[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki

[ Back to the navigation ]

You are here: start » courses » rg » 2013 » composite-activities

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:2013:composite-activities [2013/04/23 04:37]
machys 		courses:rg:2013:composite-activities [2013/09/29 21:35] (current)
machys
Line 1: Line 1:
-   - After reading first three chapters:  +   - After reading the first three chapters:  
-      * list the main parts/components/stuctures of the model.  +      * list the main parts/components/structures of the model.  
-      * Is their creation dependant on each other? +      * Is their creation dependent on other components? 
-   - Thinking about the scripts (recipes): +   - Thinking about the scripts: 
-      * What is the main reason (the biggest advantage) of using scripts? What kind of information it brings? (Hint: page 2, page 8) +      * What is the main reason (the biggest advantage) of using scripts? What kind of information does it bring? (Hint: page 2, page 8) 
-      * Truth is - authors don't get the "knowledge" of scripts straightforward - how is that "knowledge" represented in model and which (two) ways are used to get it. +      * The authors don't get the "knowledge" of scripts straightforwardly. How is that "knowledge" represented in the model and which (four) ways are used to get it? 
-   - In last paragraph of section 3 theres described method that enhances the robustness of the model (binarization of all association weights <html> <math xmlns="http://www.w3.org/1998/Math/MathML" alttext="association weights"> <msubsup><mi>w</mi><mi>i</mi><mo>z</mo></msubsup> </math> </html>). Answer one of following questions (choose one): +   - In the last paragraph of Section 3, a method is described that enhances the robustness of the model (binarization of all association weights <latex>w^z_i</latex>). Answer one of the following questions (choose one): 
-      * Why it works? (=> Why it should work the best?) +      * Why does it work? (=> Why should it work best?) 
-      * Have you any idea how to make it different way? +      * Do you have any idea how to do it differently? 
-   - In experiements part, which tools enhanced the tasks Atribute recognition and Composite activity classification "the most". Try answer why?+   - Experiments: Which "tricks"/"parts of processing" enhanced the //Attribute recognition// and //Composite activity classification// tasks "the most"? Try to answer why. 
 + 
 +====== Answers ====== 
 + 
 +   - First set 
 +      * list components [[https://docs.google.com/drawings/d/12wIoIDVDV3b6EbJAVVIzZn9h0ymIgHRHi_sTauGM08A/edit|(Google doc graph)]] 
 +      * dependance of components (the same graph) 
 +   - Scripts 
 +      * reason?: Cheap source of training data, Many combinations, unseen variants, “decsriptions” of the same thing 
 +      * four ways: 2x2: 1) direct use of words from data or 2) mapping word classes from WordNet X 3) simple word frequency or 4) TF*IDF 
 +   - There was a discussion about 3rd set of question. We are not sure why authors do that. There was strongly supported opinion that autohors do a lot unnecessary work, which is lost by binarization. 
 +   - 4th: Majority people in aswers nominated the use of TF*IDF in case of no training data as the best idea. 

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