Research map / Probabilistic and scientific modeling / Variational and generative sampling
Identifying through Flows for Recovering Latent Representations
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 iFlow's flow-based conditional marginal-likelihood objective instead of iVAE's ELBO objective | Identifiable latent representation learning · Synthetic non-stationary Gaussian sources | MCC | Helped |
| Use iFlow's flow-based conditional marginal-likelihood objective instead of iVAE's ELBO objective | Identifiable latent representation learning · Synthetic non-stationary Gaussian sources | log likelihood (energy value) | Helped |
| Use Softplus for the natural-parameter nonlinearity (ablation) | iFlow · Synthetic non-stationary Gaussian sources | MCC | Helped |
| Increase iVAE depth from 3 to 4, 5, or 6 layers (ablation) | iVAE · Synthetic data, averaged MCC | MCC: 0.4964(\pm 0.0723) -> 0.4782(\pm 0.0740); 0.4521(\pm 0.0719); 0.4276(\pm 0.0686) | Hurt |
| Increase iVAE hidden units from 50 to 100 or 200 while keeping 3 layers (ablation) | iVAE · Synthetic data, averaged MCC | MCC: 0.4964(\pm 0.0723) -> 0.4965(\pm 0.0693); 0.4961(\pm 0.0753) | No clear effect |
Closest papers by meaning
- Identifiability of deep generative models without auxiliary information (NeurIPS 2022 · similarity 0.93)
- Posterior Collapse and Latent Variable Non-identifiability (NeurIPS 2021 · similarity 0.90)
- Beyond ICA: Identifiability by Symmetry Breaking (NeurIPS 2026 · similarity 0.87)
- Multi-View Causal Representation Learning with Partial Observability (ICLR 2024 · similarity 0.87)
- Synergy Between Sufficient Changes and Sparse Mixing Procedure for Disentangled Representation Learning (ICLR 2025 · similarity 0.86)
- End-to-End Identifiable and Consistent Recurrent Switching Dynamical Systems (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.