A Commentary on the Unsupervised Learning of Disentangled Representations

Francesco Locatello, Stefan Bauer, Mario Lucic, Gunnar Rätsch, Sylvain Gelly, Bernhard Schölkopf, Olivier Bachem

AAAI 2020 pp. 13681-13684

doi:10.1609/AAAI.V34I09.7120 /aaai/2020/locatello2020aaai-commentary/

Abstract

The goal of the unsupervised learning of disentangled representations is to separate the independent explanatory factors of variation in the data without access to supervision. In this paper, we summarize the results of (Locatello et al. 2019b) and focus on their implications for practitioners. We discuss the theoretical result showing that the unsupervised learning of disentangled representations is fundamentally impossible without inductive biases and the practical challenges it entails. Finally, we comment on our experimental findings, highlighting the limitations of state-of-the-art approaches and directions for future research.

PDF AAAI Semantic Scholar

Cite

Text

Locatello et al. "A Commentary on the Unsupervised Learning of Disentangled Representations." AAAI Conference on Artificial Intelligence, 2020. doi:10.1609/AAAI.V34I09.7120

Markdown

[Locatello et al. "A Commentary on the Unsupervised Learning of Disentangled Representations." AAAI Conference on Artificial Intelligence, 2020.](https://mlanthology.org/aaai/2020/locatello2020aaai-commentary/) doi:10.1609/AAAI.V34I09.7120

BibTeX

@inproceedings{locatello2020aaai-commentary,
  title     = {{A Commentary on the Unsupervised Learning of Disentangled Representations}},
  author    = {Locatello, Francesco and Bauer, Stefan and Lucic, Mario and Rätsch, Gunnar and Gelly, Sylvain and Schölkopf, Bernhard and Bachem, Olivier},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2020},
  pages     = {13681-13684},
  doi       = {10.1609/AAAI.V34I09.7120},
  url       = {https://mlanthology.org/aaai/2020/locatello2020aaai-commentary/}
}