A deep matrix factorization method for learning attribute representations
- Submitting institution
-
Imperial College of Science, Technology and Medicine
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 2392
- Type
- D - Journal article
- DOI
-
10.1109/TPAMI.2016.2554555
- Title of journal
- IEEE Transactions on Pattern Analysis and Machine Intelligence
- Article number
- -
- First page
- 417
- Volume
- 39
- Issue
- 3
- ISSN
- 0162-8828
- Open access status
- Out of scope for open access requirements
- Month of publication
- April
- Year of publication
- 2016
- URL
-
-
- Supplementary information
-
10.1109/TPAMI.2016.2554555
- 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
- 91
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Our Deep Semi-NMF automatically learns a hierarchy of attributes of a given dataset and outperforms current Semi-NMFs on three standard image processing datasets. The Deep Semi-NMF allows for efficient one-pass signal-based deep representation learning - a major breakthrough in holistic learning. It is exploited in the ARIA-VALUSPA Horizon 2020 project (€2.9M) where Schuller and his team are partners and the ACLEW transatlantic digging in the data project (http://bit.ly/3hWLOFI; £120K). This publication is an extension of our highly-cited ICML 2014 paper (http://proceedings.mlr.press/v32/trigeorgis14.html).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -
