CFASL: Composite Factor-Aligned Symmetry Learning for Disentanglement in Variational AutoEncoder
Abstract: Symmetries of input and latent vectors have provided valuable insights for disentanglement learning in VAEs. However, only a few works were proposed as an unsupervised method, and even these works require known factor information in training data. We propose a
novel method, Composite Factor-Aligned Symmetry Learning (CFASL), which is integrated into VAEs for learning symmetry-based disentanglement in unsupervised learning without any knowledge of the dataset factor information. CFASL incorporates three novel features for learning symmetry-based disentanglement: 1) Injecting inductive bias to align latent vector dimensions to factor-aligned symmetries within an explicit learnable symmetry code-book 2) Learning a composite symmetry to express unknown factors change between two random samples by learning factor-aligned symmetries within the codebook 3) Inducing group equivariant encoder and decoder in training VAEs with the two conditions. In addition, we propose an extended evaluation metric for multi-factor changes in comparison to disentanglement evaluation in VAEs. In quantitative and in-depth qualitative analysis, CFASL demonstrates a significant improvement of disentanglement in single-factor change, and multi-factor change conditions compared to state-of-the-art methods.
Submission Length: Long submission (more than 12 pages of main content)
Supplementary Material: zip
Changes Since Last Submission: Revision: We left all the changes to the revised paper in the reviewers' comments and highlighted them.
Camera-ready version: We have addressed the final comments from reviewers AYog and 33dk. We have updated Fig. 9, fixed the layout issue on page 19 (final version), and split the Car3D images. Additionally, we have revised the notation.
Video: https://youtu.be/R03AoD3SRZ8
Code: https://github.com/GIST-IRR/CFASL
Assigned Action Editor: ~Bo_Dai1
Submission Number: 2991
Loading