Runs - Common Core 2018¶
anserini_ax¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_ax
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
e28756067300901c34c24702fd7d7fb4
- Run description: This run uses BM25 as the base ranking model and then applies Axiomatic Reranking as the feedback method.
anserini_ax17¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_ax17
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
c0540aaedc3ff565f2e67944b0dff600
- Run description: This run uses BM25 as the base ranking model and then applies Axiomatic Reranking as the feedback method. When doing the reranking we use Core 17 New York Times collection as the expansion terms pool.
anserini_bm25¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_bm25
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
aafbedfc2cf1113f6deab82a99ab595c
- Run description: This run uses BM25 as the ranking model.
anserini_ql¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_ql
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
21d13e2529a1f181908befe7f6ffa73e
- Run description: This run uses Dirichlet Language Model as the ranking model.
anserini_qlax¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_qlax
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
ddf7cee69bb3d2282935f863450706ad
- Run description: This run uses Dirichlet Language Model as the base ranking model and then applies Axiomatic Reranking as the feedback method.
anserini_qlax17¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_qlax17
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
dcce134e902decbd6c6d0d028454544b
- Run description: This run uses Dirichlet Language Model as the base ranking model and then applies Axiomatic Reranking as the feedback method. When doing the reranking we use Core 17 New York Times collection as the expansion terms pool.
anserini_qlrm3¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_qlrm3
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
1834a61218311eab214f61bef055c31e
- Run description: This run uses Dirichlet Language Model as the base ranking model and then applies RM3 Reranking model.
anserini_qlsdm¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_qlsdm
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
43d3f80e993bd7b67a1aee392e1b173b
- Run description: This run uses Dirichlet Language Model and we use the Sequential Dependency Model to model query topics. If the run or any of the information is incor
anserini_rm3¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_rm3
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
e331704ee2f7a14b44639ab914757d28
- Run description: This run uses BM25 as the base ranking model and then applies RM3 Reranking model.
anserini_sdm¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: anserini_sdm
- Participant: Anserini
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
07dd761131bcb60f58dc1aa4ce906c75
- Run description: This run uses BM25 and we use the Sequential Dependency Model to model query topics.
b-BoE¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: b-BoE
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
a34c6bed557505820e1bb6cc1651eada
- Run description: This run uses the body field of the documents, in a bag of entities document representation. BM25, Language Model with Dirichlet Smoothing, Two-Staged Smoothing, TFIDF, TF, IDF and Coordinate Match retrieval models were used to generate features, that were used to rerank the documents using Ranklib's Coordinate Ascent
b-BoW¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: b-BoW
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
b245b61c99479b82a9d4ec1d28284ad9
- Run description: This run uses the document body field in a bag of words document representation. BM25, Language Model with Dirichlet Smoothing, Two-Staged Smoothing, TFIDF, TF, IDF and Coordinate Match retrieval models were used to generate features, that were used to rerank the documents using Ranklib's Coordinate Ascent
b-BoW-t-BoWBoE¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: b-BoW-t-BoWBoE
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
3cb6b610efdedab00d16f27f621cbb53
- Run description: This run uses the document title field in a bag of words and bag of entities representation, and the body field in a bag of words document representation. BM25, Language Model with Dirichlet Smoothing, Two-Staged Smoothing, TFIDF, TF, IDF and Coordinate Match retrieval models were used to generate features, that were used to rerank the documents using Ranklib's Coordinate Ascent
b-BoWBoE-t-BoE¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: b-BoWBoE-t-BoE
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
977c66a5ee7497380e2dd70eaaa613fe
- Run description: This run uses the document body field in a bag of words and bag of entities representation, and the title field in a bag of entities document representation. BM25, Language Model with Dirichlet Smoothing, Two-Staged Smoothing, TFIDF, TF, IDF and Coordinate Match retrieval models were used to generate features, that were used to rerank the documents using Ranklib's Coordinate Ascent. Unlike Run 6, here the entities weren't boosted.
