一种基于多源数据的标注任务自适应优化方法
By performing contextual characterization and risk assessment on multi-source data annotation tasks, calculating the cost of context switching, and optimizing task scheduling based on user proficiency, the quality and efficiency issues caused by context switching in multi-source data annotation are resolved, resulting in a more efficient annotation process.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- XINGTUQIHANG ARTIFICIAL INTELLIGENCE TECHNOLOGY (ZIBO) CO LTD
- Filing Date
- 2025-11-17
- Publication Date
- 2026-07-17
AI Technical Summary
Existing technologies ignore the cost of context switching in multi-source heterogeneous data annotation tasks, resulting in a decrease in annotation quality and a loss of efficiency. In particular, the probability of errors by annotators is high during data switching, and the processing efficiency is temporarily reduced.
By generating context vectors through context feature generation of data, calculating the context switching cost of task switching, and combining user proficiency to calculate instantaneous annotation risk, an adaptive optimization strategy is executed to proactively prevent or compensate for annotation errors, thereby optimizing task scheduling and quality control.
It reduced the overall error rate of data annotation, reduced the cognitive load on annotators, increased the annotation speed, and optimized quality control costs.
Smart Images
Figure CN121543065B_ABST