Research map / Label-efficient and robust learning / Recognition and long-tail
Towards Building A Group-based Unsupervised Representation Disentanglement Framework
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 Isomorphism Loss to β-VAE | β-VAE · dSprites | DCI: 0.23 ± 0.10 -> 0.46 ± 0.085 | Helped |
| Add Isomorphism Loss to AnnealVAE | AnnealVAE · dSprites | BetaVAE score: 0.84 ± 0.050 -> 0.87 ± 0.0067 | Helped |
| Add Isomorphism Loss to β-TCVAE | β-TCVAE · Shapes3D | DCI: 0.66 ± 0.10 -> 0.72 ± 0.061 | Helped |
| Add Isomorphism Loss to AnnealVAE | AnnealVAE · Shapes3D | BetaVAE score: 0.82 ± 0.076 -> 0.89 ± 0.086 | Hurt |
| Add Isomorphism Loss to FactorVAE | FactorVAE · Shapes3D | BetaVAE score: 0.86 ± 0.055 -> 0.80 ± 0.075 | Hurt |
| Vary factor size n from 5 to 10 (ablation) | Groupified AnnealVAE · dSprites | DCI: 0.34 ± 0.062 -> 0.38 ± 0.055 | Helped |
| Add Abel Loss without Order Loss (ablation) | AnnealVAE · dSprites | DCI: 0.27 ± 0.10 -> 0.28 ± 0.11 | Helped |
| Add Order Loss without Abel Loss (ablation) | AnnealVAE · dSprites | DCI: 0.27 ± 0.10 -> 0.34 ± 0.056 | Helped |
| Add both Abel Loss and Order Loss (ablation) | AnnealVAE · dSprites | DCI: 0.27 ± 0.10 -> 0.38 ± 0.055 | Helped |
Closest papers by meaning
- Commutative Lie Group VAE for Disentanglement Learning (ICML 2021 · similarity 0.91)
- Multi-Level Variational Autoencoder: Learning Disentangled Representations From Grouped Observations (AAAI 2018 · similarity 0.90)
- Theory and Evaluation Metrics for Learning Disentangled Representations (ICLR 2020 · similarity 0.89)
- Demystifying Inductive Biases for (Beta-)VAE Based Architectures (ICML 2021 · similarity 0.88)
- $\alpha$TC-VAE: On the relationship between Disentanglement and Diversity (ICLR 2024 · similarity 0.88)
- Graph-based Unsupervised Disentangled Representation Learning via Multimodal Large Language Models (NeurIPS 2024 · similarity 0.86)
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.