AI Paper Atlas

Research map / Optimization / Optimizers

Exploring the Algorithm-Dependent Generalization of AUPRC Optimization with List Stability

Peisong Wen, Qianqian Xu, Zhiyong Yang, Yuan He, Qingming Huang

NeurIPS 2022 · Optimization › Optimizers · approach: Differentiating through embedded optimization · 5 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 unbiased stochastic estimator (prior pi) instead of biased np/(np+nn) (ablation)SOPRC with SGD · iNaturalistmAUPRC: 34.58 -> 35.84Improved
add auxiliary vector v_t to estimate positive scores (ablation)SOPRC with unbiased estimator · iNaturalistmAUPRC: 35.84 -> 35.99Improved
add semi-variance regularization L_var (ablation)SOPRC with v_t · iNaturalistmAUPRC: 35.99 -> 36.16Improved
switch optimizer from SGD to Adam (ablation)SOPRC full · iNaturalistmAUPRC: 36.16 -> 36.20No clear effect
add unbiased estimator, v_t and L_var together (ablation)SOPRC · iNaturalistRecall@4: 66.35 -> 68.22Improved
add unbiased estimator, v_t and L_var together (ablation)SOPRC · iNaturalistRecall@32: 92.72 -> 93.71Improved
use larger batch sizeSOPRC · iNaturalist validationmAUPRCImproved
optimize AUPRC directly instead of pairwise losses (XBM)SOPRC vs XBM · iNaturalistmAUPRC: 27.46 -> 36.16Improved
optimize AUPRC directly instead of pairwise losses (XBM)SOPRC vs XBM · PKU VehicleIDRecall@1: 92.78 -> 92.56Hurt
SOPRC versus best baseline SmoothAPSOPRC · Stanford Online ProductsmAUPRC: 61.65 -> 62.75Improved
SOPRC versus best baseline SmoothAPSOPRC · iNaturalistmAUPRC: 33.92 -> 36.16Improved
SOPRC versus best baseline DIRSOPRC · PKU VehicleIDmAUPRC: 72.72 -> 74.92Improved

Earlier methods this paper compares with

Named baselines from earlier accepted papers in this paper's results tables: better than 1 of 1 on most shared results.

Matched by method name in results tables read by AI. Authors' own tables tend to favour the newer method, and a comparison can be missed when a method is renamed.

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