AI Paper Atlas

Research map / Label-efficient and robust learning / Semi-supervised learning

A Data-Centric Approach for Improving Ambiguous Labels with Combined Semi-Supervised Classification and Clustering

Lars Schmarje, Monty Santarossa, Simon-Martin Schröder, Claudius Zelenka, Rainer Kiko, Jenny Stracke, Nina Volkmann, Reinhard Koch

ECCV 2022 · Label-efficient and robust learning › Semi-supervised learning · approach: Confidence-based pseudo-label selection · 10 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 argmax of l-hat as label instead of a sampled labelCE · Planktonmacro F1-Score: 86.71 -> 88.35Improved
use argmax of l-hat as label instead of a sampled labelFixMatch · Planktonmacro F1-Score: 80.29 -> 90.24Improved
use argmax of l-hat as label instead of a sampled labelCE · CIFAR-10Hmacro F1-Score: 67.71 -> 68.89Improved
remove ambiguous images from the datasetMean-Teacher · Planktonmacro F1-Score: 88.94 -> 96.00Improved
remove ambiguous images from the datasetFixMatch · CIFAR-10Hmacro F1-Score: 79.15 -> 90.37Improved
remove ambiguous images from the datasetCE · CIFAR-10Hmacro F1-Score: 68.89 -> 86.57Improved
add DC3 to Mean-Teacher at 10% labeledMean-Teacher · MiceBone variantbest inner distance d: 0.4639 -> 0.2590Improved
add DC3 to Mean-Teacher at 100% labeledMean-Teacher · MiceBone variantbest F1: 0.7254 -> 0.5692Hurt
add DC3 to Mean-Teacher at 50% labeledMean-Teacher · MiceBone variantbest F1: 0.7348 -> 0.6862Hurt
add CE-1 loss to classification and clustering (CC) (ablation)DC3 · pooled ablation (excluding FixMatch and Mice Bone)clustering inner distanceImproved
add CE-1 loss to CC to form full DC3 (ablation)CC only · pooled ablation (excluding FixMatch and Mice Bone)stability, non-degenerated runsImproved
use realistic prior ambiguity instead of p_A = 0.6 (ablation)DC3 · pooled ablation (excluding FixMatch and Mice Bone)F1-Score and inner distance dMixed

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