BM25-b-BoW¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: BM25-b-BoW
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
71e271be970385c4045b829f095cbf6a
- Run description: BM25 baseline over the document body field.
bt-BoE¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: bt-BoE
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
7a3ed3e12512f231897e36d0c0d23026
- Run description: This run uses both title and body fields of the documents, bag of entities document representation. BM25, Language Model with Dirichlet Smoothing, Two-Staged Smoothing, TFIDF, TF, IDF and Coordinate Match retrieval models were used to generate features, that were used to rerank the documents using Ranklib's Coordinate Ascent
bt-BoW¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: bt-BoW
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
4a8b15be096360bffdcf7703776ddd24
- Run description: This run uses both title and body fields of the documents, in a bag of words document representation. BM25, Language Model with Dirichlet Smoothing, Two-Staged Smoothing, TFIDF, TF, IDF and Coordinate Match retrieval models were used to generate features, that were used to rerank the documents using Ranklib's Coordinate Ascent
bt-BoWBoE¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: bt-BoWBoE
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
3125fb3f2eeb2e0c1ff89cd43607d76b
- Run description: This run uses both title and body fields of the documents, in a bag of words and bag of entities document representation. BM25, Language Model with Dirichlet Smoothing, Two-Staged Smoothing, TFIDF, TF, IDF and Coordinate Match retrieval models were used to generate features, that were used to rerank the documents using Ranklib's Coordinate Ascent
eb-boost¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: eb-boost
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
7e0db6a049debf9c27033cd75b0e8fa6
- Run description: This run uses the document body field in a bag of words and bag of entities representation, and the title field in a bag of entities document representation. BM25, Language Model with Dirichlet Smoothing, Two-Staged Smoothing, TFIDF, TF, IDF and Coordinate Match retrieval models were used to generate features, that were used to rerank the documents using Ranklib's Coordinate Ascent
feup-run1¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: feup-run1
- Participant: FEUP
- Track: Common Core
- Year: 2018
- Submission: 7/31/2018
- Type: automatic
- Task: main
- MD5:
60f800cb6e2640871d8ea1f639ef7800
- Run description: Hypergraph-of-entity with only term nodes and document hyperedges.
feup-run2¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: feup-run2
- Participant: FEUP
- Track: Common Core
- Year: 2018
- Submission: 7/31/2018
- Type: automatic
- Task: main
- MD5:
94d54394dbf142ed9487332d6c021b94
- Run description: Hypergraph-of-entity with term and entity nodes, and document, contained_in and related_to hyperedges, using external knowledge from DBpedia over named entities extracted from the first three paragraphs.
feup-run3¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: feup-run3
- Participant: FEUP
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
3c409b0afb066e883a393459f8653f10
- Run description: Hypergraph-of-entity with only term nodes and document hyperedges, reranked for diversity using a document profile based on the following features: Keywords, Named Entities, Sentiment Analysis, Reading Complexity, Emotion Categories.
feup-run4¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: feup-run4
- Participant: FEUP
- Track: Common Core
- Year: 2018
- Submission: 8/16/2018
- Type: automatic
- Task: main
- MD5:
483be35ad65b61e2218704c4bd8a963f
- Run description: Hypergraph-of-entity with term and entity nodes, and document, contained_in and related_to hyperedges, using external knowledge from DBpedia over named entities extracted from the first three paragraphs. Results are reranked for diversity using a document profile based on the following features: Sentiment Analysis, Reading Complexity. Limiting to 100 queries per topic.
feup-run5¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: feup-run5
- Participant: FEUP
- Track: Common Core
- Year: 2018
- Submission: 8/13/2018
- Type: automatic
- Task: main
- MD5:
1dad513190cb215414eb067af1e3a3a6
- Run description: Hypergraph-of-entity with only term nodes and document hyperedges, reranked for strong diversity using a document profile based on the following features: Keywords, Named Entities, Sentiment Analysis, Reading Complexity, Emotion Categories.
feup-run6¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: feup-run6
- Participant: FEUP
- Track: Common Core
- Year: 2018
- Submission: 8/16/2018
- Type: automatic
- Task: main
- MD5:
96cc9297e0cf26f0332aeaaebf91b4eb
- Run description: Hypergraph-of-entity with term and entity nodes, and document, contained_in and related_to hyperedges, using external knowledge from DBpedia over named entities extracted from the first three paragraphs. Results are reranked for diversity using a document profile based on the following features: Sentiment Analysis, Reading Complexity, Emotion Categories. Limiting to 100 queries per topic.
