Latent Space Factorisation and Manipulation via Matrix Subspace Projection
- Submitting institution
-
The University of Sheffield
- Unit of assessment
- 11 - Computer Science and Informatics
- Output identifier
- 7857
- Type
- E - Conference contribution
- DOI
-
-
- Title of conference / published proceedings
- Proceedings of the 37th International Conference on Machine Learning, PMLR
- First page
- 5916
- Volume
- 119
- Issue
- -
- ISSN
- 2640-3498
- Open access status
- Not compliant
- Month of publication
- June
- Year of publication
- 2020
- 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)
-
D - Natural Language Processing
- Citation count
- -
- Proposed double-weighted
- No
- Reserve for an output with double weighting
- No
- Additional information
- Published in the A* CORE Ranked ICML, this is the first, universal approach for disentangling many attributes simultaneously, applicable to arbitrary types of autoencoders. The work has been taken up by colleagues in Shanghai Jiao Tong University (contact: Professor Institute of Robotics) for building models for object pose tracking. Lin has given invited talk about this work at the University of Tokyo and the University of Southampton.
- Author contribution statement
- -
- Non-English
- No
- English abstract
- -