A critique of word similarity as a method for evaluating distributional semantic models
- Submitting institution
-
University of Sussex
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 78802_62044
- Type
- E - Conference contribution
- DOI
-
10.18653/v1/W16-2502
- Title of conference / published proceedings
- Proceedings of the 1st Workshop on Evaluating Vector-Space Representations for NLP
- First page
- 7
- Volume
- -
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- August
- Year of publication
- 2016
- URL
-
http://dx.doi.org/10.18653/v1/W16-2502
- 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
-
4
- Research group(s)
-
-
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- "Word similarity tasks are widely used in NLP as the primary way to measure and compare the quality of models of distributional lexical semantics. This paper raises important limitations associated with this approach, most significantly challenging the idea that a single notion of semantic similarity exists. The paper has had significant impact, being widely cited by high-profile papers, e.g [1,2], which are in turn highly cited.
[1] https://doi.org/10.18653/v1/S17-2002
[2] https://doi.org/10.1162/COLI_a_00301"
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -