Research map / Image generation / GANs
Deep Convolutional Sum-Product Networks
Cory J. Butz, Jhonatan S. Oliveira, André E. dos Santos, André L. Teixeira
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 Gaussian low-pass smoothing post-processing (ablation) | DCSPN · Olivetti Face left-completion | MSE: 455 -> 401 | Improved |
| train with differentiable MPE objective instead of NLL (ablation) | DCSPN · Olivetti Face left-completion | MSE: 455 -> 651 | Hurt |
| use height-by-width filters instead of 1-by-1 filters (no parameter sharing) (ablation) | DCSPN · image completion | MSE | Improved |
| use average-pooling, parameter sharing, batch norm, dropout (ablation) | DCSPN · image completion | MSE | No clear effect |
| keep sum instead of max in convolutional layers during MPE (ablation) | DCSPN · image completion | MSE | Improved |
| DCSPN vs DCGAN | DCGAN · Olivetti Face left-completion | MSE: 935 -> 455 | Improved |
| DCSPN vs DCGAN | DCGAN · Caltech Face bottom-completion | MSE: 1046 -> 1149 | Hurt |
| DCSPN vs DCGAN | DCGAN · Caltech Dolphin left-completion | MSE: 4096 -> 2002 | Improved |
Closest papers by meaning
- Doubly Convolutional Neural Networks (NeurIPS 2016 · similarity 0.87)
- Inductive Bias of Deep Convolutional Networks through Pooling Geometry (ICLR 2017 · similarity 0.86)
- Mixed Sum-Product Networks: A Deep Architecture for Hybrid Domains (AAAI 2018 · similarity 0.86)
- The Sum-Product Theorem: A Foundation for Learning Tractable Models (ICML 2016 · similarity 0.86)
- P-nets: Deep Polynomial Neural Networks (CVPR 2020 · similarity 0.86)
- Convolutions and More as Einsum: A Tensor Network Perspective with Advances for Second-Order Methods (NeurIPS 2024 · similarity 0.85)
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.
