Research map / Label-efficient and robust learning / Recognition and long-tail
Demystifying Inter-Class Disentanglement
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 |
|---|---|---|---|
| Use LORD instead of DrNet | class-content disentanglement · Cars3D | LPIPS: 0.095 -> 0.078 | Helped |
| Use LORD instead of DrNet | class-content disentanglement · SmallNorb | LPIPS: 0.166 -> 0.117 | Helped |
| Use LORD instead of DrNet | class-content disentanglement · SmallNorb-Poses | LPIPS: 0.152 -> 0.106 | Helped |
| Use LORD instead of ML-VAE | class-content disentanglement · CelebA | LPIPS: 0.222 -> 0.197 | Helped |
| Use LORD instead of StarGAN | facial expression translation · RaFD | facial expression classification error (%): 2.2 -> 1.8 | Helped |
| Add shared latent optimization instead of fully amortized inference (ablation) | LORD · Cars3D | LPIPS: 0.082 -> 0.078 | Helped |
| Add shared latent optimization instead of fully amortized inference (ablation) | LORD · Cars3D | classification accuracy of class labels from content codes: 0.92 -> < 0.01 | Helped |
| Add asymmetric noise instead of KL-divergence regularization (ablation) | amortized LORD · Cars3D | LPIPS: 0.094 -> 0.082 | Helped |
| Add asymmetric noise instead of KL-divergence regularization (ablation) | semi-amortized LORD · Cars3D | LPIPS: 0.095 -> 0.079 | Helped |
| Add the second-stage encoders (ablation) | LORD · Cars3D | LPIPS: 0.175 -> 0.078 | Helped |
| Add asymmetric noise regularization (ablation) | LORD · Cars3D | LPIPS: 0.095 -> 0.078 | Helped |
| Add shared latent optimization rather than amortized inference (ablation) | LORD · Cars3D | classification accuracy of content labels from class codes: 0.97 -> 0.01 | Helped |
Closest papers by meaning
- Disentangling Disentanglement in Variational Autoencoders (ICML 2019 · similarity 0.90)
- Disentanglement via Latent Quantization (NeurIPS 2023 · similarity 0.90)
- Orthogonality-Enforced Latent Space in Autoencoders: An Approach to Learning Disentangled Representations (ICML 2023 · similarity 0.89)
- Towards Principled Disentanglement for Domain Generalization (CVPR 2022 · similarity 0.89)
- Tripod: Three Complementary Inductive Biases for Disentangled Representation Learning (ICML 2024 · similarity 0.88)
- Disentangled Representation Learning via Flow Matching (NeurIPS 2026 · 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.