基于自适应生成式AI的地质灾害应急决策方法及系统
By employing an adaptive generative AI approach, the instability and uncertainty of models in landslide disaster emergency response without prior knowledge were addressed. This approach enabled efficient and reliable generation of disaster assessment reports and model self-evolution, thereby improving the accuracy of emergency response and the robustness of the system.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- CHINA UNIV OF GEOSCIENCES (WUHAN)
- Filing Date
- 2025-10-20
- Publication Date
- 2026-07-17
AI Technical Summary
In emergency response scenarios involving sudden landslide disasters where there is no prior knowledge or manual annotation, existing technologies struggle to provide stable adaptive testing. They lack the model's self-awareness, fail to effectively quantify uncertainty, and lack an explicit and updatable knowledge system, resulting in insufficient reliability and accuracy of assessment reports.
An adaptive generative AI approach is adopted, which involves data reception, semantic segmentation, feature extraction, prompt construction, and report generation. By combining uncertainty quantification and gating mechanisms, a professional knowledge base is dynamically maintained to achieve model self-evolution and efficient analysis, thereby reducing low-confidence targets.
It achieves efficient analysis in unlabeled emergency scenarios, improves the reliability and accuracy of assessment reports, reduces the frequency of low-confidence targets, and has the ability to self-evolve, continuously improving recognition accuracy and robustness.
Smart Images
Figure CN121352010B_ABST