feup-run7¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: feup-run7
- Participant: FEUP
- Track: Common Core
- Year: 2018
- Submission: 8/16/2018
- Type: automatic
- Task: main
- MD5:
dd38d543cdf038069ace1ede3bb68978
- Run description: Hypergraph-of-entity with only term nodes and document hyperedges, reranked for diversity using a document profile based on the following features: Sentiment Analysis, Reading Complexity.
feup-run8¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: feup-run8
- Participant: FEUP
- Track: Common Core
- Year: 2018
- Submission: 8/16/2018
- Type: automatic
- Task: main
- MD5:
3c3b86f8d53f4809f3f7ada39f8b3ef1
- Run description: Hypergraph-of-entity with only term nodes and document hyperedges, reranked for diversity using a document profile based on the following features: Named Entities, Sentiment Analysis, Reading Complexity, Emotion Categories. Including all possible values and weights of the features.
h2oloo_e7ax0.6¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_e7ax0.6
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: relauto
- Task: main
- MD5:
b887372d196d65ee9a7e4c09e6fda75b
- Run description: Use 7 classifiers to rerank first 25 topics, BM25+axiom to run the rest 25 topics.
h2oloo_e7ax0.7¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_e7ax0.7
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: relauto
- Task: main
- MD5:
3ebf4cf659162b421ce7fcc49e787fef
- Run description: Use 7 classifiers to rerank first 25 topics, BM25+axiom to run the rest 25 topics.
h2oloo_e7rm30.6¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_e7rm30.6
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: relauto
- Task: main
- MD5:
3dea04bb4f81f2921a1976d07de08bbf
- Run description: Use 7 classifiers to run on the first 25 topics, BM25+RM3 on the rest 25 topics
h2oloo_e7rm30.7¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_e7rm30.7
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: relauto
- Task: main
- MD5:
5838b2fc780b922cef97fb7aceb15003
- Run description: Use 7 classifiers to run on the first 25 topics, BM25+RM3 on the rest 25 topics
h2oloo_enax0.6¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_enax0.6
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/16/2018
- Type: relauto
- Task: main
- MD5:
218f71e845a440fda659cda7f6eb5836
- Run description: We used documents for same 25 topic in Robust04, Robust05, and Core17 to train classifiers.
h2oloo_enax0.7¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_enax0.7
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/16/2018
- Type: relauto
- Task: main
- MD5:
732acacab7452f9578886efec1db10ea
- Run description: We used documents for same 25 topics in Robust04, Robust05, and Core17 to train classifier and take the ensemble
h2oloo_enrm30.6¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_enrm30.6
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: relauto
- Task: main
- MD5:
f48cd8f2ca3f412d047d359236a32378
- Run description: Train 3 classifiers and ensemble on first 25 topics. BM25+RM3 for the rest 25 topics
h2oloo_LR2_rm3¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_LR2_rm3
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/16/2018
- Type: relauto
- Task: main
- MD5:
0f321ade7a3af86f3c5e4e1f64ee4aa9
- Run description: This single model was trained on 25 topics in Robust04, Robust05, Core17 with LR and rerank on BM25+RM3.
h2oloo_LR2AX0.6¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_LR2AX0.6
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/16/2018
- Type: relauto
- Task: main
- MD5:
169adb32e5aadd006b0c1a2612dd0063
- Run description: This submission contains the single model that was trained from Robust04, Robust05, Core17 over 25 topics, for the left 25 new topics, the results are just BM25+AX.
h2oloo_LRax0.6¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: h2oloo_LRax0.6
- Participant: h2oloo
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: relauto
- Task: main
- MD5:
9bd652f7b7f1a485fb49cd560bf7ae7a
- Run description: Use Logistic Regression to train on the same topics in Robust04, Robust05, and Core17. Use BM25+axiom in the other topics.
jarir_cb_ind¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: jarir_cb_ind
- Participant: JARIR
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: automatic
- Task: main
- MD5:
f73b65e0c8224d16a42651c658024d36
- Run description: A query expansion method based on natural language processing and word embedding. Word2Vec's Cbow model was trained on the Washington Post Corpus with a vector size equal to 300. Then we selected expansion terms that are similar to the individual query terms. In this run, we used only the titles of queries. The Okapi BM25 ranking model is applied with default parameters.
