de.mpg.escidoc.pubman.appbase.FacesBean
English
 
Help Guide Disclaimer Contact us Login
  Advanced SearchBrowse

Item

ITEM ACTIONSEXPORT

Released

Conference Paper

HARPY: Hypernyms and Alignment of Relational Paraphrases

MPS-Authors
http://pubman.mpdl.mpg.de/cone/persons/resource/persons137149

Grycner,  Adam
Databases and Information Systems, MPI for Informatics, Max Planck Society;

http://pubman.mpdl.mpg.de/cone/persons/resource/persons45720

Weikum,  Gerhard
Databases and Information Systems, MPI for Informatics, Max Planck Society;

Locator
There are no locators available
Fulltext (public)

C14-1207.pdf
(Any fulltext), 197KB

Supplementary Material (public)
There is no public supplementary material available
Citation

Grycner, A., & Weikum, G. (2014). HARPY: Hypernyms and Alignment of Relational Paraphrases. In J. Hajic, & J. Tsujii (Eds.), Proceedings of COLING 2014: Technical Papers (pp. 2195-2204). Dublin, Ireland: ACL. Retrieved from http://www.aclweb.org/anthology/C14-1207.


Cite as: http://hdl.handle.net/11858/00-001M-0000-0024-3329-1
Abstract
Collections of relational paraphrases have been automatically constructed from \u000Alarge text corpora, as a WordNet counterpart for the realm of binary predicates \u000Aand their surface forms.\u000AHowever, these resources fall short in their coverage of hypernymy links \u000A(subsumptions) among the synsets of phrases. \u000AThis paper closes this gap by computing a high‐quality alignment between the \u000Arelational phrases of the Patty taxonomy, one of the largest collections of \u000Athis kind, and the verb senses of WordNet. To this end, we devise judicious \u000Afeatures and develop a graph‐based alignment algorithm by adapting and \u000Aextending the SimRank random‐walk method.\u000AThe resulting taxonomy of relational phrases and verb senses, coined HARPY, \u000Acontains 20,812 synsets organized into a \em Directed Acyclic Graph (DAG)} \u000Awith 616,792 hypernymy links. \u000AOur empirical assessment, indicates that the alignment links between Patty and \u000AWordNet have high accuracy, with {\em Mean Reciprocal Rank (MRR)} score 0.7 and \u000A{\em Normalized Discounted Cumulative Gain (NDCG) score 0.73. \u000AAs an additional extrinsic value, HARPY provides fine‐grained lexical types for \u000Athe arguments of verb senses in WordNet.