AI Paper Atlas

Research map / Learning theory and algorithms / Kernels and regression theory

Posterior Refinement on Metric Matrix Improves Generalization Bound in Metric Learning

Mingda Wang, Canqian Yang, Yi Xu

ECCV 2022 · Learning theory and algorithms › Kernels and regression theory · approach: Empirical generalization measures · 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
replace trained metric matrix with an identity matrixArcFace · CUB-200-2011Precision@1: 60.12 -> 65.06Improved
replace trained metric matrix with an identity matrixProxyNCA · CUB-200-2011Precision@1: 56.99 -> 63.64Improved
replace trained metric matrix with an identity matrixTriplet · CUB-200-2011Precision@1: 54.95 -> 61.52Improved
replace trained metric matrix with an identity matrixArcFace · Cars196Precision@1: 72.88 -> 78.79Improved
replace trained metric matrix with an identity matrixProxyNCA · Cars196Precision@1: 73.91 -> 80.10Improved
replace trained metric matrix with an identity matrixTriplet · Cars196Precision@1: 64.92 -> 74.86Improved
replace trained metric matrix with an identity matrixTriplet · SOPPrecision@1: 67.96 -> 70.49Improved
replace trained metric matrix with an identity matrixArcFace · SOPPrecision@1: 71.98 -> 72.28No clear effect
replace trained metric matrix with an identity matrixArcFace · SOPMAP: 42.23 -> 42.18No clear effect
use best diagonal or random restraint instead of identityArcFace · CUB-200-2011Precision@1: 65.06 -> 64.90No clear effect
increase diagonal restraint degree r (reduce trace) (ablation)ArcFace, ProxyNCA, Triplet · CUB-200-2011MAP, Precision@1, RPImproved
increase random restraint degree r (reduce Frobenius norm) (ablation)ArcFace, ProxyNCA, Triplet · CUB-200-2011MAP, Precision@1, RPImproved

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