Research map / Label-efficient and robust learning / Recognition and long-tail
Dual-Curriculum Contrastive Multi-Instance Learning for Cancer Prognosis Analysis with Whole Slide Images
CHAO TU, YU ZHANG, Zhenyuan Ning
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 changed | Where | Measured | Effect |
|---|---|---|---|
| add multi-magnification strategy (MM) to Curriculum I (ablation) | DC_MIL · TCGA COAD/LIHC/BLCA | C-index: CI gain of 0.039 and AUC gain of 0.061 | Improved |
| add saliency-guided cross-scale encoding (SG) (ablation) | DC_MIL · TCGA COAD/LIHC/BLCA | C-index: average CI drops of 0.028 without it | Improved |
| add hierarchical transfer strategy (HT) (ablation) | DC_MIL · TCGA COAD/LIHC/BLCA | C-index: average CI drops of 0.011 without it | Improved |
| replace pre-trained ResNet-34 with Curriculum I weakly-supervised encoding (ablation) | DC_MIL · TCGA COAD | C-index: 0.660 -> 0.717 | Improved |
| add soft-bag learning to Cox baseline (ablation) | DC_MIL Curriculum II · TCGA 3 datasets avg | C-index: average CI gain of 0.024 and AUC gain of 0.031 | Improved |
| add constrained self-attention (CSA) (ablation) | DC_MIL Curriculum II · TCGA 3 datasets avg | C-index: average CI gains of 0.006 and AUC gains of 0.011 | Improved |
| add two-tier contrastive learning (TCL) (ablation) | DC_MIL Curriculum II · TCGA 3 datasets avg | C-index: average CI gains of 0.024 and AUC gains of 0.02 | Improved |
| add both CSA and TCL (ablation) | DC_MIL Curriculum II · TCGA 3 datasets avg | C-index: average CI gains of 0.032 and AUC gains of 0.04 | Improved |
| increase soft-bag size NB from 1 to 3 (ablation) | DC_MIL · TCGA COAD | C-index: 0.660 -> 0.717 | Improved |
| increase soft-bag size NB from 1 to 3 (ablation) | DC_MIL · TCGA LIHC | C-index: 0.631 -> 0.705 | Improved |
| increase soft-bag size NB from 1 to 4 (ablation) | DC_MIL · TCGA BLCA | C-index: 0.648 -> 0.686 | Improved |
| increase number of magnifications S from 1 to 3 (ablation) | DC_MIL · TCGA COAD | C-index: 0.663 -> 0.708 | Improved |
Closest papers by meaning
- DGR-MIL: Exploring Diverse Global Representation in Multiple Instance Learning for Whole Slide Image Classification (ECCV 2024 · similarity 0.89)
- Dynamic Policy-Driven Adaptive Multi-Instance Learning for Whole Slide Image Classification (CVPR 2024 · similarity 0.88)
- Interpretable Vision-Language Survival Analysis with Ordinal Inductive Bias for Computational Pathology (ICLR 2025 · similarity 0.87)
- CaMIL: Causal Multiple Instance Learning for Whole Slide Image Classification (AAAI 2024 · similarity 0.87)
- Unsupervised Foundation Model-Agnostic Slide-Level Representation Learning (CVPR 2025 · similarity 0.86)
- From Representation Space to Prognostic Insights: Whole Slide Image Generation with Hierarchical Diffusion Model for Survival Prediction (AAAI 2025 · similarity 0.86)
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.
