Research map / Label-efficient and robust learning / Recognition and long-tail
Where and What? Examining Interpretable Disentangled Representations
Xinqi Zhu, Chang Xu, Dacheng Tao
Check your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| Add PS loss and SC module to InfoGAN (ablation) | InfoGAN · CelebA | TPL: 10.9 -> 8.1 | Improved |
| Add PS loss to InfoGAN (ablation) | InfoGAN · CelebA | PPL: 43.6 -> 34.3 | Improved |
| Add SC module to InfoGAN (ablation) | InfoGAN · CelebA | FID: 6.0 -> 5.9 | Improved |
| Add PS loss and SC module to InfoGAN (ablation) | InfoGAN · FFHQ | PPL: 142.7 -> 120.1 | Improved |
| Replace SC mask with softmax mask (ablation) | InfoGAN + Both · CelebA | TPL: 8.1 -> 14.8 | Hurt |
| Set PS-loss weight to 0.1 (ablation) | InfoGAN + Both · FFHQ | FID: 13.5 -> 16.0 | Hurt |
| Set PS-loss weight to 0.001 (ablation) | InfoGAN + Both · FFHQ | FID: 13.5 -> 23.7 | Hurt |
| Use only 1 rectangle in SC masks (ablation) | InfoGAN + Both · CelebA | FID: 6.0 -> 9.0 | Hurt |
| Use 4 rectangles in SC masks (ablation) | InfoGAN + Both · CelebA | FID: 6.0 -> 5.9 | Improved |
| Add SC module on DSprites (ablation) | Ours · DSprites | FactorVAE metric: 83.54 (6.91) -> 81.47 (5.39) | Hurt |
| Select top training seeds using TPL | Ours · 3DShapes | FactorVAE metric: 92.04 (6.49) -> 93.41 (3.34) | Improved |
| Use the full method rather than InfoGAN | InfoGAN · Shoes+Edges | PPL: 2952.2 -> 246.2 | Improved |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- InfoGAN — InfoGAN: Interpretable Representation Learning by Information Maximizing Generative Adversarial Nets (NeurIPS 2016) · better on 4 of 6 shared results
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- A Sober Look at the Unsupervised Learning of Disentangled Representations and their Evaluation (JMLR 2020 · similarity 0.90)
- Challenging Common Assumptions in the Unsupervised Learning of Disentangled Representations (ICML 2019 · similarity 0.90)
- Theory and Evaluation Metrics for Learning Disentangled Representations (ICLR 2020 · similarity 0.90)
- Transferring disentangled representations: bridging the gap between synthetic and real images (NeurIPS 2024 · similarity 0.89)
- Factorized Diffusion Autoencoder for Unsupervised Disentangled Representation Learning (AAAI 2024 · similarity 0.88)
- Revisiting Disentanglement in Downstream Tasks: A Study on Its Necessity for Abstract Visual Reasoning (AAAI 2024 · similarity 0.88)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
