Research map / Image generation / GANs
Information-Theoretic GAN Compression with Variational Energy-based Model
Minsoo Kang, Hyewon Yoo, Eunhee Kang, Sehwan Ki, Hyong-Euk Lee, Bohyung Han
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 VEM mutual-information maximization | OMGD · Horse->Zebra CycleGAN | FID: 57.14 -> 50.83 | Improved |
| add VEM mutual-information maximization | OMGD · Edges->Shoes Pix2Pix | FID: 27.39 -> 24.59 | Improved |
| add VEM mutual-information maximization | OMGD · Cityscapes Pix2Pix | mIoU: 47.52 -> 49.39 | Improved |
| add VEM | OMGD · Summer->Winter CycleGAN | FID: 75.20 -> 74.04 | Improved |
| add VEM | GCC · CelebA SAGAN | FID: 27.91 -> 25.27 | Improved |
| add VEM | CAGC · FFHQ StyleGAN2 | FID: 7.90 -> 7.48 | Improved |
| add VEM (ablation) | CAT · Horse->Zebra | FID: 64.79 -> 52.44 | Improved |
| add VEM (ablation) | GAN-Compression · Horse->Zebra | FID: 59.13 -> 50.01 | Improved |
| replace VEM with Gaussian VID mutual information (ablation) | OMGD · Horse->Zebra | FID: 50.83 -> 65.73 | Hurt |
| replace VEM with contrastive CRD mutual information (ablation) | OMGD · Horse->Zebra | FID: 50.83 -> 70.59 | Hurt |
| add CRD (contrastive MI) to baseline (ablation) | OMGD · Horse->Zebra | FID: 57.14 -> 70.59 | Hurt |
| add VID to baseline (ablation) | OMGD · Edges->Shoes Pix2Pix | FID: 27.39 -> 28.02 | Hurt |
Earlier methods this paper compares with
Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.
- GAN-Compression — GAN Compression: Efficient Architectures for Interactive Conditional GANs (CVPR 2020) · better on 1 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
- Nickel and Diming Your GAN: A Dual-Method Approach to Enhancing GAN Efficiency via Knowledge Distillation (ECCV 2024 · similarity 0.90)
- Content-Aware GAN Compression (CVPR 2021 · similarity 0.89)
- Online Multi-Granularity Distillation for GAN Compression (ICCV 2021 · similarity 0.88)
- Self-Supervised Generative Adversarial Compression (NeurIPS 2020 · similarity 0.88)
- Generative Adversarial Networks for Extreme Learned Image Compression (ICCV 2019 · similarity 0.88)
- AutoGAN-Distiller: Searching to Compress Generative Adversarial Networks (ICML 2020 · 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.
