AI Paper Atlas

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

The Extreme Value Machine

Ethan M. Rudd, Lalit P. Jain, Walter J. Scheirer, Terrance E. Boult

TPAMI 2017 · Label-efficient and robust learning › Recognition and long-tail · approach: Learning from unlabeled and positive-unlabeled data · 352 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
reduce Letter training points kept to 10% via larger redundancy threshold (varsigma 0.008) (ablation)EVM · Letter (closed set)accuracy: 96% -> 92%Hurt
keep only 50% or 40% of Letter training points via model reduction (ablation)EVM · Letter (closed set)accuracy: 96% base; 50% (varsigma = 0.492) and 40% (varsigma = 0.186) comparableNo clear effect
average over k>1 largest probabilities per class instead of k=1 (ablation)EVM · Letter / OLETTERaccuracy: slight gains of 1-2% in accuracyImproved
use cosine similarity instead of Euclidean distance for margins (ablation)EVM · Open world ImageNetopen world accuracy / F1Improved
use cosine similarity instead of Euclidean distance (ablation)NNO · Open world ImageNetopen world accuracyHurt
replace NNO with EVM for open world recognitionopen world ImageNet protocol · Open world ImageNetaccuracy and F1Improved
compare EVM with W-SVM on OLETTEREVM · OLETTERF1-MeasureNo clear effect
use EVM instead of SVM models for model sizeEVM · OLETTERvector ratio: an order of magnitude smaller than that of any of the SVM modelsImproved
vary k in 1..10 and varsigma (ablation)EVM · accuracy / F1accuracy or F1: less than 2% in accuracy or F1No clear effect
use incremental KDE classifier instead of EVMopen world ImageNet · Open world ImageNetscalability: fewer than 40 samples enrolled after 24 hours; approximately 18 yearsHurt

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