Research map / Probabilistic and scientific modeling / Variational and generative sampling
MaCow: Masked Convolutional Generative Flow
Xuezhe Ma, Xiang Kong, Shanghang Zhang, Eduard Hovy
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (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 |
|---|---|---|---|
| add fine-grained multi-scale architecture to MaCow (ablation) | MaCow · CIFAR-10 | bits/dim: 3.31 -> 3.28 | Improved |
| add fine-grained multi-scale architecture to MaCow (ablation) | MaCow · ImageNet 64x64 | bits/dim: 3.78 -> 3.75 | Improved |
| add variational dequantization to MaCow (ablation) | MaCow +fine-grained · CIFAR-10 | bits/dim: 3.28 -> 3.16 | Improved |
| add variational dequantization to MaCow (ablation) | MaCow +fine-grained · ImageNet 64x64 | bits/dim: 3.75 -> 3.69 | Improved |
| replace Glow with MaCow (uniform dequantization) | Glow · CIFAR-10 | bits/dim: 3.35 -> 3.28 | Improved |
| replace Glow with MaCow (uniform dequantization) | Glow · ImageNet 64x64 | bits/dim: 3.81 -> 3.75 | Improved |
| MaCow with variational dequantization vs Flow++ Var | MaCow · CIFAR-10 | bits/dim: 3.09 -> 3.16 | Hurt |
| MaCow with variational dequantization vs Flow++ Var | MaCow · ImageNet 64x64 | bits/dim: 3.69 -> 3.69 | No clear effect |
| variational instead of uniform dequantization (ablation) | MaCow · CelebA-HQ | bits/dim: 0.95 -> 0.67 | Improved |
| variational instead of uniform dequantization (ablation) | MaCow · LSUN bedroom | bits/dim: 1.16 -> 0.98 | Improved |
| variational instead of uniform dequantization (ablation) | MaCow · LSUN tower | bits/dim: 1.22 -> 1.02 | Improved |
| variational instead of uniform dequantization (ablation) | MaCow · LSUN church | bits/dim: 1.36 -> 1.09 | Improved |
Later papers that compared with this method
2 later accepted papers report MACow as a baseline in their results tables; 2 of them report a better number on most of the results they share with it.
- VeriFlow: Modeling Distributions for Neural Network Verification (AAAI 2026 · better on 7 of 7 shared results)
- Woodbury Transformations for Deep Generative Flows (NeurIPS 2020 · better on 3 of 3 shared results)
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 3 on most shared results.
- Glow — Glow: Generative Flow with Invertible 1x1 Convolutions (NeurIPS 2018) · better on 4 of 4 shared results
- PixelSNAIL — PixelSNAIL: An Improved Autoregressive Generative Model (ICML 2018) · better on 0 of 2 shared results
- PixelCNN++ — PixelCNN++: Improving the PixelCNN with Discretized Logistic Mixture Likelihood and Other Modifications (ICLR 2017) · better on 0 of 1 shared result
Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.
Closest papers by meaning
- Analysis of Learning a Flow-based Generative Model from Limited Sample Complexity (ICLR 2024 · similarity 0.89)
- Fixed-Point Masked Generative Modeling (NeurIPS 2026 · similarity 0.87)
- Plug-and-Play Context Feature Reuse for Efficient Masked Generation (NeurIPS 2025 · similarity 0.87)
- Mean Flows for One-step Generative Modeling (NeurIPS 2025 · similarity 0.87)
- Masked Autoencoders Are Effective Tokenizers for Diffusion Models (ICML 2025 · similarity 0.87)
- Beyond Masked and Unmasked: Discrete Diffusion Models via Partial Masking (NeurIPS 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.
