Research map / Architectures and efficiency / Pruning and efficient tuning
Learning to Weight Parameters for Training Data Attribution
Shuangqi Li, Hieu Le, Jingyi Xu, Mathieu Salzmann
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 learned parameter-group weights | TracIn · ImageNet LDS (ResNet-18) | LDS (%): 11.39 -> 23.92 | Improved |
| add learned parameter-group weights | TRAK · ImageNet LDS (ViT-B/16) | LDS (%): 14.77 -> 16.74 | Improved |
| add learned parameter-group weights | EKFAC · WikiText-103 LDS | LDS (%): 15.14 -> 18.33 | Improved |
| add learned parameter-group weights | TracIn · CIFAR-2 diffusion | LDS (%): 1.39 -> 8.48 | Improved |
| add learned parameter-group weights | DAS · ArtBench-2 diffusion | LDS (%): 30.47 -> 31.58 | Improved |
| add learned parameter-group weights | TracIn · ImageNet mislabeled detection (ViT-B/16) | AUC: 71.27 -> 83.58 | Improved |
| add learned parameter-group weights | TRAK · WikiText-103 tail-patch | tail-patch score (%): 7.77 -> 7.88 | No clear effect |
| add learned parameter-group weights | TracIn · WikiText-103 tail-patch | tail-patch score (%): 4.66 -> 5.60 | Improved |
| learn weights with subject-focused query set | D-TRAK · SB-Pokemon validation | subject recall@10 (%): 19.4 -> 50.6 | Improved |
| learn weights with style-focused query set | D-TRAK · SB-Pokemon validation | style recall@10 (%): 61.8 -> 73.0 | Improved |
| learn weights with background-focused query set | D-TRAK · SB-Pokemon validation | background recall@10 (%): 17.5 -> 26.6 | Improved |
| use style weights to attribute subject (ablation) | D-TRAK · SB-Pokemon validation | subject recall@10 (%): 19.4 -> 15.2 | Hurt |
Closest papers by meaning
- Taming Hyperparameter Sensitivity in Data Attribution: Practical Selection Without Costly Retraining (NeurIPS 2025 · similarity 0.88)
- Fast Data Attribution for Text-to-Image Models (NeurIPS 2025 · similarity 0.87)
- Data Attribution for Text-to-Image Models by Unlearning Synthesized Images (NeurIPS 2024 · similarity 0.87)
- quanda: An Interpretability Toolkit for Training Data Attribution Evaluation (NeurIPS 2026 · similarity 0.86)
- Final-Model-Only Data Attribution with a Unifying View of Gradient-Based Methods (NeurIPS 2025 · similarity 0.86)
- $\texttt{dattri}$: A Library for Efficient Data Attribution (NeurIPS 2024 · 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.
