Runs - News 2019¶

cityuni_1¶

Run ID: cityuni_1
Participant: cityuni
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: 34f1032ac6620e58fb6c01f210d0034d
Run description: This run is provided based on Stochastic Hill climbing approachthat allows to recommend a set of ~20 background link for each Topic.

cityuni_ER1¶

Run ID: cityuni_ER1
Participant: cityuni
Track: News
Year: 2019
Submission: 9/22/2019
Type: auto
Task: entity
MD5: 4ecd5f19786cb8f539b0aa90d12b35bf
Run description: This run uses the wikipedia Dump along with optimisation algo to rank the entities

cityuni_ER2¶

Run ID: cityuni_ER2
Participant: cityuni
Track: News
Year: 2019
Submission: 9/22/2019
Type: auto
Task: entity
MD5: f583a53c0781ffd87347850e72302e58
Run description: This run uses the wikipedia Dump along with BM25 to rank the entities

clac_100_cos¶

Results | Participants | Input | Summary | Appendix

Run ID: clac_100_cos
Participant: CLAC_NEWS_2019
Track: News
Year: 2019
Submission: 8/19/2019
Type: auto
Task: background
MD5: 7fa9f8b7a18e89ee613d96fff159e1f5
Run description: cosine similarity on doc2vec with 100 dimensions

clac_300_cos¶

Results | Participants | Input | Summary | Appendix

Run ID: clac_300_cos
Participant: CLAC_NEWS_2019
Track: News
Year: 2019
Submission: 8/19/2019
Type: auto
Task: background
MD5: 3c956e104a0cef37542e4c58e9e76209
Run description: cosine similarity on doc2vec with 300 dimensions

CMU_NS-1-tpb¶

Run ID: CMU_NS-1-tpb
Participant: CMU
Track: News
Year: 2019
Submission: 9/23/2019
Type: auto
Task: entity
MD5: 1ce26390f3a02001ed93e7f1a3bfbe46
Run description: LTR pipeline adapted to the entity ranking. Features include TFIDF, BM25 and window operators, over the title first paragraphs and body of the document.

CMU_NS-2-tp¶

Run ID: CMU_NS-2-tp
Participant: CMU
Track: News
Year: 2019
Submission: 9/23/2019
Type: auto
Task: entity
MD5: 440b2b47327251379105f78f6027ebb9
Run description: LTR pipeline adapted to the entity ranking. Features include TFIDF, BM25 and window operators, over the title, and first paragraphs of the document.

CMU_NS-3-t¶

Run ID: CMU_NS-3-t
Participant: CMU
Track: News
Year: 2019
Submission: 9/23/2019
Type: auto
Task: entity
MD5: 4b4f0ecebf6825247a94719e1fbc5058
Run description: LTR pipeline adapted to the entity ranking. Features include TFIDF, BM25 and window operators, over the title of the document.

ICTNET_estem¶

Run ID: ICTNET_estem
Participant: ICTNET
Track: News
Year: 2019
Submission: 9/21/2019
Type: auto
Task: entity
MD5: 636fdfb78a989ae0742910998fe0814b
Run description: lower case, remove stop words, stem, top 100 for each wiki entity, top 1000 for each rank score.

ICTNET_stem¶

Run ID: ICTNET_stem
Participant: ICTNET
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: 0dc865096d7b266b14bb11fefdd28743
Run description: 'news_stem', lower case, remove stop words, all the words, title boost: 5.5

OzU_wiki¶

Run ID: OzU_wiki
Participant: OzUNLP
Track: News
Year: 2019
Submission: 9/21/2019
Type: auto
Task: entity
MD5: 490d78220bfec3b2309c708602c2f858
Run description: Doc2Vec + Wikipedia

OzU_wiki_1_ws¶

Run ID: OzU_wiki_1_ws
Participant: OzUNLP
Track: News
Year: 2019
Submission: 9/22/2019
Type: auto
Task: entity
MD5: 96d2258e969ad92efaff04a3271a68a0
Run description: Doc2Vec + Wikipedia + Most Similar News Article + Weighted Score

OzU_wiki_5_ws¶

Run ID: OzU_wiki_5_ws
Participant: OzUNLP
Track: News
Year: 2019
Submission: 9/22/2019
Type: auto
Task: entity
MD5: 6831729dbf80b3e1148758a66e779353
Run description: Doc2Vec + Wikipedia + Most Similar 5 News Articles + Weighted Score

