AI Paper Atlas

Research map / Architectures and efficiency / Neural architecture search

SparseMask: Differentiable Connectivity Learning for Dense Image Prediction

Huikai Wu, Junge Zhang, Kaiqi Huang

ICCV 2019 · Architectures and efficiency › Neural architecture search · approach: Latency-aware architecture search · 15 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 batch normalization before the connection weights (ablation)SparseMask · PASCAL VOC 2012 valmIoU: 72.21% -> 73.18%Improved
add the binary part of the sparse loss to the task loss (ablation)FDN search · PASCAL VOC 2012 valmIoU: 72.72% -> 73.23%Improved
add the sparse part (L_sp) on top of L_task + L_bp (ablation)SparseMask · PASCAL VOC 2012 valmIoU / parameters: 73.23% -> 73.18%No clear effect
add the sparse part alone to the task loss (ablation)FDN search · PASCAL VOC 2012 valmIoU: 72.72% -> 72.04%Hurt
replace the dense FDN with the searched sparse connectivitySparseMask · PASCAL VOC 2012 valFPS: 37 -> 98Improved
replace U-Net with the searched connectivityMobileNet-V2 decoder · PASCAL VOC 2012 valmIoU: 64.72% -> 73.18%Improved
replace SparseACN with the searched connectivityMobileNet-V2 decoder · PASCAL VOC 2012 valmIoU: 72.23% -> 73.18%Improved
replace FCN with the searched connectivityMobileNet-V2 decoder · PASCAL VOC 2012 valmIoU: 63.80% -> 73.18%Improved
replace Deeplab-V3 decoder with the searched connectivityMobileNet-V2 · PASCAL VOC 2012 valmIoU / FPS: 72.51% -> 73.18%; 50 -> 98 FPSImproved
replace the proposed sparse loss with L1 loss (ablation)SparseMask search · PASCAL VOC 2012 valFPS / mIoU: nearly 3 times slower at 1e-3Hurt
increase L1 loss strength from 1e-3 to 1e-2 (ablation)L1-based search · PASCAL VOC 2012 valmIoU / FPSMixed
increase MaskConnect k from 2 to 5 (ablation)MaskConnect search · PASCAL VOC 2012 valmIoUImproved

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