An active defense and access control method for deep learning models

By generating license sequences and masks through Chebyshev chaotic mapping and combining bi-state training and gradient freezing, the problem of intellectual property protection for deep learning models in unauthorized situations is solved, realizing lightweight proactive defense and access control, suitable for edge devices.

CN122027368BActive Publication Date: 2026-06-23XIANGTAN UNIV
2 Cites 0 Cited by

Patent Information

Application Number
CN202610488916.9
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2026-04-14
Publication Date
2026-06-23
Estimated Expiration
2046-04-14

AI Technical Summary

Technical Problem

Existing intellectual property protection schemes for deep learning models are difficult to prevent unauthorized access in advance. Some schemes rely on dedicated hardware or cryptographic reasoning, resulting in high deployment barriers. Lightweight model-intrinsic lock schemes struggle to balance authorized accuracy with unauthorized degradation and resistance to adaptive attacks.

Method used

By generating deterministic scalar license sequences based on Chebyshev chaotic mapping, a set of license channels is constructed and a mask is generated. Combined with bi-state training and a structure-sensitive gradient freezing mechanism, the model can switch between licensed and unlicensed states, thereby enhancing its robustness against adaptive attacks.

Benefits of technology

It enables efficient switching of the available/unavailable state of the model without the need for external encryption shells or dedicated hardware, improves robustness against attacks such as fine-tuning and pruning, maintains the performance gap between authorized and unauthorized models, and is suitable for deployment on resource-constrained edge devices.

✦ Generated by Eureka AI based on patent content.
Patent Text Reader

Abstract

The application belongs to the technical field of artificial intelligence security and model intellectual property protection, and discloses an active defense and access control method for a deep learning model, which comprises the following steps: S1, generating a scalar license sequence based on a private key parameter, and dividing a training round into multiple stages, so that the stage indexes are consistent and the license values corresponding to the indexes in the license sequence are read; S2, constructing an authorized channel set in combination with channel static importance evaluation and channel chaos score, generating an authorized mask and an unauthorized mask; S3, respectively calculating authorized task loss and unauthorized confusion loss and jointly optimizing, and simultaneously implementing reset and gradient freezing on the parameters corresponding to the authorized channel set at the beginning of the training stage; S4, during model deployment and reasoning, the authorized end reconstructs the license sequence according to the shared key and generates the authorized mask for reasoning, and the unauthorized end outputs performance degradation. The application improves the robustness to adaptive attacks and maintains the performance gap between authorized and unauthorized.
Need to check novelty before this filing date? Find Prior Art