AI Paper Atlas

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

Smart Mining for Deep Metric Learning

Ben Harwood, Vijay Kumar B G, Gustavo Carneiro, Ian Reid, Tom Drummond

ICCV 2017 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 359 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 semi-hard stochastic mining with FANNG smart miningtriplet loss · CUB-200-2011Recall@1: 42.59 -> 45.90Improved
replace semi-hard stochastic mining with FANNG smart miningtriplet loss · Cars196Recall@1: 51.54 -> 56.11Improved
add global loss (ablation)Triplet + FANNG · CUB-200-2011Recall@1: 45.90 -> 49.44Improved
add global loss (ablation)Triplet + FANNG · Cars196Recall@1: 56.11 -> 64.20Improved
add global loss (ablation)Triplet + FANNG · CUB-200-2011NMI: 58.10 -> 60.09Improved
add adaptive controller (ablation)Triplet + FANNG + Global · CUB-200-2011Recall@8: 82.85 -> 83.31Improved
add adaptive controller (ablation)Triplet + FANNG + Global · Cars196Recall@8: 88.94 -> 90.19Improved
add adaptive controller (ablation)Triplet + FANNG + Global · CUB-200-2011NMI: 60.09 -> 59.90No clear effect
add adaptive controller (ablation)Triplet + FANNG + Global · Cars196convergence speedImproved
use FANNG mined triplets in place of random triplets (global loss kept) versus Triplet + GlobalTriplet + Global · Cars196Recall@1: 61.41 -> 64.20Improved
use smaller initial kappa (kappa in {1,4,16,64}) (ablation)Triplet + FANNG + Global · CUB-200-2011Recall@K after 20 epochsHurt
use a low percentage (10-30%) of mined triplets (ablation)Triplet + FANNG + Adaptive · CUB-200-2011training error control / convergenceHurt

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