AI Paper Atlas

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

Deep Semantic Clustering by Partition Confidence Maximisation

Jiabo Huang, Shaogang Gong, Xiatian Zhu

CVPR 2020 · Label-efficient and robust learning › Recognition and long-tail · approach: Loss design and classifier theory · 202 citations (OpenAlex)

Check 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 partition-confidence clustering instead of DCCMPICA · CIFAR-10ACC: 0.623 -> 0.645Improved
use partition-confidence clustering instead of DCCMPICA · CIFAR-100ACC: 0.327 -> 0.322No clear effect
use partition-confidence clustering instead of IICPICA · STL-10ACC: 0.610 -> 0.693Improved
use partition-confidence clustering instead of DCCMPICA · ImageNet-10ACC: 0.710 -> 0.850Improved
use partition-confidence clustering instead of DCCMPICA · ImageNet-DogsACC: 0.383 -> 0.324Hurt
use partition-confidence clustering instead of DCCMPICA · Tiny-ImageNetACC: 0.108 -> 0.094Hurt
reduce entropy regularisation weight from 7 to 3 (ablation)PICA · CIFAR-10ACC: 0.246 -> 0.696Improved
reduce entropy regularisation weight from 7 to 3 (ablation)PICA · CIFAR-100ACC: 0.168 -> 0.330Improved
reduce entropy regularisation weight from 7 to 3 (ablation)PICA · ImageNet-10ACC: 0.650 -> 0.829Improved
add over-clustering head (ablation)PICA · CIFAR-10ACC: 0.582 -> 0.696Improved
add over-clustering head (ablation)PICA · STL-10ACC: 0.633 -> 0.687Improved
use the paper's standard perturbation setting rather than setting 7 (ablation)PICA · CIFAR-10ACC: 0.310 -> 0.696Improved

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