一种基于强化学习的招聘用虚假用工信息识别与预警方法及系统
By employing a reinforcement learning-based method for identifying and warning of fraudulent employment information, utilizing three-level priority data collection and four-dimensional state space modeling, combined with single-agent decision-making based on a deep Q-network architecture, the problem of iterative lag in traditional models is solved, enabling real-time identification and proactive prevention of fraudulent employment information, and reducing manual maintenance costs.
Patent Information
- Application Number
- CN202610563050.3
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Filing Date
- 2026-04-27
- Publication Date
- 2026-07-17
AI Technical Summary
In existing technologies, traditional fake employment identification models rely on training with historical samples, which cannot effectively learn new fake employment tactics. Their iteration lags behind the updates of black market tactics, resulting in low identification accuracy and requiring a lot of manual maintenance, making it difficult to deal with new fake employment behaviors.
A reinforcement learning-based approach is adopted, which constructs a four-dimensional state space by collecting data with three-level priority and parsing unstructured data in real time. A single agent with a deep Q-network architecture makes decisions and uses cold-start pre-training to achieve real-time identification and early warning of false employment information.
It improves the response speed of fraudulent employment identification to the millisecond level, proactively uncovers unknown illegal practices, reduces manual maintenance costs, transforms into proactive prevention and control, and enhances the overall timeliness and real-time nature of risk discovery cycles.
Smart Images
Figure CN122415045A_ABST
Abstract
Citation Information
Patent Citations
Intelligent network false information identification and early warning system based on multi-modal behavior map
CN120145279A