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
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 changed | Where | Measured | Effect |
|---|---|---|---|
| use hierarchical contrastive Shapley (HCDV) instead of Monte-Carlo Data-Shapley | data valuation · Criteo-1B | test AUC: 0.6175 -> 0.6269 | Improved |
| use HCDV instead of Group Shapley | data valuation · Fashion-MNIST | test accuracy: 0.868 -> 0.891 | Improved |
| use HCDV instead of Monte-Carlo Data-Shapley | data valuation · Criteo-1B | valuation time (hr): 47.5 -> 12.3 | Improved |
| use HCDV top-ranked augmented samples instead of bottom-ranked | Fashion-MNIST ConvNet · Fashion-MNIST augmentation filtering | accuracy: +2.8 pp | Improved |
| incremental HCDV update instead of full recomputation | HCDV streaming · click-stream | final AUC and compute: 0.808 -> 0.805 AUC; 2.5x less compute | Improved |
| use HCDV instead of DU-Shapley | data valuation · OpenDataVal noisy-label detection | macro-F1: +1-3 pp | Improved |
| use HCDV for revenue allocation instead of MCDS | data marketplace · UCI Adult sellers | correlation with leave-one-out: 0.89 -> 0.94 | Improved |
Closest papers by meaning
- DU-Shapley: A Shapley Value Proxy for Efficient Dataset Valuation (NeurIPS 2024 · similarity 0.85)
- Towards Attributions of Input Variables in a Coalition (ICML 2025 · similarity 0.84)
- Rethinking Data Shapley for Data Selection Tasks: Misleads and Merits (ICML 2024 · similarity 0.84)
- KAIROS: Scalable Model-Agnostic Data Valuation (NeurIPS 2025 · similarity 0.84)
- An Efficient Framework for Crediting Data Contributors of Diffusion Models (ICLR 2025 · similarity 0.84)
- Learning Distributions from Multiple Data Providers (NeurIPS 2026 · similarity 0.83)
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.
