[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki

[ Back to the navigation ]

You are here: start » courses » rg » 2013 » hmm-perc-experiments

Differences

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

Link to this comparison view

Next revision 		Previous revision
courses:rg:2013:hmm-perc-experiments [2013/03/13 01:47]
ufal vytvořeno 		courses:rg:2013:hmm-perc-experiments [2013/04/01 12:33] (current)
cechj5cm
Line 9: Line 9:
     N N X N X     N N X N X
 How would this result alter values of alfa_X,X,X, alfa_N,father? How would this result alter values of alfa_X,X,X, alfa_N,father?
-Supposing that the best tag sequence don't change, what would be your answer if "father/N" is replaced by "luke/X"?+Supposing that the best tag sequence won't change, what would be your answer if "father/N" is replaced by "luke/X"?
  
 2. Suppose this tagged sentence as the only entry in your training data: 2. Suppose this tagged sentence as the only entry in your training data:
Line 23: Line 23:
  
 (You might compare this problem to similar one in paper presented on RG in winter - http://aclweb.org/anthology-new/N/N10/N10-1069.pdf) (You might compare this problem to similar one in paper presented on RG in winter - http://aclweb.org/anthology-new/N/N10/N10-1069.pdf)
 +
 +===SHORT REPORT===
 +Intro
 +Autors said that training perceptron is quicker and easier solution.
 +Definition
 +We defined structured perceptron and parametres for special case of structured perceptron.
 +Parameters are log of conditional probability for trigram words and conditional probability for word and tag.
 +αx,y,z = log P(z | x,y)
 +αt,w = log P(w| t)
 +Logs are used for precision
 +Then speaker show how work process of learning.
 +Autors then deal with separable and inseparable data.
 +They defined condition for inseparable data.
 +
 +We spend rest of time solution of questions. During that solution were shown others fetatures of this percepton training. 

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