Research map / Image generation / Autoregressive and latent generation
DIVER: Diving Deeper into Distilled Data via Expressive Semantic Recovery
Qianxin Xia, Zhiyong Shu, Wenbo Jiang, Jiawei Du, Jielei Wang, Guoming Lu
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 DIVER stage II | MTT · ImageFruit IPC=10 unseen architectures | accuracy (%): 18.9 -> 29.8 | Improved |
| add DIVER stage II | SRe2L · ImageNet-1K IPC=50 ResNet-18 | accuracy (%): 46.8 -> 54.0 | Improved |
| add DIVER stage II as squeeze model grows stronger | SRe2L / G-VBSM · ImageNet-1K ResNet-101 | accuracy (%): 61.0 -> 63.7 | Mixed |
| add DIVER to MGD3 prototypes | MGD3 · ImageNet-1K IPC=10 | accuracy (%): 45.8 -> 46.4 | Improved |
| add Minimax fine-tuning to MGD3+DIVER | MGD3 + DIVER · ImageNet-1K IPC=10 | accuracy (%): 46.4 -> 46.9 | Improved |
| add semantic inheritance (ablation) | DIVER · ImageFruit IPC=10, MTT | accuracy (%): 23.4 -> 26.2 | Improved |
| add semantic guidance (ablation) | DIVER · ImageFruit IPC=10, MTT | accuracy (%): 23.4 -> 27.8 | Improved |
| add semantic guidance to semantic inheritance (ablation) | DIVER · ImageFruit IPC=10, MTT | accuracy (%): 26.2 -> 28.3 | Improved |
| add semantic fusion (guidance only in semantic phase) (ablation) | DIVER · ImageFruit IPC=10, MTT | accuracy (%): 28.3 -> 29.8 | Improved |
| use raw DiT without recovery modules (ablation) | DiT · ImageFruit IPC=10, MTT | accuracy (%): 18.9 -> 23.4 | Improved |
| replace distilled images with random images in DIVER (ablation) | DIVER · ImageFruit IPC=10, MTT | accuracy (%): 29.8 -> 21.7 | Hurt |
| swap SD-V1.5 for DiT as diffusion model (ablation) | DIVER · ImageNette IPC=10, MTT | accuracy (%): 26.5 -> 34.3 | Improved |
Closest papers by meaning
- D^4: Dataset Distillation via Disentangled Diffusion Model (CVPR 2024 · similarity 0.87)
- $\mathbf{D^{3}S^{2}}$: Diffusion-Guided Dataset Distillation for Semantic Segmentation (NeurIPS 2026 · similarity 0.86)
- Unlocking Dataset Distillation with Diffusion Models (NeurIPS 2025 · similarity 0.85)
- Small Scale Data-Free Knowledge Distillation (CVPR 2024 · similarity 0.85)
- Large Scale Dataset Distillation with Domain Shift (ICML 2024 · similarity 0.85)
- Boost Self-Supervised Dataset Distillation via Parameterization, Predefined Augmentation, and Approximation (ICLR 2025 · 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.
