Semi-supervised domain adaptation via Fredholm integral based kernel methods
- Submitting institution
-
Royal Holloway and Bedford New College
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 37893579
- Type
- D - Journal article
- DOI
-
10.1016/j.patcog.2018.07.035
- Title of journal
- Pattern Recognition
- Article number
- -
- First page
- 185
- Volume
- 85
- Issue
- -
- ISSN
- 0031-3203
- Open access status
- Deposit exception
- Month of publication
- August
- Year of publication
- 2018
- 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
- 14
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- In this work, we propose the first semi-supervised domain adaptation algorithm. It allows a model trained in one domain being easily adapted to another related domain. We show its superior data efficiency and performance in multiple benchmark datasets. The proposed method is widely regarded as the state-of-the-art semi-supervised domain adaptation method (c.f., Mahapatra and Ge, Pattern Recognition, 2020) and inspires follow-up semi-supervised transfer learning methods (e.g., Saboori and Ghassemian, INT J REMOTE SENSING, 2020). An earlier version of this work was presented at AAAI 2015, and the extended version is published on Pattern Recognition Journal (IF 5.9).
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -