Research map / Architectures and efficiency / RNNs and normalization
Dataset Distillation by Influence Matching
Haoru Tan, Wang Wang, Sitong Wu, Xiuzhe Wu, Yang-Tian Sun, Chirui Chang, Shaofeng Zhang, Xiaojuan Qi
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 real-data initialization (ablation) | Inf-Match · CIFAR-100 IPC=50 | accuracy (%): 52.2 -> 53.7 | Improved |
| add sampling schedule on top of real-data init (ablation) | Inf-Match · CIFAR-100 IPC=50 | accuracy (%): 53.7 -> 55.0 | Improved |
| add learnable labels on top of real-data init (ablation) | Inf-Match · CIFAR-100 IPC=50 | accuracy (%): 53.7 -> 54.6 | Improved |
| add learnable labels to real-data init plus sampling schedule (ablation) | Inf-Match · CIFAR-100 IPC=50 | accuracy (%): 55.0 -> 57.4 | Improved |
| add sampling schedule to real-data init plus learnable labels (ablation) | Inf-Match · CIFAR-100 IPC=50 | accuracy (%): 54.6 -> 57.4 | Improved |
| use outcome-influence matching instead of NCFM feature matching | Inf-Match · CIFAR-100 IPC=50 | accuracy (%): 54.7 -> 57.4 | Improved |
| use influence matching instead of MTT trajectory matching | Inf-Match · CIFAR-100 IPC=50 | accuracy (%): 47.7 -> 57.4 | Improved |
| use influence matching instead of DATM | Inf-Match · Flickr30K 500 pairs text-to-image retrieval | recall (%): 14.1 -> 14.6 | Improved |
| use influence matching instead of NCFM | Inf-Match · Tiny-ImageNet IPC=10 | accuracy (%): +4.7% | Improved |
| use influence matching instead of NCFM | Inf-Match · CIFAR-100 IPC=50 | accuracy (%): 2.7% | Improved |
Closest papers by meaning
- Breaking Class Barriers: Efficient Dataset Distillation via Inter-Class Feature Compensator (ICLR 2025 · similarity 0.88)
- SelMatch: Effectively Scaling Up Dataset Distillation via Selection-Based Initialization and Partial Updates by Trajectory Matching (ICML 2024 · similarity 0.88)
- Towards Lossless Dataset Distillation via Difficulty-Aligned Trajectory Matching (ICLR 2024 · similarity 0.88)
- Efficient Dataset Distillation for Pre-Trained Self-Supervised Models via Statistical Flow Matching (NeurIPS 2026 · similarity 0.87)
- Path-Guided Flow Matching for Dataset Distillation (NeurIPS 2026 · similarity 0.87)
- Emphasizing Discriminative Features for Dataset Distillation in Complex Scenarios (CVPR 2025 · 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.
