AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

TarGATE: Target-Aware Data Selection via Token-Attenuation Gates

Xiandi Luo, Shiwei Li, Haozhao Wang, Yihao Ouyang, Zhuoqi Hu, Yichen Li, Xiao Yang, Huning Liu, Ruixuan Li

ICML 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Data augmentation and generalization bias

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
Use joint optimization of reference loss and warmup loss (beta between 0 and 2) instead of only one objective (ablation)TarGATE · CommonsenseQA, GSM8K, MBPP (Qwen3-0.6B selector, Qwen3-1.7B target)accuracyImproved
Use smaller beta for reasoning and coding tasks, larger beta for CommonsenseQA (ablation)TarGATE · GSM8K, MBPP vs CommonsenseQAaccuracyMixed
Shrink reference set size (100% to 50% to 25%) and lower beta (ablation)TarGATE · CQA, GSM8K, MBPPaccuracyMixed
Vary regularization weight lambda away from about 1 (ablation)TarGATE · CQA, GSM8K, MBPPaccuracyHurt
Select data with TarGATE instead of training on the full 200k poolTulu-3 fine-tuning · Tulu-3 average (3 targets x 3 tasks)average accuracy: 64.9 -> 68.2Improved
Use TarGATE instead of random selection in noisy poolSynthetic noisy scenario · CQA/GSM8K averageaverage accuracy: 60.6 -> 66.5Improved
Use larger 1.7B selector instead of 0.6B selector (ablation)TarGATE · Tulu-3 averageaverage accuracy: 68.2 -> 67.6Hurt
Use 1.7B selector instead of 0.6B for gradient baseline GradSim (ablation)GradSim selecting for Llama3-8B · Tulu-3 GSM8Kaccuracy: 56.3 -> 45.7Hurt
Replace gradient-based scoring with single-forward-pass IRR scoringData selection cost · 200k samples, Qwen3-1.7BPetaFLOPs: 776.0 -> 424.9Improved

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