[ 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

Both sides previous revision Previous revision
Next revision
Previous revision
courses:rg:2012:longdtreport [2012/03/12 22:50]
longdt
courses:rg:2012:longdtreport [2012/03/12 22:59] (current)
longdt
Line 27: Line 27:
 - Use n array for n-gram model (array i-th is used for i-gram) - Use n array for n-gram model (array i-th is used for i-gram)
 - Each element in array in pair (w,c) - Each element in array in pair (w,c)
-    - w : index of that word in unigram array +     w : index of that word in unigram array  
-    c : offset pointer+     c : offset pointer
 - Sort base on w - Sort base on w
  
Line 52: Line 52:
 the underline reason is only to support binary search the underline reason is only to support binary search
  
-==== Conclusion ====+==== Decoding ==== 
 +**I. Exploiting Repetitive Queries**
  
-All in all, it was a paper worth reading, well presented, and thoroughly +The method use cache to speed up the process 
-discussed, bringing useful general ideas as well as interesting details.+This simple implementation increase performance of 300% over conventional implementation 
 +**II. Exploiting Scrolling Queries** 
 + 
 +We can quickly form the context encoding of the next query by concatenating new words with saved offset from previous query 
 + 
 +==== Conclusion ==== 
 +In summary, it was a worth reading paper, well presented, and thoroughly 
 +discussed. Paper contain many detail techniques that definitely helpful for actual implementation.  

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