AI Paper Atlas

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

ICLR 2026 · Architectures and efficiency › Pruning and efficient tuning · approach: Low-rank compression and merging · 2 citations (OpenAlex)

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 changedWhereMeasuredEffect
add learned parameter-group weightsTracIn · ImageNet LDS (ResNet-18)LDS (%): 11.39 -> 23.92Improved
add learned parameter-group weightsTRAK · ImageNet LDS (ViT-B/16)LDS (%): 14.77 -> 16.74Improved
add learned parameter-group weightsEKFAC · WikiText-103 LDSLDS (%): 15.14 -> 18.33Improved
add learned parameter-group weightsTracIn · CIFAR-2 diffusionLDS (%): 1.39 -> 8.48Improved
add learned parameter-group weightsDAS · ArtBench-2 diffusionLDS (%): 30.47 -> 31.58Improved
add learned parameter-group weightsTracIn · ImageNet mislabeled detection (ViT-B/16)AUC: 71.27 -> 83.58Improved
add learned parameter-group weightsTRAK · WikiText-103 tail-patchtail-patch score (%): 7.77 -> 7.88No clear effect
add learned parameter-group weightsTracIn · WikiText-103 tail-patchtail-patch score (%): 4.66 -> 5.60Improved
learn weights with subject-focused query setD-TRAK · SB-Pokemon validationsubject recall@10 (%): 19.4 -> 50.6Improved
learn weights with style-focused query setD-TRAK · SB-Pokemon validationstyle recall@10 (%): 61.8 -> 73.0Improved
learn weights with background-focused query setD-TRAK · SB-Pokemon validationbackground recall@10 (%): 17.5 -> 26.6Improved
use style weights to attribute subject (ablation)D-TRAK · SB-Pokemon validationsubject recall@10 (%): 19.4 -> 15.2Hurt

Closest papers by meaning

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.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy