AI Paper Atlas

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

Weighted ROC Curve in Cost Space: Extending AUC to Cost-Sensitive Learning

Huiyang Shao, Qianqian Xu, Zhiyong Yang, Peisong Wen, Gao Peifeng, Qingming Huang

NeurIPS 2023 · Learning theory and algorithms › Prediction and decision losses · approach: Data attribution and recourse · 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
Use WAUC-Gau bilevel cost-sensitive training instead of ECLcost-sensitive classifier · CIFAR-10-LT Subset1WAUC: 0.589 -> 0.679Improved
Use WAUC-Gau instead of NWAUCcost-sensitive classifier · CIFAR-100-LT Subset3WAUC: 0.631 -> 0.728Improved
Use WAUC-Log instead of ECLcost-sensitive classifier · CIFAR-100-LT Subset1bLCOST: 0.017 -> 0.012Improved
Use WAUC-Gau instead of AdaCOScost-sensitive classifier · Jane Street Market PredictionProfit: 5839 ± 34.92 -> 6526 ± 15.98Improved
Use WAUC-Gau instead of BCEcost-sensitive classifier · Jane Street Market PredictionProfit: 4955 ± 20.14 -> 6526 ± 15.98Improved
Use WAUC-Gau instead of ExAUCcost-sensitive classifier · Jane Street Market PredictionAUC-hat: 0.539 ± .004 -> 0.526 ± .002No clear effect
Use WAUC-Log instead of ExAUC (AUC-oriented method)cost-sensitive classifier · CIFAR-100-LT Subset1AUC: 0.933 -> 0.906Mixed
Use two-stage decision (train classifier or AUC, then pick cost threshold) instead of one-stage WAUC learningBCE and AUC methods · cost-sensitive benchmarksWAUC and bLCOSTHurt
Choose smaller T in the inner threshold estimation (ablation)WAUC-Gau · CIFAR benchmarksWAUCImproved
Choose larger T (ablation)WAUC-Gau · CIFAR benchmarksvariance of WAUCImproved
Vary kernel bandwidth in KDE (ablation)WAUC-Gau · CIFAR benchmarksWAUC and costMixed

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