Entity Query Feature Expansion Using Knowledge Base Links
- Submitting institution
-
University of Glasgow
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 11-04993
- Type
- E - Conference contribution
- DOI
-
10.1145/2600428.2609628
- Title of conference / published proceedings
- 37th International ACM SIGIR Conference on Research and Development in Information Retrieval
- First page
- 365
- Volume
- -
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- July
- Year of publication
- 2014
- URL
-
http://eprints.gla.ac.uk/171678/
- Supplementary information
-
-
- Request cross-referral to
- -
- Output has been delayed by COVID-19
- No
- COVID-19 affected output statement
- -
- Forensic science
- No
- Criminology
- No
- Interdisciplinary
- No
- Number of additional authors
-
2
- Research group(s)
-
-
- Citation count
- 72
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- ORIGINALITY: this work proposes new probabilistic models using entity knowledge graph features to improve the task of informational search. It is the first to use the web-scale entity link dataset, FACC1 released by Google, and the first to demonstrate significant improvements using entity representations, with particular improvements on ‘hard’ topics. Its experiments are rigorous with significant improvements on multiple standard TREC benchmark datasets. SIGNIFICANCE: now commonly used as a baseline method; has been very highly cited; contributed to a resurgence in using Knowledge Graphs for IR: a TREC track, KG4IR workshop series, and a IR Journal special issue.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -