Skip to content

Runs - Entity 2009

basewikirun

Results | Participants | Proceedings | Input | Summary

  • Run ID: basewikirun
  • Participant: Amsterdam
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: 09a923013cb1be36b71e66de2e7cd905
  • Run description: Baseline run on Wikipedia, homepages are found following outgoing links in Wikipedia.

BITDLDE09Run

Results | Participants | Proceedings | Input | Summary

  • Run ID: BITDLDE09Run
  • Participant: BIT
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: manual
  • MD5: 7a9014f1e29295e592d362141fb3e7b4
  • Run description: 1.To employ Indri to index dataset and query in passage retrival way; 2.To employ OpenEphyra to identify related entities and automatic generate queries for manual removing some wrong expanded words; 3.To employ OpenNLP's ME classifier to identify extracted entities' homepages; External resources: 1.To use wordnet to expand query; 2.To use DBPedia to extract product names as list for identifying entities;

ICTZHRun1

Results | Participants | Proceedings | Input | Summary

  • Run ID: ICTZHRun1
  • Participant: CAS_ICT_IR
  • Track: Entity
  • Year: 2009
  • Submission: 9/20/2009
  • Type: automatic
  • MD5: a86994464067726466cb14219dc0f01e
  • Run description: salient features: the similarity between the description documents for the named entities pairs (the source named entity and the named entity to be ranked). the relation index between the source named entity and the named entity to be ranked

ilpsEntBL

Results | Participants | Proceedings | Input | Summary

  • Run ID: ilpsEntBL
  • Participant: UAms
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 00bc0c05faef7a2512a2e346eaa17188
  • Run description: Named entities are recognized using an of the shell named entity tagger. This tagger was trained on the CoNLL 2003 english corpus. We use standard IR on entity co-occurence models to obtain a ranked list of candidates. We use the wikipedia category structure to filter candidate entities by type. This list is reranked based on the association between the input entity and candidate entities.

ilpsEntcf

Results | Participants | Proceedings | Input | Summary

  • Run ID: ilpsEntcf
  • Participant: UAms
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 6f8eb3d1aa187846b74b57bcdce702a2
  • Run description: Named entities are recognized using an of the shell named entity tagger. This tagger was trained on the CoNLL 2003 english corpus. We use standard IR on entity co-occurence models to obtain a ranked list of candidates. We determine the wikipedia categories related to a topic and then filter candidates if they are not in one of those associated categories.

ilpsEntcr

Results | Participants | Proceedings | Input | Summary

  • Run ID: ilpsEntcr
  • Participant: UAms
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 5a6fe27d9260a9a07c57aaddc8021557
  • Run description: Named entities are recognized using an of the shell named entity tagger. This tagger was trained on the CoNLL 2003 english corpus. We use standard IR on entity co-occurence models to obtain a ranked list of candidates. We use the wikipedia category structure to filter candidate entities by type. This list is reranked based on the wikipedia categories associated with a candidate entity and the categories associated with a topic.

ilpsEntem

Results | Participants | Proceedings | Input | Summary

  • Run ID: ilpsEntem
  • Participant: UAms
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 65a6f7e294f9d2e3e986c2345e90d4ed
  • Run description: Named entities are recognized using an of the shell named entity tagger. This tagger was trained on the CoNLL 2003 english corpus. We use standard IR on entity co-occurence models to obtain a ranked list of candidats. We use the wikipedia category structure to filter candidate entities by type.

KMR1PU

Results | Participants | Proceedings | Input | Summary

  • Run ID: KMR1PU
  • Participant: Purdue_Si
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: 1d1d97cf4c83d7f4bed5a6e0c6a27b89
  • Run description: We use Wikipedia to verify whether the target entities satisfy the constraints specified in the narratives.

