arXivDaily arXiv每日学术速递 周一至周五更新
arXiv周末暂无论文更新,休息一下吧,周末愉快~~
arXiv 2608.21203cs.AI

SENTRY:面向IT变更管理的确定性智能风险评估

SENTRY: Deterministic, Intelligent Risk Assessment for IT Change Management

  • Royal Bank of Canada(加拿大皇家银行)
  • RBC Borealis

机构由 AI 辅助整理,请以论文原文为准。

Daniel Arulpragasam, Christer Henrysson, Ella Ly, Deepika Anbalagan, Leo Feng

AI总结:

SENTRY是一款替代问卷评分的IT变更管理风险评估平台,采用XGBoost与RAG构建的确定性流水线,在企业数据上AUC达0.87,高风险变更检测速率为现有流程的3.25倍。

AI中文摘要:

大型金融机构的技术变更管理依赖于准确、一致且可审计的风险评估。实际中,许多机构仍依赖自填式问卷,这些问卷具有主观性、易被操纵,且难以区分常规变更与后续触发重大事件的变更。本文提出SENTRY,这是一款风险评估平台,它用由梯度提升决策树(XGBoost)和混合检索增强生成(RAG)构建的确定性机器学习流水线取代基于问卷的评分。该系统结合结构化运营元数据、应用依赖图、历史事件记录,以及对历史变更请求的混合语义与词汇搜索;检索步骤从非结构化变更请求文本中捕获风险信号,再将该信号压缩为单个标量特征供模型推理。此设计使模型保持确定性,并通过SHAP值保留每一次预测的可解释性。在企业级变更数据上评估,SENTRY的ROC AUC达0.87,整体准确率为85%,检测高风险变更的速率约为现有流程的3.25倍。最后,本文探讨该设计背后的架构权衡及其对受监管变更管理中机器学习应用的启示。

英文摘要:

Technology change management in large financial institutions depends on risk assessments that are accurate, consistent, and auditable. In practice, many institutions still rely on self-reported questionnaires. Those questionnaires are subjective, easy to game, and poor at separating routine changes from the ones that later trigger major incidents. This paper presents SENTRY, a risk assessment platform that replaces questionnaire-based scoring with a deterministic machine learning pipeline built from gradient-boosted decision trees (XGBoost) and hybrid retrieval-augmented generation (RAG). The system combines structured operational metadata, application dependency graphs, and historical incident records with a hybrid semantic and lexical search over historical change requests. The retrieval step captures the risk signal in unstructured change request text, then compresses that signal into a single scalar feature before model inference. That design keeps the model deterministic and preserves per-prediction explainability via SHAP values. Evaluated on enterprise-scale change data, SENTRY achieves a ROC AUC of 0.87 and 85% overall accuracy, and it detects high-risk changes at roughly 3.25 times the rate of the existing process. We close by examining the architectural trade-offs behind this design and what they imply for the use of machine learning in regulated change management.

补充信息

↑