[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki


[ Back to the navigation ]

Differences

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

Link to this comparison view

Next revision
Previous revision
courses:rg:2014:wr [2014/11/04 01:24]
hoangt created
courses:rg:2014:wr [2014/11/04 14:41] (current)
popel
Line 2: Line 2:
  
 1. In three main types of word representations described in the paper, to which types the following two samples belong: 1. In three main types of word representations described in the paper, to which types the following two samples belong:
-a) dog -0.087099201783 -0.136966257697 0.106813367913 [47 more numbers]+a)  
 +   dog -0.087099201783 -0.136966257697 0.106813367913 [47 more numbers]
    cat -0.103287428163 -0.0066971301398 -0.0346911076188 [47 more numbers]    cat -0.103287428163 -0.0066971301398 -0.0346911076188 [47 more numbers]
-b) dog 11010111010+b)  
 +   dog 11010111010
    cat 11010111010    cat 11010111010
-    +2. Section 4.1 defines a corrupted (or noise) n-gram, but there is a tiny error/typo in the definition. Try nitpicking and point it out.
-2. Section 4.1 defines a corrupted (or noise) n-gram, but there is a tiny error/typo in the definition. Try to be nitpicking and point it out.+
  
 3. Section 7.4 states that "word representations in NER brought larger gains on the out-of-domain data than on the in-domain data." Try to guess what is the reason. 3. Section 7.4 states that "word representations in NER brought larger gains on the out-of-domain data than on the in-domain data." Try to guess what is the reason.
Line 15: Line 16:
 b) Does it contain any compound feature with word representations?​ b) Does it contain any compound feature with word representations?​
 c) Give an example of a possible compound feature with word representations for the NER task.  c) Give an example of a possible compound feature with word representations for the NER task. 
 +
 +5.
 +Consider the C&W embedding vectors with 50 dimensions. Guess which word has the embedding vector most similar (by Euclidean distance) to the following vector:
 +a) vector(king) - vector(man) + vector(woman)
 +b) vector(dollars) - vector(dollar) + vector(mouse)
  
 **Hint** : The paper is 11-page long. You can skip section 2 and section 3.2 which are the literature review. **Hint** : The paper is 11-page long. You can skip section 2 and section 3.2 which are the literature review.
  
  

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