Wikipedia - RTE Users

From ACL Wiki
Revision as of 06:20, 22 December 2009 by Celct (Talk | contribs)

(diff) ← Older revision | Latest revision (diff) | Newer revision → (diff)
Jump to: navigation, search


Participants* Campaign Version Specific usage description Evaluations / Comments
BIU RTE5 Used for the creation of WikiRules!, a resource of lexical reference rules Cfr. evaluation of WikiRules!
Cswhu RTE5 Named entity relations extraction Ablation test performed. Positive impact of the resosurce: +1.33% on two-way, +3.34% on three-way task
FBKirst RTE5 Rules extracted from WP using Latent Semantic Analysis (LSA) Ablation test performed. Positive impact of the resosurce: +1% on two-way task
PeMoZa RTE5 17 October 2009 Similarity measure based on WikiPedia, using jLSA (java Latent Semantic Analysis) No ablation test performed.

It is worth mentioning that the similarity measure over Wikipedia, not only covers all the existing pairs, but its computation is also about 100 times faster than the previously used tool based on WordNet

QUANTA RTE5 Wikipedia redirection set is employed as a synonymous entity set to detect the entity entailment. No ablation test performed.
UAIC RTE5 FIRST USE: Extraction of a new resource based on named entity relation mining using Wikipedia (Iftene and Balahur-Dobrescu, 2008)

SECOND USE: Extraction of a new resource aimed to identify a "distance" between one name entity from hypothesis and name entities from text

FIRST USE: Ablation test performed. Positive impact of the resource: +0.17% on two-way, +0.50% on three-way task.

SECOND USE: Ablation test performed (Wikipedia+NER's+Perl patterns). Positive impact of the whole NE module: +6.17% on two-way, +5% on three-way task.

BIU RTE4 Used for the creation of WikiRules!, a resource of lexical reference rules Cfr. evaluation of WikiRules!
UAIC RTE4 Data taken from the RTE4 proceedings. Participants are recommended to add further information.
UPC RTE4 Data taken from the RTE4 proceedings. Participants are recommended to add further information.
New user Participants are encouraged to contribute.
Total: 9


[*] For further information about participants, click here: RTE Challenges - Data about participants

   Return to RTE Knowledge Resources