Research map / Label-efficient and robust learning / Recognition and long-tail
Discovering Interpretable Representations for Both Deep Generative and Discriminative Models
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 |
|---|---|---|---|
| Interpretable lens (ILVM) on top of an existing VAE latent instead of joint optimization | VAE latent space · MNIST | interpretability metric (%): 95.2 vs 89.8 (JLVM) and 83.3 (InfoGAN) | Helped |
| Joint optimization of reconstruction and interpretability (JLVM) instead of ILVM lens | latent space trained from scratch · SVHN | interpretability metric (%): 90.1 vs 85.7 (ILVM) and 83.9 (InfoGAN) | Helped |
| Joint optimization (JLVM) vs ILVM lens | latent space trained from scratch · 3D Chairs | interpretability metric (%): 89.8 vs 87.4 (ILVM) and 85.2 (InfoGAN) | Helped |
| Active learning acquisition (mutual information criterion) for choosing side information vs random acquisition and max entropy (ablation) | ILVM · MNIST and SVHN | interpretability metric vs training size | Helped |
| Interpretable lens applied to a hidden layer of a classifier network (MNIST-rot) rather than a VAE latent | hidden layer of discriminative network · MNIST-rot | qualitative similarity of orientation in latent | Helped |
Closest papers by meaning
- Restyling Unsupervised Concept Based Interpretable Networks with Generative Models (ICLR 2025 · similarity 0.89)
- Enhancing Pre-trained Representation Classifiability can Boost its Interpretability (ICLR 2025 · similarity 0.88)
- Enhancing Performance of Explainable AI Models with Constrained Concept Refinement (ICML 2025 · similarity 0.88)
- Interpretable Generative Models through Post-hoc Concept Bottlenecks (CVPR 2025 · similarity 0.86)
- Intrinsic User-Centric Interpretability through Global Mixture of Experts (ICLR 2025 · similarity 0.86)
- Predictive Concept Decoders: Training Scalable End-to-End Interpretability Assistants (NeurIPS 2026 · 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.