Research map / Label-efficient and robust learning / Recognition and long-tail
Modeling Long-Tail Relations in the Operating Room via In-Context Multimodal Learning
Boqiang Xu, Wei Zhang, Ding Ma, Jian Liang, Zhenan Sun, Zhen Lei
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 selective routing (Adaptive Router) so ICL is applied only to long-tail samples (ablation) | SGG-ICL · MM-OR (OR scene graph generation) | macro-average F1: 0.412 -> 0.546 | Improved |
| add Adaptive Router (ablation) | SGG-ICL · MM-OR (OR scene graph generation), head classes | F1: 0.622 -> 0.693 | Improved |
| add MLLM Reranker to pick ICL examples (ablation) | SGG-ICL · MM-OR (OR scene graph generation), long-tail classes | F1: 0.298 -> 0.358 | Improved |
| add MLLM Reranker (ablation) | SGG-ICL · MM-OR (OR scene graph generation) | top-1 example accuracy (%): 48.9 -> 72.8 | Improved |
| increase number of ICL examples top-k from 1 to 3 (ablation) | SGG-ICL · MM-OR (OR scene graph generation) | F1 | Hurt |
| raise margin threshold Gamma to filter low-similarity ICL queries (ablation) | SGG-ICL · MM-OR (OR scene graph generation) | F1 | Mixed |
| apply SGG-ICL instead of MM2SG | MM2SG · MM-OR (OR scene graph generation), long-tail classes | F1: 0.289 -> 0.358 | Improved |
| apply SGG-ICL instead of MM2SG | MM2SG · MM-OR (OR scene graph generation), head classes | F1: 0.699 -> 0.693 | Hurt |
| apply SGG-ICL instead of MM2SG | MM2SG · MM-OR (OR scene graph generation), Cutting | F1: 0.000 -> 0.000 | No clear effect |
| apply SGG-ICL instead of MM2SG | MM2SG · MM-OR (OR scene graph generation), Suturing | F1: 0.000 -> 0.209 | Improved |
Closest papers by meaning
- MM-OR: A Large Multimodal Operating Room Dataset for Semantic Understanding of High-Intensity Surgical Environments (CVPR 2025 · similarity 0.92)
- EgoExOR: An Ego-Exo-Centric Operating Room Dataset for Surgical Activity Understanding (NeurIPS 2025 · similarity 0.85)
- Multi-scale Temporal Prediction via Incremental Generation and Multi-agent Collaboration (NeurIPS 2025 · similarity 0.84)
- Procedure-Aware Surgical Video-language Pretraining with Hierarchical Knowledge Augmentation (NeurIPS 2024 · similarity 0.83)
- MedMO: Grounding and Understanding Multimodal Large Language Model for Medical Images (CVPR 2026 · similarity 0.83)
- SurgCoT: Advancing Spatiotemporal Reasoning in Surgical Videos through a Chain-of-Thought Benchmark (CVPR 2026 · similarity 0.83)
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.