KMR2PU

Results | Participants | Proceedings | Input | Summary

  • Run ID: KMR2PU
  • Participant: Purdue_Si
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: c644bf2bf8b8d01fd84d1b34d09ad8dc
  • Run description: We use Wikipedia to verify the specific type of the returned entities

KMR3PU

Results | Participants | Proceedings | Input | Summary

  • Run ID: KMR3PU
  • Participant: Purdue_Si
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: manual
  • MD5: 1f5fb6be2d24edfee2e490d4f85903f3
  • Run description: We design a table extraction method to obtain the answers from tables.

NiCTm1

Results | Participants | Proceedings | Input | Summary

  • Run ID: NiCTm1
  • Participant: NiCT
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: 678de8fb0b0d3ac3e8ab9f529f07b404
  • Run description: Use unsupervised SVM based on RBF kernel to ranking entities generated by UIUC NER tool; Beside NER tool, the other external resouce used in this RUN is Google search research.

NiCTm2

Results | Participants | Proceedings | Input | Summary

  • Run ID: NiCTm2
  • Participant: NiCT
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: 9c23977883e49c73d73d6bed14d8ce8e
  • Run description: Use Document language model to ranking entities generated by UIUC NER tool; Beside NER tool, the other external resouce used in this RUN is Google search research.

NiCTm3

Results | Participants | Proceedings | Input | Summary

  • Run ID: NiCTm3
  • Participant: NiCT
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: d61d8554f6aea6d58bc9136cbece2510
  • Run description: Use Ranking SVM to ranking entities generated by UIUC NER tool; Beside NER tool, the other external resouce used in this RUN is Google search research.

NiCTm4

Results | Participants | Proceedings | Input | Summary

  • Run ID: NiCTm4
  • Participant: NiCT
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: 4b95a971bf3c0557ff7d6877ab6ff003
  • Run description: Use Document language model to ranking entities generated by UIUC NER tool; NER tool is used as external resouce, Google is used to find homepages of entities.

PRIS1

Results | Participants | Proceedings | Input | Summary

  • Run ID: PRIS1
  • Participant: buptpris___2009
  • Track: Entity
  • Year: 2009
  • Submission: 9/23/2009
  • Type: automatic
  • MD5: 9fed977f67b256b74abc6d10adef03f7
  • Run description: Using original queries.

PRIS2

Results | Participants | Proceedings | Input | Summary

  • Run ID: PRIS2
  • Participant: buptpris___2009
  • Track: Entity
  • Year: 2009
  • Submission: 9/23/2009
  • Type: manual
  • MD5: c54e5ef97747264d9ba00f39968a7b06
  • Run description: Using original queries and narratives.

PRIS3

Results | Participants | Proceedings | Input | Summary

  • Run ID: PRIS3
  • Participant: buptpris___2009
  • Track: Entity
  • Year: 2009
  • Submission: 9/23/2009
  • Type: manual
  • MD5: fe88fbae889ccd7c77314567d11688ee
  • Run description: Terms of queris must appear ordered.

PRIS4

Results | Participants | Proceedings | Input | Summary

  • Run ID: PRIS4
  • Participant: buptpris___2009
  • Track: Entity
  • Year: 2009
  • Submission: 9/23/2009
  • Type: manual
  • MD5: ac9bd8fdcde67dc3c28f2568c6649c14
  • Run description: Retrievaling queries in title field.

tudpw

Results | Participants | Proceedings | Input | Summary

  • Run ID: tudpw
  • Participant: tudelft
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: 7a6d95058b999487be447da4f19ccb4b
  • Run description: First, we find the primary wikipedia page, then we rank all wikipedia pages that it links to

tudpwkntop

Results | Participants | Proceedings | Input | Summary

  • Run ID: tudpwkntop
  • Participant: tudelft
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: b79189e401f497d1453bb9b0577d9b55
  • Run description: First, we find the primary wikipedia page, then we rank all wikipedia pages that it links to, but give preference to those which names appear in the top-k relevant web-documents, plus we additionally give preference to the pages that are properly classified using DBPedia and Yago ontologies.