jarir_cb_re¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: jarir_cb_re
- Participant: JARIR
- Track: Common Core
- Year: 2018
- Submission: 8/19/2018
- Type: automatic
- Task: main
- MD5:
0e856c66133cd0f0a49e4024b1ef4144
- Run description: A query expansion method based on natural language processing and word embedding. Word2Vec's Cbow model was trained on the Washington Post Corpus with a vector size equal to 300. Then we selected expansion terms that are similar to the entire query. The new query terms are reweighed. In this run, we used only the titles of queries. The Okapi BM25 ranking model is applied with default parameters.
jarir_cbow¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: jarir_cbow
- Participant: JARIR
- Track: Common Core
- Year: 2018
- Submission: 8/19/2018
- Type: automatic
- Task: main
- MD5:
772e683585e6c0cc71df63d1aefaeb53
- Run description: A query expansion method based on natural language processing and word embedding. Word2Vec's Cbow model was trained on the Washington Post Corpus with a vector size equal to 300. Then we selected expansion terms that are similar to the entire query. In this run, we used only the titles of queries. The Okapi BM25 ranking model is applied with default parameters.
jarir_sg_ind¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: jarir_sg_ind
- Participant: JARIR
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: automatic
- Task: main
- MD5:
233176c23dc6a448f9d1726f515696c6
- Run description: A query expansion method based on natural language processing and word embedding. Word2Vec's Skip-Gram model was trained on the Washington Post Corpus with a vector size equal to 300. Then we selected expansion terms that are similar to the individual query terms. In this run, we used only the titles of queries. The Okapi BM25 ranking model is applied with default parameters.
jarir_sg_re¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: jarir_sg_re
- Participant: JARIR
- Track: Common Core
- Year: 2018
- Submission: 8/19/2018
- Type: automatic
- Task: main
- MD5:
0225a6bb265876deb5fd89fb54b69810
- Run description: A query expansion method based on natural language processing and word embedding. Word2Vec's Skip-Gram model was trained on the Washington Post Corpus with a vector size equal to 300. Then we selected expansion terms that are similar to the entire query. The new query terms are reweighed. In this run, we used only the titles of queries. The Okapi BM25 ranking model is applied with default parameters.
jarir_skipgram¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: jarir_skipgram
- Participant: JARIR
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: automatic
- Task: main
- MD5:
9a77a09ec574aae7a59fe462ef0faade
- Run description: A query expansion method based on natural language processing and word embedding. Word2Vec's Skip-Gram model was trained on the Washington Post Corpus with a vector size equal to 300. Then we selected expansion terms that are similar to the entire query. In this run, we used only the titles of queries. The Okapi BM25 ranking model is applied with default parameters.
RMITEXTGIGADA5¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: RMITEXTGIGADA5
- Participant: RMIT
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: relauto
- Task: main
- MD5:
cf019907e4e7b1ccd747f8d8f5bc2b7a
- Run description: External query expansion using the Gigaword + Tipster corpus.
RMITFDA4¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: RMITFDA4
- Participant: RMIT
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: relauto
- Task: main
- MD5:
13d0e12d6815dd90169a0af37ac0779b
- Run description: Title query runs on Indri and Terrier with query expansion, and external expansion runs from Gigaword and Tipster fused into a single run using RRF. Query expansion parameters taken from NYT judgments (collection-wide, not per-topic). A baseline for how query variations compare to titles.
RMITUQVBestDM2¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: RMITUQVBestDM2
- Participant: RMIT
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: manual
- Task: main
- MD5:
311e877001c11893f700fd8bace81a21
- Run description: Authors formed a judgment pool to the top-5 of RMITUQVDBFDM3 and a title query language model run. These judgments were used to select the best title-only query without fusion using the same systems as in RMITUQVDBFDM3.
