Compositional morphology for word representations and language modelling
- Submitting institution
-
University of Oxford
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 1935
- Type
- E - Conference contribution
- DOI
-
-
- Title of conference / published proceedings
- International Conference on Machine Learning, ICML 2014
- First page
- 1899
- Volume
- 32
- Issue
- 2
- ISSN
- 2640-3498
- Open access status
- Out of scope for open access requirements
- Month of publication
- January
- 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
-
1
- Research group(s)
-
-
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- This paper helped to initiate the study of large-scale language modelling with neural networks. It won a best paper award at ICML’14. It introduced the concept of compositionality in vector representations of the meaning and syntactic categories of words; for instance, the algorithm it introduced allowed the representations of 'cat', 'cats' and 'dogs' to all share common basis vectors, rather than being independent as they were in previous models. The paper’s contribution was discussed in Yoav Goldberg’s 2016 JAIR article on neural network models for NLP , and it remains influential (see e.g. recent work by Bojanowski et al., https://doi.org/10.1162/tacl_a_00051).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -