Research map / Label-efficient and robust learning / Knowledge distillation
FYI: Flip Your Images for Dataset Distillation
Byunggwan Son, Youngmin Oh, Donghyeon Baek, Bumsub Ham
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 horizontal-flip counterparts of synthetic images (FYI) | DM · CIFAR-10 (1 IPC) | top-1 accuracy (%): 26.0 -> 28.7 | Improved |
| add FYI | MTT · CIFAR-100 (1 IPC) | top-1 accuracy (%): 24.3 -> 28.2 | Improved |
| add FYI | FTD · CIFAR-10 (1 IPC) | top-1 accuracy (%): 46.8 -> 49.5 | Improved |
| add FYI | FTD · CIFAR-100 (50 IPC) | top-1 accuracy (%): 50.7 -> 50.8 | No clear effect |
| add FYI | IDC · CIFAR-10 (50 IPC) | top-1 accuracy (%): 74.5 -> 75.2 | Improved |
| add FYI with fewer synthetic images (smaller IPC) | DM, MTT, DC, DSA · CIFAR-10/100 | top-1 accuracy (%) | Improved |
| add FYI | MTT · ImageSquawk (10 IPC) | top-1 accuracy (%): 52.3 -> 61.6 | Improved |
| use vertical flip instead of horizontal flip (ablation) | DC · CIFAR-10 (50 IPC) | top-1 accuracy (%): 53.9 -> 32.2 | Hurt |
| use scale augmentation instead of horizontal flip (ablation) | DM · CIFAR-10 (50 IPC) | top-1 accuracy (%): 63.0 -> 58.1 | Hurt |
| use rotate augmentation instead of horizontal flip (ablation) | DSA · CIFAR-10 (50 IPC) | top-1 accuracy (%): 60.6 -> 62.1 | Improved |
| add horizontal flip (FYI) rather than other augmentations (ablation) | DSA · CIFAR-10 (50 IPC) | top-1 accuracy (%): 60.6 -> 63.7 | Improved |
Closest papers by meaning
- Dataset Distillation for Pre-Trained Self-Supervised Vision Models (NeurIPS 2025 · similarity 0.89)
- Emphasizing Discriminative Features for Dataset Distillation in Complex Scenarios (CVPR 2025 · similarity 0.87)
- Large Scale Dataset Distillation with Domain Shift (ICML 2024 · similarity 0.87)
- Boost Self-Supervised Dataset Distillation via Parameterization, Predefined Augmentation, and Approximation (ICLR 2025 · similarity 0.87)
- D^4: Dataset Distillation via Disentangled Diffusion Model (CVPR 2024 · similarity 0.87)
- Efficient Dataset Distillation for Pre-Trained Self-Supervised Models via Statistical Flow Matching (NeurIPS 2026 · 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.
