AI Paper Atlas

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

Deep Hashing with Active Pairwise Supervision

Ziwei Wang, Quan Zheng, Jiwen Lu, Jie Zhou

ECCV 2020 · Label-efficient and robust learning › Semi-supervised learning · approach: Holistic semi-supervised pipelines · 12 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
increase labeled-sample ratio from 1% to 30% (ablation)DH-APS · CIFAR-10 (32-bit)MAP (%): 32.6 -> 63.5Improved
increase labeled ratio from 50% to 100% (ablation)DH-APS · CIFAR-10 (32-bit)MAP (%): 64.7 -> 66.1No clear effect
add active sample selection at 1% labels vs 0% (ablation)DH-APS · CIFAR-10 (32-bit)MAP (%): 18.1 -> 32.6Improved
use all three acquisition terms (uncertainty, representativeness, diversity) (ablation)DH-APS · CIFAR-10 (32-bit, 1% labeled)MAP (%): 25.5 -> 32.6Improved
add uncertainty term to acquisition function (ablation)DH-APS · CIFAR-10 (32-bit)MAP (%)Improved
add representativeness term (ablation)DH-APS · CIFAR-10 (32-bit)MAP (%)Improved
add diversity term (ablation)DH-APS · CIFAR-10 (32-bit)MAP (%)Improved
use DH-APS as the sampler for labeled instances in PTS3HPTS3H · CIFAR-10 (32-bit)MAP (%): 83.5 -> 86.7Improved
use DH-APS as the sampler for labeled instances in PTS3HPTS3H · ImageNet-100 (32-bit)MAP (%): 68.0 -> 70.3Improved
use DH-APS with 1% labels instead of unsupervised GraphBithashing · CIFAR-10 (32-bit)MAP (%): 27.0 -> 32.6Improved

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