AI Paper Atlas

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

Contrastive Multi-Task Dense Prediction

Siwei Yang, Hanrong Ye, Dan Xu

AAAI 2023 · Label-efficient and robust learning › Recognition and long-tail · approach: Contrastive representation theory · 9 citations (OpenAlex)

Read the paper

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
Apply CCR to InvPTInvPT · NYUD-v2SemSeg mIoU: 52.84 -> 53.80Improved
Apply CCR to InvPTInvPT · Pascal-ContextNormal mErr: 14.15 -> 13.93Improved
Add basic cross-task contrastive regularization (ablation)MT Baseline · NYUD-v2SemSeg mIoU: 38.901 -> 39.499Improved
Add a shared feature projector (ablation)CCR-Basic · NYUD-v2SemSeg mIoU: 39.499 -> 39.956Improved
Add semi-hard pixel sampling (ablation)CCR + Proj · NYUD-v2SemSeg mIoU: 39.956 -> 40.779Improved
Add contrastive task-pair selection (ablation)CCR + Proj + SS · NYUD-v2SemSeg mIoU: 40.779 -> 41.275Improved
Share projectors among source tasks rather than leaving projectors unshared (ablation)CCR with SS and CTS · NYUD-v2SemSeg mIoU: 40.414 -> 41.275Improved
Share projectors among target tasks (ablation)CCR with SS and CTS · NYUD-v2SemSeg mIoU: 40.414 -> 40.273Hurt
Set the top-k factor to 128 (ablation)CCR · NYUD-v2Overall multi-task performanceImproved
Increase the sampling ratio γ (ablation)CCR · NYUD-v2Overall multi-task performanceImproved
Add CCRPAD-Net · NYUD-v2SemSeg mIoU: 35.406 -> 36.925Improved
Add CCRMTI-Net · NYUD-v2SemSeg mIoU: 37.456 -> 39.192Improved

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