Research map / Label-efficient and robust learning / Recognition and long-tail
Data-Free Knowledge Amalgamation via Group-Stack Dual-GAN
Check 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 data-free knowledge amalgamation instead of DAFL | TargetNet · VOC2007 randomly selected 10-label set | mAP: 73.6 vs 46.8 | Helped |
| Use data-free knowledge amalgamation instead of DAFL | TargetNet · VOC2007 whole 20-label set | mAP: 87.2 vs 54.9 | Helped |
| Use data-free knowledge amalgamation instead of DAFL | TargetNet · MS-COCO | O-F1: 68.2 vs 36.6 | Helped |
| Add both dual-generator input streams (ablation) | Dual-GAN · VOC2007 single-teacher ablation | mAP: 70.2 to 73.6; 59.8 to 73.6 | Helped |
| Remove the generator-to-dual-generator link and use only F_in^1 (ablation) | Dual-GAN · VOC2007 single-teacher ablation | mAP: 73.6 vs 70.2 | Hurt |
| Use only F_in^2 and omit I_gan as an input (ablation) | Dual-GAN · VOC2007 single-teacher ablation | mAP: 73.6 vs 59.8 | Hurt |
| Add the discrete loss L_dis (ablation) | Group-stack GAN · VOC2007 single-teacher ablation | label diversity | Helped |
Closest papers by meaning
- Hybrid Data-Free Knowledge Distillation (AAAI 2025 · similarity 0.89)
- Amalgamating Knowledge towards Comprehensive Classification (AAAI 2019 · similarity 0.86)
- Small Scale Data-Free Knowledge Distillation (CVPR 2024 · similarity 0.86)
- MCL-GAN: Generative Adversarial Networks with Multiple Specialized Discriminators (NeurIPS 2022 · similarity 0.86)
- Taming the Tail in Class-Conditional GANs: Knowledge Sharing via Unconditional Training at Lower Resolutions (CVPR 2024 · similarity 0.85)
- Impartial Adversarial Distillation: Addressing Biased Data-Free Knowledge Distillation via Adaptive Constrained Optimization (AAAI 2024 · similarity 0.84)
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.