[ 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:2011-report-parser [2012/09/27 11:03]
ufal
courses:rg:2011-report-parser [2012/09/27 11:22] (current)
ufal
Line 1: Line 1:
-====== Using a Wikipedia-based Semantic Relatedness Measure for Document Clustering ======+====== A Fast, Accurate, Non-Projective, Semantically-Enriched Parser ======
  
 written by Stephen Tratz and Eduard Hovy (Information Sciences Institute, University of Southern Carolina) written by Stephen Tratz and Eduard Hovy (Information Sciences Institute, University of Southern Carolina)
Line 32: Line 32:
           * they reduced the number of generic "dep/DEP" relation           * they reduced the number of generic "dep/DEP" relation
               * Stanford tags are hierarchical and "dep/DEP" is the top-most one               * Stanford tags are hierarchical and "dep/DEP" is the top-most one
-          * 1.3% of arcs are non-projective (out of 8.1% of all non-projective arcs) because of the following conversion (agreement can be a motivation for this, i.e. in Czech): +          * correcting of POS using the syntactic info + additional rules for specific word forms 
-            {{:courses:rg:dependency-conversion.png|}} +          * 1.3% of arcs are non-projective (out of 8.1% of all non-projective arcs) because of the following conversion (agreement can be a motivation for this, i.e. in Czech): {{:courses:rg:dependency-conversion.png|}} 
-            +      - additional changes and the final conversion from the intermediate output to a dependency structure 
 ==== Parser ==== ==== Parser ====
  
Line 52: Line 53:
     * "move" operation:     * "move" operation:
 {{:courses:rg:move_operation.png|}} {{:courses:rg:move_operation.png|}}
 +
 +==== Features ====
 +  * Brown et al. clusters - they are surprisingly used rarely
 +==== Features ====
 +  * not much discussed in the paper
 +==== Evaluation ====
 +  * they make the same transformation as they did in Section 2
 +==== Shallow semantic annotation ====
 +  * 4 optional modules
 +      * preposition sense disambiguation
 +      * noun compound interpretation
 +      * possesives interpretation
 +      * PropBank semantic role labelling
 +          * (Hajič et al., 2009) is not in the list of references
 +===== Conclusion =====
 +  * non-directional easy-first parsing
 +  * new features - Brown et al. clusters
 +  * fast and accurate
 +  * modified Penn converter
 +      * changes in 9.500 POS tags
 +      * labels copula, coordination
 +  * semantic info

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