Research map / Label-efficient and robust learning / Recognition and long-tail
Scalable Neural Data Server: A Data Recommender for Transfer Learning
Tianshi Cao, Sasha Doubov, David Acuna, Sanja Fidler
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 SNDS-selected data instead of random sample, 2% budget | SNDS-50 · 9 fine-grained targets average | accuracy: 55.75 -> 59.07 | Improved |
| use SNDS-selected data instead of random sample, 5% budget | SNDS-50 · 9 fine-grained targets average | accuracy: 63.95 -> 64.96 | Improved |
| use SNDS-selected data instead of random sample, 10% budget | SNDS-50 · 9 fine-grained targets average | accuracy: 67.10 -> 68.33 | Improved |
| SNDS (indexing on public data) versus NDS (experts on each source), 2% budget | SNDS-50 · 9 targets average | accuracy: 59.53 -> 59.07 | Mixed |
| SNDS versus NDS at 10% budget | SNDS-50 · 9 targets average | accuracy: 68.33 -> 68.33 | No clear effect |
| replace stochastic sampling with greedy selection, 2% budget (ablation) | SNDS · 9 targets average | accuracy: 59.07 -> 57.55 | Hurt |
| replace stochastic sampling with greedy selection, 5% budget (ablation) | SNDS · 9 targets average | accuracy: 64.96 -> 64.45 | Hurt |
| greedy selection lowers pretraining task accuracy, 2% samples (ablation) | SNDS · 601-way OpenImages pretraining | accuracy: 29.44 -> 25.02 | Hurt |
| greedy selection lowers pretraining task accuracy, 5% samples (ablation) | SNDS · 601-way OpenImages pretraining | accuracy: 34.40 -> 31.61 | Hurt |
| use 10 experts instead of 50, 2% budget (ablation) | SNDS · 9 targets average | accuracy: 59.07 -> 58.32 | Hurt |
| use 10 experts instead of 50, 5% budget (ablation) | SNDS · 9 targets average | accuracy: 64.96 -> 64.00 | Hurt |
| drop public clusters overlapping target classes (Pets, CUB200) (ablation) | SNDS · Oxford Pets, CUB200 | accuracy | No clear effect |
Closest papers by meaning
- Evaluating Transferability in Retrieval Tasks: An Approach Using MMD and Kernel Methods (CVPR 2024 · similarity 0.85)
- Knowledge Transfer from Vision Foundation Models for Efficient Training of Small Task-specific Models (ICML 2024 · similarity 0.84)
- SAVA: Scalable Learning-Agnostic Data Valuation (ICLR 2025 · similarity 0.84)
- Behavior Importance-Aware Graph Neural Architecture Search for Cross-Domain Recommendation (AAAI 2025 · similarity 0.83)
- Scaling Laws for Data Filtering-- Data Curation cannot be Compute Agnostic (CVPR 2024 · similarity 0.83)
- Insights from the Use of Previously Unseen Neural Architecture Search Datasets (CVPR 2024 · similarity 0.83)
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.
