Researchers have developed a new method for intrusion detection in network traffic that utilizes conformal prediction to provide statistical validity guarantees. This approach, termed traffic-aware conformal prediction, calibrates on data from potential attacker mechanisms to maintain coverage even when adversaries perturb network features. For more adaptive attackers, the system excludes controllable features to ensure an exact coverage guarantee, though this may result in a slight decrease in clean accuracy. AI
排序理由 The cluster contains a research paper published on arXiv detailing a new method for network intrusion detection. [lever_c_demoted from research: ic=1 ai=1.0]
- arXiv
- Attack-Orbit Invariance
- Conformal Intrusion Detection
- Conformal prediction
- Hugging Face
- large-language models
- Traffic-Aware Calibration
AI 生成摘要 · Google Gemini · 来自 1 个来源。 我们如何撰写摘要 →
