3 results about "False rejection rate" patented technology

This invention discloses a method and system for suppressing false rejection behavior and selectively forgetting harmful semantics in large language models across vertical domains based on feature subspace decoupling, belonging to the field of fine-tuning and alignment technology for large artificial intelligence models. The method includes steps for target definition and data construction, feature subspace localization, key level optimization, orthogonal decoupling weight correction, constrained norm renormalization, selective fine-tuning, false rejection calibration, and closed-loop verification monitoring and document output. The system includes modules for concept discovery and data construction, representation subspace modeling, false rejection localization and calibration, and verifiable evaluation and continuous monitoring. This invention can significantly reduce the false rejection rate and maintain a safety baseline while preserving the model's generalizability, meeting the compliance requirements of generative artificial intelligence services.

This invention discloses a dynamic role-based access control method based on dual-source trust fusion to address the high risk of unauthorized access caused by static role authorization in dynamic network environments. The method first receives user access requests and extracts contextual risk features. Then, it extracts historical user interaction indicators and calculates dynamic behavior trust scores based on time decay. External feedback scores are obtained, outlier scores are removed, and weighted aggregation yields robust feedback trust scores. Then, based on data consistency and divergence, adaptive weighting is applied to derive a comprehensive trust score. Finally, the comprehensive trust score is coupled with the contextual risk value to determine the trust gap, and based on this, full authorization, partial authorization downgrade based on sensitivity truncation, or access denial are performed. This invention abandons static fixed weights, enhancing the robustness of trust assessment in extreme scenarios; and it pioneers a partial authorization pruning mechanism, effectively reducing the false rejection rate of legitimate users while ensuring security, achieving a balance between system availability and the principle of least privilege.

The application discloses a finger vein recognition method and system based on dynamic balance of false recognition rate and false rejection rate, and the method comprises the following steps: obtaining a to-be-recognized finger vein image, and performing image preprocessing and ROI extraction processing to obtain an ROI region of the finger vein image; a finger vein recognition model is constructed, three-stage hybrid training is performed in combination with the ROI region of the finger vein image, and a trained finger vein recognition model is obtained; a balanced error rate evaluation index is constructed to evaluate the trained finger vein recognition model, and an evaluated finger vein recognition model is obtained; and finger vein recognition is performed on the to-be-recognized finger vein image based on the evaluated finger vein recognition model, and a finger vein recognition result is obtained. The application can meet the finger vein recognition requirements in multiple scenes and improve the finger vein recognition accuracy. The application can be widely applied to the technical field of finger vein recognition as a finger vein recognition method and system based on dynamic balance of false recognition rate and false rejection rate.