AI Paper Atlas

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

Deep Comprehensive Correlation Mining for Image Clustering

Jianlong Wu, Keyu Long, Fei Wang, Chen Qian, Cheng Li, Zhouchen Lin, Hongbin Zha

ICCV 2019 · Label-efficient and robust learning › Recognition and long-tail · approach: Correspondence and causal representation learning · 221 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 local robustness (transformed-sample pseudo-graph and loss) (ablation)DCCM · CIFAR-10ACC: 0.405 -> 0.512Improved
add local robustness (ablation)DCCM · CIFAR-10NMI: 0.304 -> 0.412Improved
add pseudo-label loss (ablation)DCCM · CIFAR-10ACC: 0.512 -> 0.583Improved
add pseudo-label loss (ablation)DCCM · CIFAR-10NMI: 0.412 -> 0.448Improved
add triplet mutual information loss (ablation)DCCM · CIFAR-10ACC: 0.583 -> 0.623Improved
add triplet mutual information loss (ablation)DCCM · CIFAR-10ARI: 0.358 -> 0.408Improved
raise pseudo-label confidence threshold (ablation)DCCM · CIFAR-10clustering performanceImproved
sample negatives randomly instead of farthest negatives for MI pairs (ablation)DCCM · CIFAR-10classification ACC (Y64): 0.713 -> 0.744Improved
sample positives nearest instead of random for MI pairs (ablation)DCCM · CIFAR-10classification ACC (Y64): 0.731 -> 0.744Improved
sample uniformly in the minibatch instead of top-n confident positives (ablation)DCCM · CIFAR-10classification ACC (Y64): 0.698 -> 0.744Improved
replace DAC with DCCM under same architectureDCCM · CIFAR-10clustering ACC: 0.514 -> 0.623Improved
replace DAC with DCCMDCCM · CIFAR-100 (20 superclasses)ACC: 0.238 -> 0.327Improved

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