tudwebtop

Results | Participants | Proceedings | Input | Summary

  • Run ID: tudwebtop
  • Participant: tudelft
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: d30caf6483bd066ea5fa91f8bebbe32c
  • Run description: Just scoring outlinks from the given entity page

tudwtop

Results | Participants | Proceedings | Input | Summary

  • Run ID: tudwtop
  • Participant: tudelft
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 7bb6c935505d5e6d0fbe450692a9e558
  • Run description: We first select the top wiki pages and then give the preference to those which names we can find on the given entity page

UALRCB09r1

Results | Participants | Proceedings | Input | Summary

  • Run ID: UALRCB09r1
  • Participant: UALR_CB
  • Track: Entity
  • Year: 2009
  • Submission: 9/20/2009
  • Type: automatic
  • MD5: d1ddd3fac393b66716aa75f46ac251df
  • Run description: This run was based on the Indri query language. It was automatically done by the Indri Search Engine by giving the queries to the system. We indexed all the WARC pages into our INDRI search engine. First the supportive pages were found and then by finding the entities we found the home pages of every entity by giving the queries.

UALRCB09r2

Results | Participants | Proceedings | Input | Summary

  • Run ID: UALRCB09r2
  • Participant: UALR_CB
  • Track: Entity
  • Year: 2009
  • Submission: 9/20/2009
  • Type: automatic
  • MD5: 9b846e02deda14afb961c90d76efe098
  • Run description: This run was based on the Indri query language. It was automatically done by the Indri Search Engine by giving the queries to the system. We indexed all the WARC pages into our INDRI search engine. First the supportive pages were found and then by finding the entities we found the home pages of every entity by giving the queries.

UALRCB09r3

Results | Participants | Proceedings | Input | Summary

  • Run ID: UALRCB09r3
  • Participant: UALR_CB
  • Track: Entity
  • Year: 2009
  • Submission: 9/20/2009
  • Type: manual
  • MD5: e1755e8ac714c5bfd3dc123d3f5efdca
  • Run description: Using the keywords from the queries of each topic to query by Indri.

UALRCB09r4

Results | Participants | Proceedings | Input | Summary

  • Run ID: UALRCB09r4
  • Participant: UALR_CB
  • Track: Entity
  • Year: 2009
  • Submission: 9/20/2009
  • Type: automatic
  • MD5: 577b84314d3058c8dacdb70d33fe9888
  • Run description: This run was based on the Indri query language. It was automatically done by the Indri Search Engine by giving the queries to the system. We indexed all the WARC pages into our INDRI search engine. First the supportive pages were found and then by finding the entities we found the home pages of every entity by giving the queries.

UAmsER09Ab1

Results | Participants | Proceedings | Input | Summary

  • Run ID: UAmsER09Ab1
  • Participant: Amsterdam
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: b502a0b9eaf98a02a33e6c122d463804
  • Run description: Run on anchor text index with linear length prior. First three retrieved web pages are first 3 HP results, first Wikipedia article is first WP result, next three web pages are next 3 HP results, etc. Pages ranked after the first 300 web pages are pooled as support documents.

UAmsER09Co

Results | Participants | Proceedings | Input | Summary

  • Run ID: UAmsER09Co
  • Participant: Amsterdam
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: c102f4f1ed59a953d8856557811d2f2e
  • Run description: Documents co-cited with the example are returned. Co-cited documents receive score from the citing document based on the inverse of the outdegree of citing document. Only example entities and the link graph are used.

UdSmuCM

Results | Participants | Proceedings | Input | Summary

  • Run ID: UdSmuCM
  • Participant: EceUdel
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: 96ee2ce04847dbfd589a7544d485d2b2
  • Run description: Candidate generation is used to compute the score of the result entity on segment level with 100 words in length.

UdSmuCM50

Results | Participants | Proceedings | Input | Summary

  • Run ID: UdSmuCM50
  • Participant: EceUdel
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: 38f00b108c3ef822b168fd48a67f45a8
  • Run description: We use the candidate generation model on segment level with 50 words in length.

