Research map / Vision-language and multimodal / Cross-modal learning
Scalable Medical Multimodal Fusion via Symmetric Consistency Modeling
Xiaowen Sun, Hui Liu, Gongguan Chen, Ning Mao
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 the proposed multi-view consistency fusion instead of an MLP aggregator (ablation) | Ours · CMMD | ACC (%): 91.86 -> 92.37 | Improved |
| add the proposed multi-view consistency fusion instead of an MLP aggregator (ablation) | Ours · Breast Cancer (self-constructed) | ACC (%): 70.52 -> 71.64 | Improved |
| add the proposed fusion instead of cross-attention aggregation (ablation) | Ours · CMMD | ACC (%): 92.01 -> 92.37 | Improved |
| add the proposed fusion instead of cross-attention aggregation (ablation) | Ours · Breast Cancer (self-constructed) | ACC (%): 70.88 -> 71.64 | Improved |
| use deterministic linear reconstruction instead of MLP nonlinear mapping at token level (ablation) | Ours · CMMD | ACC (%): 91.68 -> 92.37 | Improved |
| use deterministic linear reconstruction instead of MLP nonlinear mapping at token level (ablation) | Ours · Breast Cancer (self-constructed) | ACC (%): 70.93 -> 71.64 | Improved |
| add Gaussian semantic uncertainty modeling (ablation) | Ours · CMMD | t-SNE separability | Improved |
| replace cosine-similarity token alignment with linear-reconstruction token consistency (ablation) | Ours · Breast Cancer (self-constructed) | channel activation consistency | Improved |
| add signal-noise decomposition with denoising (ablation) | Ours · Breast Cancer (self-constructed) | per-class ROC AUC | Improved |
| add multi-view fusion instead of direct concatenation (ablation) | Ours · Breast Cancer (self-constructed) | per-class ROC | Improved |
| increase modality set from {1,2,3} to {1,2,3,4} | Ours · Breast Cancer (self-constructed) | ACC (%): 73.92 -> 75.38 | Improved |
| increase modality set from {1,2} to {1,2,3} | Ours · Gastric Cancer (self-constructed) | ACC (%): 73.11 -> 74.62 | Improved |
Closest papers by meaning
- Modalities Contribute Unequally: Enhancing Medical Multi-modal Learning through Adaptive Modality Token Re-balancing (ICML 2025 · similarity 0.91)
- DrFuse: Learning Disentangled Representation for Clinical Multi-Modal Fusion with Missing Modality and Modal Inconsistency (AAAI 2024 · similarity 0.88)
- HEALNet: Multimodal Fusion for Heterogeneous Biomedical Data (NeurIPS 2024 · similarity 0.87)
- Multimodal Lego: Model Merging and Fine-Tuning Across Topologies and Modalities in Biomedicine (ICLR 2025 · similarity 0.87)
- Flex-MoE: Modeling Arbitrary Modality Combination via the Flexible Mixture-of-Experts (NeurIPS 2024 · similarity 0.87)
- Multi-modal Medical Diagnosis via Large-small Model Collaboration (CVPR 2025 · 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.