OzU_wiki_top1¶

Run ID: OzU_wiki_top1
Participant: OzUNLP
Track: News
Year: 2019
Submission: 9/22/2019
Type: auto
Task: entity
MD5: f8b5177d0febad4039c6a40020b9235a
Run description: Doc2Vec + Wikipedia + Most Similar News Article

OzU_wiki_top5¶

Run ID: OzU_wiki_top5
Participant: OzUNLP
Track: News
Year: 2019
Submission: 9/22/2019
Type: auto
Task: entity
MD5: a2a292e8170bd2804331b37cd27d3bac
Run description: Doc2Vec + Wikipedia + Most Similar 5 News Articles

ql¶

Run ID: ql
Participant: ICTNET
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: 6a7d3a88ae580e6d6333b47fe3dafe67
Run description: query likelihood with Jelinek Mercer Similarity.

QU_KCore¶

Run ID: QU_KCore
Participant: QU
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: 293e78eb9d3fb3df7fdfeb0f94fef9e6
Run description: The query article is converted into a graph of words (using a tuned window size). Each node/word is assigned a weight using K-Core graph analysis method. The top K nodes were then selected to form a weighted query that is eventually issued against the given news collection. Returned articles are restricted to precede the query article.

QU_KCore_F¶

Run ID: QU_KCore_F
Participant: QU
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: ef1ccd5dd976859366b85e97e5e2e77f
Run description: The query article is converted into a graph of words (using a tuned window size). Each node/word is assigned a weight using K-Core graph analysis method. The top K nodes were then selected to form a weighted query that is eventually issued against the given news collection. Returned articles are not restricted to precede the query article, i.e., follow up articles can appear in the returned articles.

QU_KTruss¶

Run ID: QU_KTruss
Participant: QU
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: 9ade5e132ed47dfb99e7f155c1e6c97d
Run description: The query article is converted into a graph of words (using a tuned window size). Each node/word is assigned a weight using K-Truss graph analysis method. The top K nodes were then selected to form a weighted query that is eventually issued against the given news collection. Returned articles are restricted to precede the query article.

QU_KTruss_F¶

Run ID: QU_KTruss_F
Participant: QU
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: da86d7e5ba2811cee2ec1d3268dcd33e
Run description: The query article is converted into a graph of words (using a tuned window size). Each node/word is assigned a weight using K-Truss graph analysis method. The top K nodes were then selected to form a weighted query that is eventually issued against the given news collection. Returned articles are not restricted to precede the query article, i.e., follow up articles can appear in the returned articles.

rm3¶

Run ID: rm3
Participant: ICTNET
Track: News
Year: 2019
Submission: 8/19/2019
Type: auto
Task: background
MD5: 13dfe60cf0db4346e7005c601fd90490
Run description: query likelihood with RM3 relevance model.

rocchio¶

Run ID: rocchio
Participant: ICTNET
Track: News
Year: 2019
Submission: 8/19/2019
Type: auto
Task: background
MD5: c8e86533425c63b4cfe12f9a4b283b46
Run description: Use bm25 to retrieve candidate documents and rerank them by rocchio algorithm.

ru-ent-90-10-df¶

Results | Participants | Input | Summary | Appendix

Run ID: ru-ent-90-10-df
Participant: RUIR
Track: News
Year: 2019
Submission: 8/19/2019
Type: auto
Task: background
MD5: 9a0d0d6e1d31260d9ed8b217477c21c7
Run description: BM25 RM3 ELR Used TAGME API as a external resource: https://tagme.d4science.org

ru-ent-95-05-df¶

Results | Participants | Input | Summary | Appendix

Run ID: ru-ent-95-05-df
Participant: RUIR
Track: News
Year: 2019
Submission: 8/19/2019
Type: auto
Task: background
MD5: 8bb45a40ee9c491bec122a4399abc9a6
Run description: BM25 RM3 ELR Used TAGME API as a external resource: https://tagme.d4science.org

ru-invwiki¶

Results | Participants | Input | Summary | Appendix

Run ID: ru-invwiki
Participant: RUIR
Track: News
Year: 2019
Submission: 9/20/2019
Type: auto
Task: entity
MD5: 4d885a9aa4f01def52aaf63ffbcfb7fe
Run description: Length of the wikipedia pages of the entities. Python wiki api is used.

ru-m-order¶

Results | Participants | Input | Summary | Appendix

Run ID: ru-m-order
Participant: RUIR
Track: News
Year: 2019
Submission: 9/20/2019
Type: auto
Task: entity
MD5: 6f19fb27081c2083dedfe9ca9a757e16
Run description: Order in which entities are mentioned in the topic article.

