Research map / Label-efficient and robust learning / Recognition and long-tail
Exploring 3D Dataset Pruning
Xiaohan Zhao, Xinyi Shang, Jiacheng Liu, Zhiqiang Shen
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 |
|---|---|---|---|
| use KD soft labels instead of hard labels under rebalancing policies (ablation) | PointNet++ on FL-RBF subset · ShapeNet55 | OA: 79.10 -> 81.82 | Improved |
| use KD soft labels instead of hard labels under rebalancing policies (ablation) | PointNet++ on FL-RBF subset · ShapeNet55 | mAcc: 60.47 -> 64.24 | Improved |
| select with embedding geometry rather than scalar scores (Loss) (ablation) | subset selection · ShapeNet55 | head/tail score overlap: 26% -> 96% | Improved |
| select with embedding geometry rather than scalar scores (Loss) (ablation) | subset selection · ShapeNet55 | selection imbalance ratio: 40.94x -> 1.88x | Improved |
| add seeded stratified floor (k=0.6) to FL-RBF (ablation) | FL-RBF · ShapeNet55 (m=550) | mAcc: +8.13 | Improved |
| add seeded stratified floor (k=0.4) to FL-RBF (ablation) | FL-RBF · ShapeNet55 (m=1100) | OA: -0.30 | No clear effect |
| add seeded stratified floor (k=0.4) to FL-RBF (ablation) | FL-RBF · ShapeNet55 (m=1100) | mAcc: +4.65 | Improved |
| add seeded stratified floor (k=0.4) to FL-RBF (ablation) | FL-RBF · ScanObjectNN (m=150) | mAcc: +3.34 | Improved |
| add seeded stratified floor (k=0.4) to FL-RBF (ablation) | FL-RBF · ModelNet40 (m=1600) | OA: -0.24 | No clear effect |
| add seeded stratified floor (k=0.6) to FL-RBF (ablation) | FL-RBF · ShapeNet55 (m=1100) | OA: -1.82 | Hurt |
| add seeded stratified floor (k=0.6) to FL-RBF (ablation) | MeshMAE FL-RBF · ModelNet40 mesh (m=400) | mAcc: +2.48 | Improved |
| add seeded stratified floor (k=0.4) to FL-RBF (ablation) | MeshMAE FL-RBF · ModelNet40 mesh (m=400) | OA: -0.58 | Hurt |
Closest papers by meaning
- Spanning Training Progress: Temporal Dual-Depth Scoring (TDDS) for Enhanced Dataset Pruning (CVPR 2024 · similarity 0.87)
- DRoP: Distributionally Robust Data Pruning (ICLR 2025 · similarity 0.87)
- Exploring Learning Complexity for Efficient Downstream Dataset Pruning (ICLR 2025 · similarity 0.87)
- MDP: Multidimensional Vision Model Pruning with Latency Constraint (CVPR 2025 · similarity 0.87)
- $\mathbb{D}^2$ Pruning: Message Passing for Balancing Diversity & Difficulty in Data Pruning (ICLR 2024 · similarity 0.87)
- Lightweight Dataset Pruning without Full Training via Example Difficulty and Prediction Uncertainty (ICML 2025 · similarity 0.86)
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.
