发表机构
University of Dhaka; Tennessee State University; Military Institute of Science and Technology; Huston–Tillotson University; Howard University（达卡大学; 田纳西州立大学; 军事科学技术学院; 休斯顿-蒂洛森大学; 霍华德大学）
机构由 AI 辅助整理，请以论文原文为准。AI 中文总结
针对LOTL APT攻击，提出融合BERT、字符级CNN、注意力机制和自编码器的多路径架构SENTINEL，在Volt Typhoon基准上对混淆命令检测准确率达91.2%，优于基线方法。
AI 中文摘要
Living-Off-the-Land (LOTL) 是高级持续性威胁（APT）行为者使用的主要规避技术，利用合法的Windows实用程序进行恶意操作，而无需部署自定义恶意软件，并使国家支持的攻击活动能够在军事和关键国防基础设施中长时间保持持久访问。现有检测方法无法应对混淆命令和多阶段攻击序列，正如Volt Typhoon APT攻击活动所证明的那样，该活动仅使用签名的Windows实用程序，在未被发现的情况下访问美国关键基础设施超过18个月。我们提出了SENTINEL，一种多路径架构，集成了基于BERT的语义编码、用于混淆不变性的字符级CNN、用于多阶段模式识别的命令间注意力机制，以及基于自编码器的异常评分。在从Microsoft和CISA威胁情报公告中得出的平衡Volt Typhoon基准上进行评估，SENTINEL在记录在案的国家支持的攻击命令上达到92.0%的准确率，在混淆变体上达到91.2%，而独立BERT分别为74.0%和72.0%。按类别分析显示，在不平衡数据上实现超过98%总体验证准确率的模型，在平衡对抗集上仅表现出44-58%的恶意召回率。字符级处理贡献了5.6个百分点的混淆不变性，与仅增强基线相比8.0个百分点的差距证实了结构架构价值超越了仅数据驱动的鲁棒性。
英文摘要
Living-Off-the-Land (LOTL) is the dominant evasion technique of Advanced Persistent Threat (APT) actors, exploiting legitimate Windows utilities to conduct malicious operations without deploying custom malware and enabling state-sponsored campaigns to maintain persistent access within military and critical defense infrastructure for extended periods. Existing detection methods fail against obfuscated commands and multi-stage attack sequences, as demonstrated by the Volt Typhoon APT campaign, which maintained undetected access to U.S. critical infrastructure for over 18 months using exclusively signed Windows utilities. We present SENTINEL, a multi-pathway architecture integrating BERT-based semantic encoding, character-level CNN for obfuscation invariance, inter-command attention for multi-stage pattern recognition, and autoencoder-based anomaly scoring. Evaluated on a balanced Volt Typhoon benchmark derived from Microsoft and CISA threat intelligence advisories, SENTINEL achieves 92.0% accuracy on documented state-sponsored attack commands and 91.2% on obfuscated variants, compared to 74.0% and 72.0% for standalone BERT. Per-class analysis reveals that models achieving over 98% overall validation accuracy on imbalanced data exhibit only 44-58% malicious recall on balanced adversarial sets. Character-level processing contributes 5.6 percentage points of obfuscation invariance, and the 8.0 percentage point gap over augmentation-only baselines confirms structural architectural value beyond data-driven robustness alone.
CommentsAccepted for publication in the Proceedings of the 2026 IEEE Military Communications Conference (MILCOM 2026). This is the authors' accepted version; the final published version will appear in IEEE Xplore. 7 pages, 6 figures, 1 table