ru-t-order¶

Results | Participants | Input | Summary | Appendix

Run ID: ru-t-order
Participant: RUIR
Track: News
Year: 2019
Submission: 9/20/2019
Type: auto
Task: entity
MD5: 73df933c46407ed60598ef1653485eec
Run description: topic order

ru-tf-invwiki¶

Results | Participants | Input | Summary | Appendix

Run ID: ru-tf-invwiki
Participant: RUIR
Track: News
Year: 2019
Submission: 9/20/2019
Type: auto
Task: entity
MD5: cb5061213598ff0459f989261f43bab2
Run description: Count of entity mentions. Length of the wikipedia pages of the entities. Python wiki api is used.

ru-tf-m-ord¶

Results | Participants | Input | Summary | Appendix

Run ID: ru-tf-m-ord
Participant: RUIR
Track: News
Year: 2019
Submission: 9/20/2019
Type: auto
Task: entity
MD5: 84e7dce1fec51a5e97660cca9bc6ce9b
Run description: Count of entity mentions. Order in which entities are mentioned in the topic article.

ru_bm25_rm3¶

Results | Participants | Input | Summary | Appendix

Run ID: ru_bm25_rm3
Participant: RUIR
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: 2819b7f46514cd151a64f5896ba84fe4
Run description: bm25 rm3

ru_bm25_rm3_fil¶

Results | Participants | Input | Summary | Appendix

Run ID: ru_bm25_rm3_fil
Participant: RUIR
Track: News
Year: 2019
Submission: 8/16/2019
Type: auto
Task: background
MD5: 0921e54ffba0876884389d8428e1fb74
Run description: BM25 RM3 date filter

ru_sdm_rm3_fil¶

Results | Participants | Input | Summary | Appendix

Run ID: ru_sdm_rm3_fil
Participant: RUIR
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: 57d7a03ab225ca85511e10fd4939f55c
Run description: sdm rm3 datefilter

sils_news_run1¶

Run ID: sils_news_run1
Participant: UNC_SILS
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: ba86ba975f777db0e0ab812607b933e1
Run description: Spacy for NER

sils_news_run2¶

Run ID: sils_news_run2
Participant: UNC_SILS
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: bb1856f39934f119ed244959ae061c79
Run description: Spacy for NER

sils_news_run3¶

Run ID: sils_news_run3
Participant: UNC_SILS
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: 63aed52127750e3be96eaba5f0b74a0a
Run description: Spacy for NER

sils_news_run4¶

Run ID: sils_news_run4
Participant: UNC_SILS
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: 0d0d09ac996453bdd78dcf9e8bc586a3
Run description: Spacy for NER

smith_base¶

Run ID: smith_base
Participant: Smith
Track: News
Year: 2019
Submission: 8/16/2019
Type: auto
Task: background
MD5: 9f79f42bc87c4e09935f2dceb149f87c
Run description: - Language Modeling Classifier -> BM25 Weighted Query - Date Information - Entropy & Clickbait Probabilities (https://github.com/bhargaviparanjape/clickbait)

smith_full¶

Run ID: smith_full
Participant: Smith
Track: News
Year: 2019
Submission: 8/16/2019
Type: auto
Task: background
MD5: 8a759c32ad5fd882816d83e9c6aa17d4
Run description: smith_base LTR model + keywords + poetry categories

smith_keyword¶

Run ID: smith_keyword
Participant: Smith
Track: News
Year: 2019
Submission: 8/16/2019
Type: auto
Task: background
MD5: 086961b155545edae97f3b1fdbf6937e
Run description: smith_base LTR model + Keyword extraction techniques (e.g., TextRank) -> SDM-BM25 weighted phrases

smith_poetry¶

Run ID: smith_poetry
Participant: Smith
Track: News
Year: 2019
Submission: 8/16/2019
Type: auto
Task: background
MD5: 097ee8545f72054d72f2cb7926b35999
Run description: - Language Modeling Classifier -> BM25 Weighted Query - Date Information - Poetry Classifications derived from PoetryFoundation.org toplevel categories. - Entropy & Clickbait Probabilities (https://github.com/bhargaviparanjape/clickbait)

UDInfolab_all¶

Run ID: UDInfolab_all
Participant: udel_fang
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: a266f91f9a008101269b9cec4da03957
Run description: The run used dbpedia-spotlight to conduct entity annotation on all documents in the collction. Entities were indexed as a single word. All words (terms and entities included) in a query document was used as the query. A simple time filter was used to ensure that only documents published before the query document could be retrieved.

