AI Paper Atlas

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

From Points to Coalitions: Hierarchical Contrastive Shapley Values for Prioritizing Data Samples

Canran Xiao, Jiabao Dou, Zhiming Lin, Zong Ke, Liwei Hou

AAAI 2026 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 19 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
use hierarchical contrastive Shapley (HCDV) instead of Monte-Carlo Data-Shapleydata valuation · Criteo-1Btest AUC: 0.6175 -> 0.6269Improved
use HCDV instead of Group Shapleydata valuation · Fashion-MNISTtest accuracy: 0.868 -> 0.891Improved
use HCDV instead of Monte-Carlo Data-Shapleydata valuation · Criteo-1Bvaluation time (hr): 47.5 -> 12.3Improved
use HCDV top-ranked augmented samples instead of bottom-rankedFashion-MNIST ConvNet · Fashion-MNIST augmentation filteringaccuracy: +2.8 ppImproved
incremental HCDV update instead of full recomputationHCDV streaming · click-streamfinal AUC and compute: 0.808 -> 0.805 AUC; 2.5x less computeImproved
use HCDV instead of DU-Shapleydata valuation · OpenDataVal noisy-label detectionmacro-F1: +1-3 ppImproved
use HCDV for revenue allocation instead of MCDSdata marketplace · UCI Adult sellerscorrelation with leave-one-out: 0.89 -> 0.94Improved

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