Research map / Vision-language and multimodal / Cross-modal learning
Unify, Align and Refine: Multi-Level Semantic Alignment for Radiology Report Generation
Yaowei Li, Bang Yang, Xuxin Cheng, Zhihong Zhu, Hongxiang Li, Yuexian Zou
This paper tried a common change: Add a contrastive loss
add triplet contrastive loss · IU-Xray · CIDEr: 0.271 -> 0.409 Improved
What 1,317 accepted papers report for this change
- 1,253 improved
- 22 no clear effect
- 16 mixed
- 26 hurt
- Beyond Looking Up, Try Looking Around: Harmonizing Global Structure and Local Consistency in Optimal Transport for Short Text Clustering ICML 2026 · add contrastive learning · SearchSnippets · ACC: 56.65 -> 80.12Improved
- SleepLM: Natural-Language Intelligence for Human Sleep ICML 2026 · add contrastive objective to captioning (CoCa vs Cap) · sleep event IoU · IoU: 29.2 -> 27.4Hurt
- LIMSSR: LLM-Driven Sequence-to-Score Reasoning under Training-Time Incomplete Multimodal Observations ICML 2026 · add contrastive loss L_con · FS1000 average · Spearman correlation: 0.759 -> 0.789Improved
See this change in detailMatched by the wording of each paper's own result, read by AI from the full text. Papers report gains more often than failures, so this shows what gets published, not the chance a change works.
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 |
|---|---|---|---|
| replace ResNet101 with dVAE discrete visual tokens (LSU) (ablation) | Transformer-Base · IU-Xray | BLEU-4: 0.144 -> 0.177 | Improved |
| add LSU (ablation) | Base · IU-Xray | alignment score: 36% -> 49% | Improved |
| add CRA feature integration (orthogonal basis + dual gate) (ablation) | Base · IU-Xray | CIDEr: 0.271 -> 0.330 | Improved |
| add triplet contrastive loss (ablation) | Base · IU-Xray | CIDEr: 0.271 -> 0.409 | Improved |
| add LSU to CRA model (ablation) | Base + CRA · IU-Xray | CIDEr: 0.393 -> 0.410 | Improved |
| add TIR (ablation) | Base + LSU + CRA · IU-Xray | CIDEr: 0.410 -> 0.460 | Improved |
| add two-stage training (ablation) | LSU + CRA + TIR · IU-Xray | CIDEr: 0.460 -> 0.501 | Improved |
| use orthogonal subspace instead of uniform or normal (ablation) | CRA · IU-Xray | CIDEr: 0.428 / 0.427 -> 0.501 | Improved |
| use dual gate instead of no gate or addition fusion (ablation) | CRA · IU-Xray | CIDEr: 0.405 / 0.376 -> 0.501 | Improved |
| UAR versus CMM+RL (reinforcement learning) | UAR · MIMIC-CXR | BLEU-4: 0.109 vs 0.107 | No clear effect |
Closest papers by meaning
- Automatic Radiology Reports Generation via Memory Alignment Network (AAAI 2024 · similarity 0.91)
- S2D-Align: Shallow-to-Deep Auxiliary Learning for Anatomically-Grounded Radiology Report Generation (AAAI 2026 · similarity 0.91)
- DART: Disease-aware Image-Text Alignment and Self-correcting Re-alignment for Trustworthy Radiology Report Generation (CVPR 2025 · similarity 0.90)
- RadZero: Similarity-Based Cross-Attention for Explainable Vision-Language Alignment in Chest X-ray with Zero-Shot Multi-Task Capability (NeurIPS 2025 · similarity 0.90)
- Bootstrapping Large Language Models for Radiology Report Generation (AAAI 2024 · similarity 0.88)
- CARZero: Cross-Attention Alignment for Radiology Zero-Shot Classification (CVPR 2024 · similarity 0.88)
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.
