Word Embedding as Maximum A Posteriori Estimation
- Submitting institution
-
The University of Kent
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 14915
- Type
- E - Conference contribution
- DOI
-
10.1609/aaai.v33i01.33016562
- Title of conference / published proceedings
- Proceedings of the AAAI Conference on Artificial Intelligence
- First page
- 6562
- Volume
- 33
- Issue
- 1
- ISSN
- 2374-3468
- Open access status
- Compliant
- Month of publication
- July
- Year of publication
- 2019
- URL
-
https://kar.kent.ac.uk/70009/
- 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
- 1
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper contains a novel model that uses priors when learning word embeddings. It is significant because it has the advantage of encoding the level of the informativeness that a word can convey. Previous methods rely on raw count data which are prone to noise. Furthermore, automatic parameter inference makes our method practically applicable because a user does not have to understand the technicalities of the model. Our experimental results, across three varied standard evaluation tasks, show that our model consistently outperforms several major state-of-the-art methods and popular baselines.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
