AI Paper Atlas

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

Jaccard Metric Losses: Optimizing the Jaccard Index with Soft Labels

Zifu Wang, Xuefei Ning, Matthew B. Blaschko

NeurIPS 2023 · Label-efficient and robust learning › Semi-supervised learning · approach: Pseudo-labeling for specialized tasks · 14 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
Replace CE-only training with JML plus CEDL3-R18 segmentation · CityscapesmIoU: 72.88±0.44 -> 75.55±0.13Improved
Replace CE-only training with JML plus CEDL3-R18 segmentation · PASCAL VOCmIoU: 72.47±0.33 -> 74.42±0.52Improved
Add boundary label smoothing to JMLDL3-R18 segmentation · PASCAL VOCmIoU: 74.42±0.52 -> 75.60±0.24Improved
Add JML-based teacher distillationDL3-R18 segmentation · CityscapesmIoU: 75.55±0.13 -> 76.68±0.33Improved
Use JML-based SSL with boundary label smoothingAugSeg, DL3+-R101 · PASCAL VOC 92-label splitmIoU: 71.09 -> 76.43Improved
Set CE/JML loss weights to 0.25/0.75 rather than CE-only (ablation)DL3-R101 JML · PASCAL VOCmIoU: 78.39±0.09 -> 80.26±0.45Improved
Set CE/JML loss weights to 0.50/0.50 instead of 0.25/0.75 (ablation)DL3-R101 JML · PASCAL VOCmIoU: 80.26±0.45 -> 79.92±0.12Hurt
Add the teacher's CE and JML distillation loss terms (ablation)DL3-R18 student with supervised JML · CityscapesmIoU: 75.55±0.13 -> 76.16±0.37Improved
Train the teacher with JML-BLS (ablation)DL3-R18 JML-KD · PASCAL VOCmIoU: 75.05±0.31 -> 75.89±0.29Improved
Filter classes based on teacher confidence instead of including all classes (ablation)DL3-R18 JML-KD · PASCAL VOCmIoU: 75.21±0.33 -> 75.89±0.29Improved
Use only the teacher's maximum-confidence class rather than confidence-filtered classes (ablation)DL3-R18 JML-KD · PASCAL VOCmIoU: 75.89±0.29 -> 74.50±0.25Hurt
Add supervised JML loss to the supervised-plus-unlabeled CE-SSL objective (ablation)DL3+-R101 JML-SSL · PASCAL VOC 92-label splitmIoU: 71.09 -> 72.89Improved

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