UdSmuTP

Results | Participants | Proceedings | Input | Summary

  • Run ID: UdSmuTP
  • Participant: EceUdel
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: de038de7ad31a58ca20f8a354dfff186
  • Run description: This method use the topic generation model with nonuniform priory to compute the score of the result entity.

UdSmuTU

Results | Participants | Proceedings | Input | Summary

  • Run ID: UdSmuTU
  • Participant: EceUdel
  • Track: Entity
  • Year: 2009
  • Submission: 9/21/2009
  • Type: automatic
  • MD5: d0b50e7802046a9f114d79fb0d7cf974
  • Run description: Topic generation model with uniform priory is used to compute the score of entities.

UIauto

Results | Participants | Proceedings | Input | Summary

  • Run ID: UIauto
  • Participant: UIUC
  • Track: Entity
  • Year: 2009
  • Submission: 9/23/2009
  • Type: automatic
  • MD5: 050bb21f6f21deb9e55cf09909e1a04d
  • Run description: Automatic run will query relation formulation and extended relation retrieval for snippets.

UIqryForm

Results | Participants | Proceedings | Input | Summary

  • Run ID: UIqryForm
  • Participant: UIUC
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: manual
  • MD5: 99535183094837d65b5357320bafbadc
  • Run description: Uses query formulation using WordNet. All retrieved entities are returned.

UIqryForm3

Results | Participants | Proceedings | Input | Summary

  • Run ID: UIqryForm3
  • Participant: UIUC
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: manual
  • MD5: 7f0f84823ca71ef9f1af61a42f53a78b
  • Run description: Uses query formulation using WordNet. Query reformulated to remove poor expansions. Three result per query.

uogTrEbl

Results | Participants | Proceedings | Input | Summary

  • Run ID: uogTrEbl
  • Participant: uogTr
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 5dbbfc6ffb9a1975704aa95c187eedc1
  • Run description: DFR parameter-free weighting model, Voting Model.

uogTrEc3

Results | Participants | Proceedings | Input | Summary

  • Run ID: uogTrEc3
  • Participant: uogTr
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 8fa0f0e9c1cc3a9978a9dad899b66d19
  • Run description: DFR parameter-free weighting model, Voting Model, within a novel framework for exploiting entity relationships.

uogTrEdi

Results | Participants | Proceedings | Input | Summary

  • Run ID: uogTrEdi
  • Participant: uogTr
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 1537f5e6bcfb153250d34da7294fe168
  • Run description: DFR parameter-free weighting model, Voting Model, and document centrality.

uogTrEpr

Results | Participants | Proceedings | Input | Summary

  • Run ID: uogTrEpr
  • Participant: uogTr
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 43d225c36168ca725fdc5a9bba61bf01
  • Run description: DFR parameter-free weighting and proximity models, Voting Model.

uwaterlooRun

Results | Participants | Input | Summary

  • Run ID: uwaterlooRun
  • Participant: Waterloo
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 05aa35e928e18d991f857f38bf5fb276
  • Run description: We did not use any external resource(s). Entity name and terms in the narrative were used as query terms. Passages were retrieved and expanded into a fixed sliding window size. 10-gram phrases were obtained from the passages. Terms in the retrieved passages were assigned weights using a tf-idf method described in Clarke et al. (2001). N-grams were assigned scores using the term weights. Only N-grams having a page on Wikipedia were presented in the run submitted (although, another run not submitted did not include this feature).

wikiruncats

Results | Participants | Proceedings | Input | Summary

  • Run ID: wikiruncats
  • Participant: Amsterdam
  • Track: Entity
  • Year: 2009
  • Submission: 9/22/2009
  • Type: automatic
  • MD5: 00c858358277f5364d357693899344bd
  • Run description: Ranks Wikipedia pages based on contents and category and follows links to discover homepages.