Evaluating Neural Word Representations in Tensor-Based Compositional Settings
- Submitting institution
-
Queen Mary University of London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 401
- Type
- E - Conference contribution
- DOI
-
10.3115/v1/d14-1079
- Title of conference / published proceedings
- Proceedings of the 2014 Conference on Empirical Methods in Natural Language Processing (EMNLP)
- First page
- 708
- Volume
- abs/1408.6179
- Issue
- -
- ISSN
- -
- Open access status
- -
- Month of publication
- October
- Year of publication
- 2014
- URL
-
-
- 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
-
3
- Research group(s)
-
-
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Showed that vector-space models of meaning based in interpretable statistics can compete with state-of-the-art models learned using neural networks. Developed in the ConCreTe project (EU-FP7 611733 ?3.2m 2013-16) into a model for understanding context-dependent, figurative & metaphorical language, leading to 4 journal papers (JAGI 2015, ACM-CSUR 2019, Frontiers 2016/2019), 3 conference papers (INLG 2016, IWCS 2017/2019) and supporting workshop papers. This led to further follow-on funding (EMBEDDIA, EU-H2020 825153 ?3m 2019-21) and collaboration with IJS Slovenia via the FORMICA project 2016-18 (http://kt.ijs.si/proj, senja.pollak@ijs.si). Two associated PhDs now on research careers: Milajevs at NIST (https://scholar.google.com/citations?user=CScje3kAAAAJ), McGregor in industry at Action.AI (stephen@action.ai).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
