AI Paper Atlas

Research map / Multi-agent systems and games / Human-AI and goal-conditioned

Data Agent: Learning to Select Data via End-to-End Dynamic Optimization

Suorong Yang, Fangjian Su, Hai Gan, Ziqi Ye, Jie Li, Baile Xu, Furao Shen, Soujanya Poria

ICML 2026 · Multi-agent systems and games › Human-AI and goal-conditioned · approach: Imitation and skill learning · 0 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 difficulty reward (ablation)Data Agent · Tiny-ImageNet (ResNet-50), 30%accuracy: 41.5 -> 42.1Improved
add difficulty reward (ablation)Data Agent · Tiny-ImageNet (ResNet-50), 50%accuracy: 42.8 -> 45.0Improved
add confidence reward (ablation)Data Agent · Tiny-ImageNet (ResNet-50), 30%accuracy: 41.5 -> 42.2Improved
add confidence reward (ablation)Data Agent · Tiny-ImageNet (ResNet-50), 70%accuracy: 43.1 -> 48.4Improved
add confidence reward to difficulty reward (ablation)Data Agent · Tiny-ImageNet (ResNet-50), 30%accuracy: 42.1 -> 42.5Improved
add adaptive reward weighting r (ablation)Data Agent · Tiny-ImageNet (ResNet-50), 30%accuracy: 42.5 -> 44.9Improved
add adaptive reward weighting r (ablation)Data Agent · Tiny-ImageNet (ResNet-50), 50%accuracy: 46.1 -> 47.0Improved
add adaptive reward weighting r (ablation)Data Agent · Tiny-ImageNet (ResNet-50), 70%accuracy: 48.6 -> 49.4Improved
add PPO agent over greedy top-scoring selection by composite reward (ablation)Data Agent · CIFAR-100 and Tiny-ImageNetaccuracyImproved
use data agent instead of full dataset trainingResNet-50 · ImageNet-1k, 60% selectionaccuracy: 76.4 -> 76.8Improved
use data agent instead of full dataset trainingResNet-50 · ImageNet-R, 60% selectionaccuracy: 36.2 -> 38.0Improved
use data agent instead of full dataset trainingResNet-50 · ImageNet-O, 60% selectionAUPR: 13.2 -> 15.7Improved

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