AI Paper Atlas

Research map / Label-efficient and robust learning / Recognition and long-tail

Correcting the Triplet Selection Bias for Triplet Loss

Baosheng Yu, Tongliang Liu, Mingming Gong, Changxing Ding, Dacheng Tao

ECCV 2018 · Label-efficient and robust learning › Recognition and long-tail · approach: Group-robust classification · 107 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
add distribution matching loss (lambda=0.01)Triplet loss · CARS196R@1: 0.7781 -> 0.7917Improved
add distribution matching loss (lambda=0.005)Triplet loss · CARS196R@1: 0.7781 -> 0.7912Improved
increase lambda to 0.1Triplet loss · CARS196R@1: 0.7917 -> 0.7631Hurt
add distribution matching loss (lambda=0.01)Triplet loss · CUB200-2011R@1: 0.4450 -> 0.4660Improved
add distribution matching loss (lambda=0.005)Triplet loss · CUB200-2011R@1: 0.4450 -> 0.4439No clear effect
increase lambda to 0.5Triplet loss · CUB200-2011R@1: 0.4660 -> 0.4483Hurt
add distribution matching loss (lambda=0.01)Triplet loss · Stanford Online ProductsR@1: 0.6274 -> 0.6303No clear effect
add distribution matching loss (lambda=0.01)Triplet loss · Stanford Online ProductsR@100: 0.9050 -> 0.9071No clear effect
add distribution matching loss (lambda=0.01)Triplet loss · CARS196R@20: 0.9716 -> 0.9745Improved
add distribution matching loss (lambda=0.01)Triplet loss · CUB200-2011R@20: 0.8893 -> 0.9024Improved
add distribution matching loss to triplet lossTriplet loss · MNISTtest accuracyImproved
vary lambda over 0, 0.1, 0.5, 1.0, 2.0, 5.0Triplet loss · MNISTtest accuracyMixed

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