Research map / Probabilistic and scientific modeling / Variational and generative sampling
SurVAE Flows: Surjections to Bridge the Gap between VAEs and Flows
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 |
|---|---|---|---|
| Replace the initial sorting surjection with stochastic permutation layers | SortFlow on SpatialMNIST · SpatialMNIST | PPLL (per-point log-likelihood): -5.53 -> -5.30 | Helped |
| Use max pooling surjections instead of tensor slicing (ablation) | Image flow · CIFAR-10 | bits/dim: 3.08 -> 3.09 | Hurt |
| Use max pooling surjections instead of tensor slicing (ablation) | Image flow · ImageNet32 | bits/dim: 4.00 -> 4.01 | Hurt |
| Use max pooling surjections instead of tensor slicing (ablation) | Image flow · ImageNet64 | bits/dim: 3.70 -> 3.74 | Hurt |
| Add an absolute-value inference surjection | AbsFlow on symmetric synthetic 2D data · Synthetic symmetric 2D datasets | density modeling | Helped |
| Add a sorting surjection or stochastic permutation layer | Flow on exchangeable point-cloud data · SpatialMNIST | permutation invariance | Helped |
| Use max pooling surjections instead of tensor slicing (ablation) | Image flow · CIFAR-10, ImageNet32, ImageNet64 | Inception score and FID | Helped |
Closest papers by meaning
- Embedded-model flows: Combining the inductive biases of model-free deep learning and explicit probabilistic modeling (ICLR 2022 · similarity 0.88)
- Densely connected normalizing flows (NeurIPS 2021 · similarity 0.88)
- Woodbury Transformations for Deep Generative Flows (NeurIPS 2020 · similarity 0.88)
- Lifting Architectural Constraints of Injective Flows (ICLR 2024 · similarity 0.87)
- Normalizing Flows are Capable Generative Models (ICML 2025 · similarity 0.86)
- Kernelised Normalising Flows (ICLR 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.