---
title: "SENTRY：面向IT变更管理的确定性智能风险评估系统"
source: "arXiv AI/CL/LG（论文）"
source_tier: "t2"
url: "https://arxiv.org/abs/2608.21203v1"
permalink: "https://feed.astralor.com/items/itm_bdac0884ed712696501fa5d0"
published: "2026-08-21T15:18:21.000Z"
score: 72
channel: "tech"
tags: ["arXiv","IT变更管理","SENTRY","XGBoost","可解释AI","检索增强生成"]
---

# SENTRY：面向IT变更管理的确定性智能风险评估系统

## AI 点评

> 这篇论文值得关注，因为它把RAG和XGBoost以一种克制的、可审计的方式嵌入受监管流程，而不是堆砌大模型。对金融科技、IT运维负责人和风控团队尤其有参考价值：它展示了一个不过度复杂、又能显著提升风险检出率的落地范例。

## 速览

大型金融机构的IT变更管理依赖风险评分，但传统问卷式评估主观、易钻空子，难以区分常规变更与高风险变更。本文提出SENTRY平台，用XGBoost梯度提升决策树与混合检索增强生成（RAG）替代问卷打分，融合结构化运营元数据、应用依赖图和历史事件记录，并通过混合语义与词法检索将非结构化变更文本压缩为单一标量特征，在保持模型确定性的同时用SHAP值实现可解释性。在企业级数据上，SENTRY的ROC AUC达0.87，准确率85%，高风险变更检出率约为现有流程的3.25倍。

## 关键点

- SENTRY用XGBoost与混合RAG替代问卷式风险评分，实现确定性ML管道
- 通过混合检索将非结构化变更文本压缩为标量特征，保持模型可解释性
- 企业级数据上ROC AUC 0.87、准确率85%，高风险检出率达现有流程3.25倍
- 探讨了受监管变更管理中使用机器学习的架构权衡

## 全文

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.

---

引用前请核对原文;本文由星际流动整理:https://feed.astralor.com/items/itm_bdac0884ed712696501fa5d0