KG-AST 如何用预算化测试选择与 AI 审查降低 CI/CD 成本
KG-AST for Cheaper CI/CD: Budgeted Test Selection and AI Bug Review
AI 导读
KG-AST 基于提交历史和代码图特征为每个 commit 打风险分，将 CI 测试和 AI 审查预算优先分配给高风险提交。在 11 种语言的 155 个开源项目上，对风险最高的 20% 提交测试可覆盖 49-64% 的引入 bug 的提交，优于随机选择的 20% 和最大变更优先的 40-55%；在 20 个 Java 项目中以 15.1 个百分点优势胜出 19 个。
正文 · 原文
Abstract:Continuous integration and AI-assisted review spend the same budget on every commit, although few commits introduce bugs. We study KG-AST, which scores each commit risk from git history and code-graph features before heavy tests or AI review run, so the budget goes to the riskiest commits first, and which hands the AI reviewer only a graph slice around the change. On 155 open-source projects in 11 languages, a single cross-language model trained on other projects ranks commits so that heavy testing of the riskiest 20% covers 49-64% of bug-introducing commits depending on the language, against 20% for random selection and 40-55% for testing the biggest change first.
Every language beats the biggest-change baseline, by 3.8 to 13.2 points. In 20 Java projects the gain at a 20% budget is 15.1 points (64.6% vs 49.5%, 95% interval +11.6 to +18.4), and the model wins in 19 of the 20 projects. About twelve training projects suffice; richer features, tuning and cleaner labels did not help. An AI reviewer that sees only a commit's diff finds 26% of later-fixed defects exactly (32% with graph context), and about half of its findings that did not match a later fix were judged real by independent AI judges. A static call-graph selector runs about 8% of a test suite while including a test written for the fix in 85% of 67 fix commits. The Poincare embedding used in our retrieval work adds nothing measurable here: commit-risk AUC, bug review and severity ranking are unchanged or worse. The labels are noisy and the budget metric assumes a heavy test catches the bug it covers; no CI minutes were measured.
| Subjects: | Software Engineering (cs.SE); Artificial Intelligence (cs.AI) |
| Cite as: | arXiv:2611.06950 [cs.SE] |
| (or arXiv:2611.06950v1 [cs.SE] for this version) | |
| https://doi.org/10.48550/arXiv.2611.06950 arXiv-issued DOI via DataCite |
Submission history
From: Henry Yeung [view email]
[v1]
Tue, 3 Nov 2026 16:28:29 UTC (116 KB)
来源：arXiv cs.AI · arxiv.org