AI Paper Atlas

Research map / Learning theory and algorithms / Prediction and decision losses

Is Data Shapley Not Better than Random in Data Selection? Ask NASH

Xiao Tian, Jue Fan, Rachael Hwee Ling Sim, Wang Zixuan, Nancy F. Chen, Bryan Kian Hsiang Low

ICML 2026 · Learning theory and algorithms › Prediction and decision losses · approach: Shapley value attribution · 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
select data with NASH instead of top-m Data ShapleyData Shapley selection · WD/PO/PM-LR, MR-BT, MP-BT, RT-LM general selectionvalidation accuracyImproved
select data with NASH instead of top-m Data ShapleyData Shapley selection · label-noise-injected MR-BT and MP-BT (10-30% flip)validation accuracyImproved
use NASH with alternative utility (validation loss, negated MSE)Data Shapley selection · WD-LR loss, PP-RR, AU-RRutilityImproved
pair NASH with Data Banzhaf and other semivaluesvanilla semivalues · semivalue compatibilityvalidation accuracyImproved
use Data Shapley or Banzhaf instead of LOO within NASH (ablation)NASH · semivalue ablationvalidation accuracyImproved
use Beta(4,1) semivalue within NASH at small selection size (ablation)NASH · semivalue ablation, small selectionvalidation accuracyImproved
use Beta(1,4) semivalue within NASH at large selection size (ablation)NASH · semivalue ablation, large selectionvalidation accuracyHurt

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