AI 中文总结
针对商业广告账户被入侵的检测难题，本文提出将LLM智能体作为信号提取器，结合归纳逻辑编程、朴素贝叶斯校准和矛盾层的神经符号裁决器，显著提升MCC和精确率，同时保持规则可解释。
AI 中文摘要
在数字广告领域，检测被入侵的商业广告账户是一项挑战，因为攻击者利用被劫持的账户发起欺诈性广告活动。大型语言模型（LLM）智能体在完整性维护方面展现出潜力，但在困难案例上产生的幻觉性错误会造成业务摩擦。在一项研究中，我们发现自主智能体是一个强大的、以召回率为导向的信号提取器，但作为最终裁决者并不可靠，在模糊决策上会牺牲精确率。因此，我们将智能体保留为调查者，它输出一个结构化的、可解释的信号向量，并将裁决权交给神经符号阶段：由归纳逻辑编程（FOIL-IE）发现的符号规则、一个朴素贝叶斯校准层和一个数据调优的矛盾层。在过采样的人口中以及带有领域专家标注的现实低患病率样本上进行评估，这种裁决者替换将马修斯相关系数（MCC）从0.295提升至0.435（ΔMCC +0.139，95%置信区间[+0.026, +0.245]，p=0.018，配对自助法），精确率从0.250提升至0.446（1.8倍），召回率成本为从0.920降至0.660。在相同条件下进行基准测试，它还优于树集成方法（0.386）。这些规则编码了领域标签，同时保持可解释性和可审计性。
英文摘要
Detecting compromised business ad accounts is a challenge in digital advertising, as attackers exploit hijacked accounts to launch fraudulent campaigns. Large Language Model (LLM) agents show promise for integrity enforcement, but hallucinated mistakes on hard cases create business friction. In a study we find the autonomous agent is a strong, recall-heavy signal extractor but an unreliable final arbiter, conceding precision on ambiguous decisions. We therefore keep the agent as an investigator that emits a structured, interpretable signal vector, and delegate the verdict to a neuro-symbolic stage: symbolic rules discovered by Inductive Logic Programming (FOIL-IE), a Naïve Bayes calibration layer, and a data-tuned contradiction layer. Evaluating on a compromise-over-sampled population and a realistic low-prevalence sample with subject-matter-expert labels, this arbiter substitution raises MCC from 0.295 to 0.435 (ΔMCC +0.139, 95% CI [+0.026, +0.245], p=0.018, paired bootstrap), lifting precision from 0.250 to 0.446 (1.8x) at a recall cost (0.920 to 0.660). Benchmarked under identical conditions, it also edge tree ensembles (0.386).The rules encode domain w labels while remaininginterpretable and auditable.
CommentsAccepted at EMNLP 2026. 13 pages, 2 figures, 10 tables
