linguistic knowledge
- log-linear model
- method
- model combination
- morphological information
- n-best list
- n-gram
- n-grams
- noisy channel
- normalization factor
- optimization criterion
- oracle
- part-of-speech
- part-of-speech information
- perplexity
- phrase
- pos information
- precision
- probabilities
- probability
- probability distributions
- probability estimates
- procedure
- run-time
- search procedure
- semantic
- semantic features
- sentence
- sentences
- source language
- stem
- system performance
- target language
- technique
- text
- toolkit
- training
- training data
- training set
- translation model
- translation models
- translation task
- translations
- trigram
- trigram language model
- trigram model
- unigram
- utterance
- word
- word alignments
- word classes
- word error rate
- word trigram
- word trigram model
- word-based model
- words
