基于域自适应的分类模型构建方法与网络入侵检测方法
By constructing a domain-adaptive classification model, utilizing open-source dataset label information and a reconstruction error loss function, the problem of unlabeled datasets in real network environments is solved, achieving efficient intrusion detection in the target domain.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Patents(China)
- Current Assignee / Owner
- COMP NETWORK INFORMATION CENT CHINESE ACADEMY OF SCI
- Filing Date
- 2022-07-06
- Publication Date
- 2026-07-17
AI Technical Summary
Existing intrusion detection classifiers based on open-source datasets struggle to achieve usable performance in real-world network environments, and the difficulty in labeling datasets in real-world network environments makes model training challenging.
We adopt a domain-adaptive classification model construction method. By combining encoder, decoder, classifier and domain discriminator, we use the label information of open source dataset to construct a hybrid target domain reconstruction dataset. We constrain the encoder mapping process by reconstruction error and classification loss function to retain the original data information to facilitate classification tasks.
In the case of unlabeled target domain, the intrusion detection model is trained using label information from open-source datasets, which solves the problem of difficult model training. Furthermore, by constraining the reconstruction error and classification loss function, the classification performance of the model in real network environments is improved.
Smart Images
Figure CN117435948B_ABST