ACL RD-TEC 1.0 Summarization of W06-1672
Paper Title:
DISCRIMINATIVE METHODS FOR TRANSLITERATION
DISCRIMINATIVE METHODS FOR TRANSLITERATION
Authors: Dmitry Zelenko and Chinatsu Aone
Primarily assigned technology terms:
- algorithm
- beam search
- binary classification
- binary classifier
- candidate scoring
- classification
- classifier
- classifier learning
- classifiers
- computational linguistics
- cross-lingual information extraction
- decoding
- decomposition
- discriminative approach
- discriminative modeling
- em algorithm
- extraction systems
- feature representation
- global transliteration modeling
- good-turing smoothing
- information extraction
- language modeling
- language processing
- learning
- learning algorithm
- machine learning
- maximum likelihood
- modeling
- name transliteration
- natural language processing
- perceptron
- perceptron algorithm
- perceptron training
- probabilistic modeling
- probabilistic modeling approach
- processing
- re-ranking
- relative distance
- scoring
- search
- smoothing
- spelling
- transcription
- transliteration
- transliteration modeling
- viterbi
- voted perceptron
- voted perceptron algorithm
Other assigned terms:
- alphabet
- approach
- approach to transliteration
- association for computational linguistics
- beam
- binary classification problem
- case
- characters
- classification problem
- conditional probability
- dictionary
- distribution
- edit distance
- empty string
- feature
- feature space
- foreign language
- joint probability
- language model
- language pairs
- learning problem
- likelihood
- linguistics
- log-likelihood
- mapping
- maps
- method
- names
- natural language
- noise
- probabilistic approach
- probabilistic model
- probabilities
- probability
- probability distribution
- probability estimates
- process
- russian
- self-contained transliteration model
- substring
- symbol
- symbols
- terms
- text
- training
- training examples
- transliteration candidate
- transliteration model
- transliteration performance
- weight vector
- words
