Research map / Learning theory and algorithms / Prediction and decision losses
Validation Free and Replication Robust Volume-based Data Valuation
Xinyi Xu, Zhaoxuan Wu, Chuan-Sheng Foo, Bryan Kian Hsiang Low
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 |
|---|---|---|---|
| replace volume with robust volume (cube compression) as Shapley characteristic function | volume Shapley value · replicated data, Trip Advisor i.i.d. | replication robustness | Improved |
| use RVSV instead of VSV | data valuation · CaliH i.i.d. similarity to VLSV | Pearson correlation rp: -0.886 -> 0.767 | Improved |
| use RVSV instead of VSV | data valuation · CaliH supersets 1 | Pearson correlation rp: -0.660 -> 0.918 | Improved |
| discretization width omega 0.1 instead of 0.05 (ablation) | RVSV · CaliH disjoint 0 | Pearson correlation rp: 0.700 -> 0.351 | Hurt |
| discretization width omega 0.1 instead of 0.05 (ablation) | RVSV · CaliH i.i.d. | Pearson correlation rp: 0.767 -> 0.767 | No clear effect |
| add dataset with larger robust volume first | training data selection · credit card, Uber Lyft | train loss | Improved |
| larger volume for test loss | training data selection · credit card, Uber Lyft | test loss | Mixed |
| change validation set preference | LOO · UK used car, credit card | valuation consistency | Hurt |
Closest papers by meaning
- Distributionally Robust Data Valuation (ICML 2024 · similarity 0.88)
- SAVA: Scalable Learning-Agnostic Data Valuation (ICLR 2025 · similarity 0.85)
- KAIROS: Scalable Model-Agnostic Data Valuation (NeurIPS 2025 · similarity 0.85)
- DeRDaVa: Deletion-Robust Data Valuation for Machine Learning (AAAI 2024 · similarity 0.84)
- GMValuator: Similarity-based Data Valuation for Generative Models (ICLR 2025 · similarity 0.84)
- Convex Dataset Valuation for Post-Training (ICML 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.
