[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki

[ Back to the navigation ]

You are here: start » courses » rg » 2012 » encouraging-consistent-translation

Differences

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

Link to this comparison view

Both sides previous revision Previous revision
Next revision Both sides next revision
courses:rg:2012:encouraging-consistent-translation [2012/10/16 15:15]
dusek 		courses:rg:2012:encouraging-consistent-translation [2012/10/17 11:40]
dusek
Line 37: Line 37:
       * Beware, this notion of grouping is not well-defined, does not create equivalence classes: "old hostages" = "new hostages" = "completely new hostages" but "old hostages" != "completely new hostages" (we hope this didn't actually happen)       * Beware, this notion of grouping is not well-defined, does not create equivalence classes: "old hostages" = "new hostages" = "completely new hostages" but "old hostages" != "completely new hostages" (we hope this didn't actually happen)
     * Cases where //only one translation variant prevails// are //discarded// (this is the case of "Korea")     * Cases where //only one translation variant prevails// are //discarded// (this is the case of "Korea")
 +
 +==== Sec. 4. Approach ====
 +
 +The actual experiments begin only now; the used data is different.
 +
 +**Choice of features**
 +  * They define 3 features that are designed to be biased towrds consistency -- or are they?
 +    * If e.g. two variants are used 2 times each, they will have roughly the same score
 +  * The new features need two passes through the data
 +  * The BM25 function is a refined version of a [[http://en.wikipedia.org/wiki/TF-IDF|TF-IDF]] score
 +  * The exact parameter values are probably not tuned, left at a default value (and maybe they don't have much influence anyway)
 +   * See NPFL103 for details on Information retrieval, it's largely black magic
 +
 +**Feature weights**
 +  * The usual model in MT is scoring the hypotheses according to the feature values (''f'') and their weights (''lambda''): 
 +    * ''score(H) = exp( sum( lambda_i * f_i(H)) )''
 +  * The feature weights are trained on a heldout data set using [[http://acl.ldc.upenn.edu/acl2003/main/pdfs/Och.pdf|MERT]] (or, here: [[http://en.wikipedia.org/wiki/Margin_Infused_Relaxed_Algorithm|MIRA]])
 +  * The resulting weights are not mentioned, but if the weight is < 0, will this favor different translation choices?
 +
 +**Differences in features**
 +  * C1 indicates that a certain Hiero rule was used frequently
 +    * but rules are very similar, so we also need something less fine-grained
 +  * C2 is a target-side feature, just counts the target side tokens
 +    * It may be compared to Language Model features, but is trained only on the target part of the bilingual training data.
 +

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