NASARI: integrating explicit knowledge and corpus statistics for a multilingual representation of concepts and entities
- Submitting institution
-
Cardiff University / Prifysgol Caerdydd
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 96998200
- Type
- D - Journal article
- DOI
-
10.1016/j.artint.2016.07.005
- Title of journal
- Artificial Intelligence
- Article number
- -
- First page
- 36
- Volume
- 240
- Issue
- -
- ISSN
- 0004-3702
- Open access status
- Deposit exception
- Month of publication
- August
- Year of publication
- 2016
- URL
-
https://doi.org/10.1016/j.artint.2016.07.005
- 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)
-
A - Artificial intelligence and data analytics
- Citation count
- 53
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- We present an effective method to overcome the inability of word embeddings to handle ambiguity (e.g. mouse can be an animal or a computer device). This article generalizes and extends two conference papers published at ACL and NAACL, and was awarded the 2016 best article prize in Computer Science at Sapienza University (Italy). The NASARI vector embeddings are publicly available (http://lcl.uniroma1.it/nasari/) and have been used for diverse applications such as sentiment analysis (https://doi.org/10.1016/j.eswa.2018.08.044), ontology learning (https://www.aclweb.org/anthology/L18-1034.pdf), disambiguation (https://doi.org/10.1016/j.knosys.2019.105030), computer vision (https://doi.org/10.1109/ICRA.2017.7989323), and multilingual similarity (https://openreview.net/forum?id=H196sainb; https://doi.org/10.18653/v1/D18-1330). NASARI is currently integrated into BabelNet, a large multilingual semantic network.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -