A bidding evaluation and review detection method, device, equipment and storage medium
By constructing a multimodal knowledge graph through multimodal parsing and multi-agent collaboration mechanisms, the existing bidding and tendering review system is unable to identify low-level image features and causal logic contradictions, thus realizing comprehensive intelligent review of bidding and tendering documents and improving the accuracy and reliability of detection.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- SHENZHEN XUNCE TECH CO LTD
- Filing Date
- 2026-03-18
- Publication Date
- 2026-07-17
AI Technical Summary
Existing bidding and tendering review systems struggle to identify the underlying physical features of images, cannot recognize carefully spliced and tampered discrepancies between images and text, cannot uncover causal logical contradictions behind data, and are unable to effectively identify AI-generated false performance descriptions and heavily rewritten tender documents, lacking the ability to analyze the origin of the data.
By acquiring bidding documents, tender documents, and external dynamic data streams through multimodal parsing, a multimodal knowledge graph is constructed. Adversarial game reasoning is then conducted using physical evidence-gathering agents, causal reasoning agents, and topology analysis agents to generate bidding review reports.
It enables comprehensive intelligent review of bidding documents, improves the identification rate of complex fraudulent behaviors, ensures professional division of labor and in-depth collaboration in review results, and enhances the accuracy and reliability of detection.
Smart Images
Figure CN122415202A_ABST