RMITUQVDBFDM3¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: RMITUQVDBFDM3
- Participant: RMIT
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: manual
- Task: main
- MD5:
dfc3526dae0803858c37e708279470c9
- Run description: Query variations for the original TREC topics were generated by the authors. All query variations were run on systems with parameters shown to be effective on NYT using Indri and Terrier with query expansion, as well as external corpus query expansion using Gigaword and Gigaword+Tipster. This was fused to make a single run using RRF k=60. Documents found with duplicates were included in-place in the ranked list.
RMITUQVDBFNZDM1¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: RMITUQVDBFNZDM1
- Participant: RMIT
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: manual
- Task: main
- MD5:
d7c2e97f39eca796c0e5ddcd392b11e9
- Run description: Authors formed a judgment pool to the top-5 of RMITUQVDBFDM3 and a title query language model run. These judgments were used to remove any query variations with a zero score prior to rank fusion. The reduced set of queries using RMITUQVDBFDM3's approach, where documents found to have duplicates were included in-place in the ranked list.
sab18aqv45nytE1¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sab18aqv45nytE1
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
c87e39450ec777b17877dd3169fb35a9
- Run description: expansion by all terms occurring in known relevant documents of aquaint,v45,nyt.
sab18aqv45nytO1¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sab18aqv45nytO1
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/22/2018
- Type: relauto
- Task: main
- MD5:
a7bcdd494ea3f2f8e24f1129a98ba534
- Run description: optimization with 250 terms occurring in known relevant documents of aquaint,v45,nyt.
sab18coreA¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sab18coreA
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: automatic
- Task: main
- MD5:
8539c640d3b7c97787421c36823c1453
- Run description: Standard SMART Lnu.ltu weighted vector run
sab18coreE1¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sab18coreE1
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
28164589ff674db7da181e490ae56acf
- Run description: Massive Rocchio feedback expansion by all terms occurring in known relevant documents of v45. Exact duplicate of run done in Robust and Core 2017.
sab18coreO1¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sab18coreO1
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
805ef351d44ef2b5e39da4c154aa0575
- Run description: optimization of weights for 250 terms occurring in known relevant documents of v45. Exact duplicate of run done in Robust and Core 2017.
sab18nyt_E1¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sab18nyt_E1
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
50f72fda273f182167a7bd7a9bd58f99
- Run description: massive expansion run with all terms occurring in known relevant documents of nyt (Core 2017).
sab18nyt_O1¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sab18nyt_O1
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
d753057baaec3b2f053eb13ac34314c2
- Run description: optimization of 250 terms occurring in known relevant documents of nyt (Core 2017).
sab18nyt_O1.50¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sab18nyt_O1.50
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
d4ac31ef4d75ee09ece87d96cd8093a6
- Run description: optimization of 50 terms occurring in known relevant documents of nyt (Core 2017).
sabaqv45nytO150¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sabaqv45nytO150
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
562abc59c6922686335698e3163da1b8
- Run description: optimization of 50 terms occurring in known relevant documents of aquaint,v45,nyt.
sabopt50v45¶
Results
| Participants
| Input
| Summary
| Appendix
- Run ID: sabopt50v45
- Participant: Sabir
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
689680f1ae9029774f787fed3c8c71d1
- Run description: optimization of weights for 250 terms occurring in known relevant documents of v45. Exact duplicate of run done in Robust and Core 2017.
t-BoE¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: t-BoE
- Participant: NOVASearch
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: relauto
- Task: main
- MD5:
d7e9386dcecac9715a34282fec36df20
- Run description: This run uses the title field of the document, in a bag of entities document representation. BM25, Language Model with Dirichlet Smoothing, Two-Staged Smoothing, TFIDF, TF, IDF and Coordinate Match retrieval models were used to generate features, that were used to rerank the documents using Ranklib's Coordinate Ascent
umass_bsdm¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: umass_bsdm
- Participant: UMass
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: automatic
- Task: main
- MD5:
4358af5872163b294dc5bbc538e30e16
- Run description: BSDM model is a Sequential Dependence Model built on BM25 scoring.
umass_cbsdm¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: umass_cbsdm
- Participant: UMass
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: automatic
- Task: main
- MD5:
575442362009dd51e06d5a1608cdfbac
- Run description: CBSDM model is a Sequential Dependence Model built on BM25 scoring with cheaper (more-efficient) feature functions than the traditional model and smart stopword removal (in unigrams and unordered features) using the inquery stopword list.
umass_ql¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: umass_ql
- Participant: UMass
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: automatic
- Task: main
- MD5:
caaaa34b0f1eab84ca11c2daf1c79155
- Run description: umass_ql is a straight-up implementation of the Query Likelihood Model.
umass_sdm¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: umass_sdm
- Participant: UMass
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: automatic
- Task: main
- MD5:
e116906fb5f3b3066f2d822957a73784
- Run description: umass_sdm is a straight-up implementation of the Sequential Dependence Model.
