Research map / Vision-language and multimodal / Cross-modal learning
Together, Then Apart: Balancing Alignment and Distinctiveness for Multimodal Survival Analysis
Wenjing Liu, Qin Ren, Wen Zhang, Yuewei Lin, Chenyu You
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 Together stage (shared prototypes + UOT alignment) (ablation) | TTA baseline · TCGA average over 5 cohorts | C-index: 0.643 -> 0.662 | Improved |
| add Apart stage (anchor refiner + contrastive objective) (ablation) | TTA baseline · TCGA average over 5 cohorts | C-index: 0.643 -> 0.666 | Improved |
| add both Together and Apart (ablation) | TTA baseline · TCGA average over 5 cohorts | C-index: 0.643 -> 0.693 | Improved |
| add Apart stage on CRC (ablation) | Together-only TTA · TCGA CRC | C-index: 0.587 -> 0.685 | Improved |
| use shared prototypes instead of modality-specific (ablation) | TTA Together stage · TCGA average over 5 cohorts | C-index: 0.681 -> 0.684 | Improved |
| add joint UOT over concatenated prototypes (ablation) | TTA Together stage · TCGA average over 5 cohorts | C-index: 0.684 -> 0.693 | Improved |
| add anchor refinement alone (ablation) | TTA Apart stage · TCGA average over 5 cohorts | C-index: 0.662 -> 0.669 | Improved |
| add contrastive regularization alone (ablation) | TTA Apart stage · TCGA average over 5 cohorts | C-index: 0.662 -> 0.672 | Improved |
| combine anchor refinement and contrastive regularization (ablation) | TTA Apart stage · TCGA average over 5 cohorts | C-index: 0.662 -> 0.693 | Improved |
| use UOT assignment instead of hard KMeans (ablation) | TTA · TCGA average over 5 cohorts | C-index: 0.681 -> 0.693 | Improved |
| use UOT instead of general OT (ablation) | TTA · TCGA average over 5 cohorts | C-index: 0.683 -> 0.693 | Improved |
| add curriculum mass rho (ablation) | TTA UOT · TCGA average over 5 cohorts | C-index: 0.686 -> 0.693 | Improved |
Closest papers by meaning
- Prototypical Information Bottlenecking and Disentangling for Multimodal Cancer Survival Prediction (ICLR 2024 · similarity 0.90)
- CA-MLIF: Cross-Attention and Multimodal Low-Rank Interaction Fusion Framework for Tumor Prognostic Prediction (AAAI 2025 · similarity 0.89)
- Multimodal Prototyping for cancer survival prediction (ICML 2024 · similarity 0.88)
- HEALNet: Multimodal Fusion for Heterogeneous Biomedical Data (NeurIPS 2024 · similarity 0.88)
- Distilled Prompt Learning for Incomplete Multimodal Survival Prediction (CVPR 2025 · similarity 0.87)
- When Are Multimodal Predictions Biologically Supported? A Diagnostic Evaluation Framework (NeurIPS 2026 · similarity 0.87)
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.
