Abstract
We tackle the problem disentangling the latent space of an autoencoder in order to separate labelled attribute information from other characteristic information. This then allows us to change selected attributes while preserving other information. Our method, matrix subspace projection, is much simpler than previous approaches to latent space factorisation, for example not requiring multiple discriminators or a careful weighting among their loss functions. Furthermore our new model can be applied to autoencoders as a plugin, and works across diverse domains such as images or text. We demonstrate the utility of our method for attribute manipulation in autoencoders trained across varied domains, using both human evaluation and automated methods. The quality of generation of our new model (e.g. reconstruction, conditional generation) is highly competitive to a number of strong baselines.
Original language | English |
---|---|
Title of host publication | 37th International Conference on Machine Learning, ICML 2020 |
Editors | Hal Daume, Aarti Singh |
Publisher | International Machine Learning Society (IMLS) |
Pages | 5872-5882 |
Number of pages | 11 |
ISBN (Electronic) | 9781713821120 |
Publication status | Published - 2020 |
Event | 37th International Conference on Machine Learning, ICML 2020 - Virtual, Online Duration: 13 Jul 2020 → 18 Jul 2020 |
Conference
Conference | 37th International Conference on Machine Learning, ICML 2020 |
---|---|
City | Virtual, Online |
Period | 13/07/20 → 18/07/20 |
Bibliographical note
Funding Information:We would like to thank all the anonymous reviewers for their insightful comments. This work is supported by the award made by the UK Engineering and Physical Sciences Research Council (Grant number: EP/P011829/1).
Publisher Copyright:
© International Conference on Machine Learning, ICML 2020. All rights reserved.
Keywords
- cs.LG
- stat.ML