Research map / Detection and segmentation / Open-set and few-shot detection
From Objects to Events: Unlocking Complex Visual Understanding in Object Detectors via LLM-guided Symbolic Reasoning
Yuhui Zeng, Haoxiang Wu, Wenjie Nie, Guangyao Chen, Xiawu Zheng, Yunhang Shen, Jun Peng, Yonghong Tian, Rongrong Ji
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 symbolic regression pattern discovery over manual logic (ablation) | SymbolicDet · average over datasets | performance (%): 67.00% to 85.36% | Improved |
| add SymbolicDet to the detector | APE · ERA BALL | AUROC (%): 55.36 -> 94.91 | Improved |
| add SymbolicDet to the detector | YOLO-World · ERA PersonCrowd | AUROC (%): 55.00 -> 85.11 | Improved |
| add SymbolicDet to the detector | APE · Helmet-Mac | AUROC (%): 67.41 -> 83.18 | Improved |
| add SymbolicDet to the detector | GLIP · Multi-rods Fishing | AUROC (%): 50.00 -> 71.11 | Improved |
| add SymbolicDet to the detector | APE · Multi-rods Fishing | AUROC (%): 66.82 -> 75.16 | Improved |
| add SymbolicDet on top of LoRA fine-tuning | APE + LoRA · Helmet-Mac | AUROC (%): 84.11 -> 95.67 | Improved |
| add SymbolicDet on top of LoRA fine-tuning | APE + LoRA · Multi-rods Fishing | AUROC (%): 75.86 -> 78.44 | Improved |
| apply prompt tuning to the detector | APE · Helmet-Mac | AUROC (%): 67.41 -> 67.42 | No clear effect |
| apply LoRA fine-tuning instead of training-free SymbolicDet | APE · Helmet-Mac | AUROC (%): 83.18 -> 84.11 | Mixed |
| scale up the LLM from 7B to 72B guidance model | SymbolicDet search · computational cost | cost time (s): 45.92 -> 26.07 | Improved |
| increase search scale (ablation) | SymbolicDet · Helmet and Fishing | AUROC | Improved |
Closest papers by meaning
- LEOD: Label-Efficient Object Detection for Event Cameras (CVPR 2024 · similarity 0.86)
- PASS: Path-selective State Space Model for Event-based Recognition (NeurIPS 2025 · similarity 0.85)
- FlexEvent: Towards Flexible Event-Frame Object Detection at Varying Operational Frequencies (NeurIPS 2025 · similarity 0.85)
- Black Swan: Abductive and Defeasible Video Reasoning in Unpredictable Events (CVPR 2025 · similarity 0.85)
- OpenESS: Event-based Semantic Scene Understanding with Open Vocabularies (CVPR 2024 · similarity 0.85)
- ExACT: Language-guided Conceptual Reasoning and Uncertainty Estimation for Event-based Action Recognition and More (CVPR 2024 · similarity 0.84)
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.
