Research map / Probabilistic and scientific modeling / Variational and generative sampling
Invertible Convolutional Flow
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 circular or symmetric convolution coupling flows instead of affine coupling/autoregressive flows | CONF · CIFAR10 (FC conditioning) | bits/dim: 4.31 -> 3.78 | Helped |
| use symmetric convolution instead of circular convolution (ablation) | CONF · Density estimation (Table 1) | NLL: MNIST 1.26 vs 1.25; CIFAR10 3.78 vs 3.82 | Mixed |
| use CNN conditioning networks with CONF versus Glow | S-CONF · CIFAR10 (CNN conditioning) | bits/dim: 3.35 -> 3.34 | Helped |
| use CONF as VAE flow with larger size (F=16, M=16) versus smaller flows (ablation) | VAE on Frey Faces · Frey Faces | NLL: 4.31 -> 4.29 | Helped |
| CONF(4-8) versus larger O-SNF(16-32) | VAE MNIST · MNIST | NLL: 80.64 vs 80.22 | Mixed |
Closest papers by meaning
- Bidirectional Normalizing Flow: From Data to Noise and Back (CVPR 2026 · similarity 0.92)
- Woodbury Transformations for Deep Generative Flows (NeurIPS 2020 · similarity 0.91)
- CDFlow: Building Invertible Layers with Circulant and Diagonal Matrices (NeurIPS 2025 · similarity 0.89)
- Representational aspects of depth and conditioning in normalizing flows (ICML 2021 · similarity 0.89)
- STARFlow: Scaling Latent Normalizing Flows for High-resolution Image Synthesis (NeurIPS 2025 · similarity 0.87)
- Normalizing Flows are Capable Generative Models (ICML 2025 · similarity 0.87)
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.