UWaterMDS_DS_A¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UWaterMDS_DS_A
- Participant: UWaterlooMDS
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: manual
- Task: main
- MD5:
01033c97a160fc732c02261fc28e51c7
- Run description: Manual search and judgments are initially used to find dozens of relevant documents. Then dynamic sampling is applied to select 300 sampled documents for human assessors to judge. The final machine learner ranks the judged relevant documents according to their relevance scores from high to low.
UWaterMDS_DS_B¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UWaterMDS_DS_B
- Participant: UWaterlooMDS
- Track: Common Core
- Year: 2018
- Submission: 8/20/2018
- Type: manual
- Task: main
- MD5:
5a5b682b510ce46a755ef7a85a38e296
- Run description: Manual search and judgments are initially used to find dozens of relevant documents. Then dynamic sampling is applied to select 300 sampled documents for human assessors to judge. The final machine learner ranks the judged relevant documents according to their relevance scores from low to high.
UWaterMDS_Rank¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UWaterMDS_Rank
- Participant: UWaterlooMDS
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: manual
- Task: main
- MD5:
909920b948fe0dc1c1c182d0207b190e
- Run description: Build a machine learner model from the judgments of ad-hoc search and dynamic sampling. Rank all the documents using the model according to their relevance scores.
UWaterMDS_SEQ¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: UWaterMDS_SEQ
- Participant: UWaterlooMDS
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: manual
- Task: main
- MD5:
4e7203f54535c82da20b65138e5043f4
- Run description: The run is generated according to the order of our judgments. The earlier judged documents are ranked prior than the other documents.
uwmrg¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: uwmrg
- Participant: MRG_UWaterloo
- Track: Common Core
- Year: 2018
- Submission: 8/12/2018
- Type: automatic
- Task: main
- MD5:
a4ddddaae38b61f73bdd4caf0d70bb94
- Run description: Logistic Regression scoring, based on (up to) 11 pseudo-relevant documents: Google SERP page and top-10 documents linked-to by SERP page. For each topic, the pseudo-relevant documents for the other 49 topics were used as pseudo-nonrelevant documents.
uwmrgx¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: uwmrgx
- Participant: MRG_UWaterloo
- Track: Common Core
- Year: 2018
- Submission: 8/12/2018
- Type: automatic
- Task: main
- MD5:
d1962acdcc2fb51a121493115703048e
- Run description: Logistic Regression scoring, based on 1 pseudo-relevant document: Google SERP containing ten blue links and summaries. For each topic, the pseudo-relevant documents for the other 49 topics were used as pseudo-nonrelevant documents. This run did not use the linked-to documents, just the Google SERP.
webis-argument¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: webis-argument
- Participant: Webis
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: automatic
- Task: main
- MD5:
a5e647d9be8cf65f339df7b8886ca090
- Run description: - argumentation re-ranking with a higher priority for the queries demanding argumentation in arguments - argumentation extractor - textrank summarization
webis-baseline¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: webis-baseline
- Participant: Webis
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: automatic
- Task: main
- MD5:
f1cc46b8a406b7ded4fb35d8fc5e2330
- Run description: - ranking with no argumentation priority - multi-field ranking
webis-bm25f¶
Results
| Participants
| Proceedings
| Input
| Summary
| Appendix
- Run ID: webis-bm25f
- Participant: Webis
- Track: Common Core
- Year: 2018
- Submission: 8/21/2018
- Type: automatic
- Task: main
- MD5:
028ae72a093b56af56052ae444635dcf
- Run description: - argumentative re-ranking with a low argumentation priority - multi-field ranking function - textrank