Abstract | For training, we derive growth transformations for phrase and lexicon translation probabilities to iteratively improve the objective. |
Abstract | For effective optimization, we derive updating formulas of growth transformation (GT) for phrase and lexicon translation probabilities . |
Abstract | Then the phrase translation probabilities were estimated based on the phrase alignments. |
Decoding | In the topic-specific lexicon translation model, given a source document, it first calculates the topic-specific translation probability by normalizing the entire lexicon translation table, and then adapts the lexical weights of rules correspondingly. |
Estimation | The process of rule-topic distribution estimation is analogous to the traditional estimation of rule translation probability (Chiang, 2007). |
Experiments | The lexicon translation probability is adapted by: |
Introduction | Such models first estimate word translation probabilities conditioned on topics, and then adapt lexical weights of phrases |
Head-Driven HPB Translation Model | o Phd_h7- and Phde- (3|t), translation probabilities for HD—HRs; |
Head-Driven HPB Translation Model | Plea, and Plea, (3|t), lexical translation probabilities for HD—HRs; |
Head-Driven HPB Translation Model | Ptyhcbhr = escp (—1), rule penalty for HD—HRs; PW (t|s), translation probability for NRRs; |