UDInfolab_ent¶

Run ID: UDInfolab_ent
Participant: udel_fang
Track: News
Year: 2019
Submission: 8/18/2019
Type: auto
Task: background
MD5: e75b88a63e78282aabc2ecdccfa4a038
Run description: The run used dbpedia-spotlight to conduct entity annotation on all documents in the collction. Entities were indexed as a single word. Entities in a query document was weighted based on the kl-divergence between the language model of the whole document and the "context" language model of the entity, which was generated based on the words occur before and after the entities. Top 50 entities based on their weights as well as top 50 non-entity terms based on their occurrence in the query documet, were used as the query. A simple time filter was used to ensure that only documents published before the query document could be retrieved.

unh-trema-news¶

Run ID: unh-trema-news
Participant: TREMA-UNH
Track: News
Year: 2019
Submission: 9/23/2019
Type: auto
Task: entity
MD5: f9d78612282e83b6a60a3e2b96a70a59
Run description: For each paragraph in each article, DBpedia spotlight annotates entities in the paragraph. In this method, initially the query relevant feedback paragraphs are retrieved using BM25 retrieval method. URL given in the input file is considered as query here. A candidate entity list of all the entities annotated using DBpedia spotlight in the feedback paragraphs is generated. For every entity present in the candidate entity list, an entity-pair is created with every other entity present in the list. Check the existence of every entity-pair in the feedback paragraphs, if the pair is present then the rank of the paragraph is considered in calculating the score of the entity-pair. The score of an entity is calculated by taking the average of entity-pairs. Take the list of entities given in the input file and check if the generated list of entities are present in the input list. If it is present then take the score of that entity from the generated entity list else 0 as the final score of entity.

UQ_count¶

Run ID: UQ_count
Participant: UQ
Track: News
Year: 2019
Submission: 8/22/2019
Type: auto
Task: entity
MD5: 5a7e03ce8eca3480b03ebfc7b8dc81f4
Run description: The entities are ranked based on the number of entities' occurrences.

UQ_count_sent¶

Run ID: UQ_count_sent
Participant: UQ
Track: News
Year: 2019
Submission: 8/22/2019
Type: auto
Task: entity
MD5: 739f200e9663c99bf07d9bbfa7966a4e
Run description: The entities are ranked based on the number of entities' occurrences and the mean similarity scores which are derived by comparing the containing sentence with the whole document.

UQ_sent¶

Run ID: UQ_sent
Participant: UQ
Track: News
Year: 2019
Submission: 8/22/2019
Type: auto
Task: entity
MD5: 3b59c7d8e9661ee8511f2272b20b07b3
Run description: The entities are ranked based on the mean similarity scores which are derived by comparing the containing sentence with the whole document.

UQ_wiki¶

Run ID: UQ_wiki
Participant: UQ
Track: News
Year: 2019
Submission: 8/22/2019
Type: auto
Task: entity
MD5: 42e011391ab64ca7d36255086f6b9ee9
Run description: The entities are ranked based on the similarity score between the wikipedia representation and the whole document.

UQ_wiki_count¶

Run ID: UQ_wiki_count
Participant: UQ
Track: News
Year: 2019
Submission: 8/22/2019
Type: auto
Task: entity
MD5: 4826bf83df890a05b182ec6dc9f76235
Run description: The entities are ranked based on the similarity score between the wikipedia representation and the whole document combined with the sentences' score and entities' occurrences.

WHUirteam_run1¶

Results | Participants | Input | Summary | Appendix

Run ID: WHUirteam_run1
Participant: YQW2018CGroup
Track: News
Year: 2019
Submission: 8/18/2019
Type: manual
Task: background
MD5: 07400de58a1bf9c059d60fda402ebddd
Run description: Three features were used in, including the main title, the keywords of different paragraphs and the most important entities. And these results of the three features were sequenced by the merging results ranks of each one.

WHUirteam_run2¶

Results | Participants | Input | Summary | Appendix

Run ID: WHUirteam_run2
Participant: YQW2018CGroup
Track: News
Year: 2019
Submission: 8/18/2019
Type: manual
Task: background
MD5: 3a80885a57b1af3dcc41df815debdf10
Run description: Three features were the same as the WHUirteam_run1. Just the ranks were followed by the scores more than the positions of each one.