[ Skip to the content ]

Institute of Formal and Applied Linguistics Wiki

• ÚFAL Wiki Home
• ÚFAL Home
• Sitemap
• RSS
• ÚFAL calendar
• Recent changes
• Login
• Hledat:

[ Back to the navigation ]

Differences

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

View differences: Side by SideInline Go

Link to this comparison view

Both sides previous revision Previous revision 2012/09/27 11:22 ufal 2012/09/27 11:21 ufal 2012/09/27 11:20 ufal 2012/09/27 11:17 ufal 2012/09/27 11:15 ufal 2012/09/27 11:12 ufal 2012/09/27 11:09 ufal 2012/09/27 11:03 ufal 2012/09/27 10:58 ufal 2012/09/27 10:54 ufal 2012/09/27 10:39 ufal 2012/09/27 10:36 ufal 2012/09/27 10:14 ufal 2012/09/27 09:41 ufal 2012/09/27 09:31 ufal 2012/09/26 13:57 ufal 2012/09/26 13:56 ufal 2012/09/26 13:49 ufal 2012/09/26 13:46 ufal vytvořeno Go Next revision		Previous revision 2012/09/27 11:22 ufal 2012/09/27 11:21 ufal 2012/09/27 11:20 ufal 2012/09/27 11:17 ufal 2012/09/27 11:15 ufal 2012/09/27 11:12 ufal 2012/09/27 11:09 ufal 2012/09/27 11:03 ufal 2012/09/27 10:58 ufal 2012/09/27 10:54 ufal 2012/09/27 10:39 ufal 2012/09/27 10:36 ufal 2012/09/27 10:14 ufal 2012/09/27 09:41 ufal 2012/09/27 09:31 ufal 2012/09/26 13:57 ufal 2012/09/26 13:56 ufal 2012/09/26 13:49 ufal 2012/09/26 13:46 ufal vytvořeno Go
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 ]