Difference between revisions of "POS Induction (State of the art)"

From ACL Wiki
Jump to: navigation, search
(Software: fix mkcls link)
 
(11 intermediate revisions by 6 users not shown)
Line 1: Line 1:
 +
==Evaluation==
 +
 +
'''Many-to-1:''' Map every induced label to a gold standard tag greedily (45 labels to 45 tags of the Penn tag set). Use the mapping to compute tag accuracy on the Wall Street Journal portion of the Penn TreeBank.
 +
 +
 +
==Results==
 +
'''Listed in order of decreasing accuracy'''
 +
 +
 
{| border="1" cellpadding="5" cellspacing="1" width="100%"
 
{| border="1" cellpadding="5" cellspacing="1" width="100%"
 
|-
 
|-
Line 7: Line 16:
 
! Many-to-1
 
! Many-to-1
 
|-
 
|-
| Prototype-based+Brown
+
| UPOS
 +
| Learning Syntactic Categories Using Paradigmatic Representations of Word Context
 +
| Yatbaz et al. (2012)
 +
| [https://github.com/ai-ku/upos/tree/emnlp2012 upos]
 +
| 80.2%
 +
|-
 +
| Brown+proto
 
| MRF initialized with Brown prototypes
 
| MRF initialized with Brown prototypes
 
| Christodoulopoulos, Goldwater and Steedman (2010)
 
| Christodoulopoulos, Goldwater and Steedman (2010)
 
|  
 
|  
 
| 76.1%
 
| 76.1%
 +
|-
 +
|
 +
| Logistic regression with features and LBFGS
 +
| Berg-Kirkpatrick et al. (2010)
 +
|
 +
| 75.5%
 +
|-
 +
| Clark DMF
 +
| Distributional clustering + morphology + frequency
 +
| Clark (2003)
 +
| [http://www.cs.rhul.ac.uk/home/alexc/pos2.tar.gz alexc]
 +
| 71.2%*
 
|-
 
|-
 
|}
 
|}
 +
 +
<nowiki>*</nowiki> according to Christodoulopoulos, Goldwater and Steedman (2010)
  
  
 
== References ==
 
== References ==
 +
'''Listed alphabetically.'''
 +
 +
* Berg-Kirkpatrick, Taylor, Alexandre Bouchard-Cote, John DeNero, and Dan Klein. 2010. [http://www.aclweb.org/anthology/N/N10/N10-1083.pdf Painless Unsupervised Learning with Features]. NAACL 2010.
 +
* Christodoulopoulos, Christos, Sharon Goldwater and Mark Steedman. 2010. [http://www.aclweb.org/anthology/D/D10/D10-1056.pdf Two Decades of Unsupervised POS induction: How far have we come?] In Proceedings of EMNLP 2010.
 +
* Clark, Alexander. 2003. [http://www.aclweb.org/anthology/E/E03/E03-1009.pdf Combining distributional and morphological information for part of speech induction]. In Proceedings of EACL 2003, pages 59–66, Morristown, NJ, USA.
 +
* Yatbaz, Mehmet Ali, Enis Sert and Deniz Yuret. 2012. [http://aclweb.org/anthology//D/D12/D12-1086.pdf Learning Syntactic Categories Using Paradigmatic Representations of Word Context]. In Proceedings of EMNLP 2012, pages 940–951.
  
* [http://www.aclweb.org/anthology/D/D10/D10-1056.pdf Christos Christodoulopoulos, Sharon Goldwater and Mark Steedman. 2010. Two Decades of Unsupervised POS induction: How far have we come? In Proceedings of EMNLP 2010]
+
== Software ==
 +
* [http://www.cs.rhul.ac.uk/home/alexc/pos2.tar.gz alexc]
 +
* [https://github.com/percyliang/brown-cluster brown-cluster]
 +
* [https://code.google.com/p/giza-pp/ mkcls]
 +
* [http://wortschatz.uni-leipzig.de/~cbiemann/software/unsupos.html unsupos]
 +
* [https://github.com/ai-ku/upos upos]
  
 
== See also ==
 
== See also ==

Latest revision as of 17:40, 7 March 2014

Evaluation

Many-to-1: Map every induced label to a gold standard tag greedily (45 labels to 45 tags of the Penn tag set). Use the mapping to compute tag accuracy on the Wall Street Journal portion of the Penn TreeBank.


Results

Listed in order of decreasing accuracy


System name Short description Main publications Software Many-to-1
UPOS Learning Syntactic Categories Using Paradigmatic Representations of Word Context Yatbaz et al. (2012) upos 80.2%
Brown+proto MRF initialized with Brown prototypes Christodoulopoulos, Goldwater and Steedman (2010) 76.1%
Logistic regression with features and LBFGS Berg-Kirkpatrick et al. (2010) 75.5%
Clark DMF Distributional clustering + morphology + frequency Clark (2003) alexc 71.2%*

* according to Christodoulopoulos, Goldwater and Steedman (2010)


References

Listed alphabetically.

Software

See also