[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki

[ Back to the navigation ]

You are here: start » courses » rg » 2011 » deciphering_foreign_language

Differences

This shows you the differences between two versions of the page.

Link to this comparison view

Next revision 		Previous revision
Next revision Both sides next revision
courses:rg:2011:deciphering_foreign_language [2011/12/06 09:52]
tran vytvořeno 		courses:rg:2011:deciphering_foreign_language [2012/01/07 13:25]
tran
Line 3: Line 3:
 Scriber: Ke. T Scriber: Ke. T
  
 +The talk is about how to tackle MT without parallel training data.
 +
 +==== Section 1 ====
 +Given sentence pairs (e,f) where e is an English sentence and f is a foreign sentence, the translation model estimates parameter 
 +<latex>\theta</latex> such that
 +<latex>
 +\mathop {\arg \max }\limits_\theta  \prod\limits_\theta  {p_\theta  (f|e)} 
 +</latex>
 +
 +In case we do not have parallel data, we observe foreign text and try to maximize likelihood 
 +<latex>
 +\mathop {\arg \max }\limits_\theta  \prod\limits_f {p_\theta  (f)} 
 +</latex>
 +
 +Treating English translation as hidden alignment, our task is to find the parameter <latex>\theta</latex> that
 +<latex>
 +\mathop {\arg \max }\limits_\theta  \prod\limits_f {\sum\limits_e {P(e) \times \sum\limits_a {P_\theta  (f,a|e)} } } 
 +</latex>

[ Back to the navigation ] [ Back to the content ]