Complex embeddings for simple link prediction
- Submitting institution
-
University College London
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 14141
- Type
- E - Conference contribution
- DOI
-
-
- Title of conference / published proceedings
- 33rd International Conference on Machine Learning, ICML 2016
- First page
- 3021
- Volume
- 5
- Issue
- -
- ISSN
- 2640-3498
- Open access status
- Compliant
- Month of publication
- June
- Year of publication
- 2016
- 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
-
4
- Research group(s)
-
-
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Previous work in link prediction has either been unable to model specific phenomena (e.g. asymmetry), used too many parameters for it or needed very complex architectures. This work addresses all these issues by taking the simplest existing model and operating it in complex vector space. This lead to state-of-the-art results at the time. Due to its simplicity and scalability it is still, today, one of the most used baselines in the field and very hard to beat.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -