paperbot · PL 论文追踪

全部论文 待读 收藏

P1 Learning probabilistic models for static analysis alarms

学习静态分析告警的概率模型
ICSE Proceedings of the 44th International Conference on Software Engineering2022匹配度 7/10 · 论文研究静态分析告警的概率模型学习，属于你明确感兴趣的贝叶斯/概率程序分析方向，且问题驱动、兼顾模型学习与工具评估。但它不涉及 LLM/Agent，与你当前主攻的 AI→PL（用 LLM 改进静态分析）及贝叶斯程序分析与 LLM 的交叉仍有距离。引用 20 inbox
Hyunsu Kim, Mukund Raghothaman, Kihong Heo
程序分析与抽象解释机器学习与程序语言
crossref: 2026-09-12dblp: 2026-09-12

AI 速览

一句话：提出 BayeSmith，自动学习静态分析告警概率模型的结构与权重，以更好排序真实缺陷并减少误报。

现有静态分析告警的概率推理技术虽能结合外部反馈降低人工检查负担，但其模型结构预定义，难以跨程序学习或迁移知识，且常对外部反馈过度泛化，导致真实缺陷被误抑制。BayeSmith 从初始模型和带缺陷标签的训练程序出发，自动学习概率模型的结构与权重，并据此优化对真实缺陷的优先级排序。作者在 C 程序集上使用两种静态分析进行评估，证明所学模型显著提升三种先进概率推理系统的性能。

问题
现有概率推理方法受限于预定义模型结构，无法跨程序学习或迁移知识，且可能过度泛化外部反馈而误抑制真实缺陷。
方法
BayeSmith 从初始模型与带缺陷标签的训练程序自动学习概率模型的结构与权重，并细化模型以优先排序真实缺陷。
评估
在 C 程序上结合两种静态分析进行评估，所学模型显著提升三种先进概率推理系统的性能。
原文摘要（Abstract）

We present BayeSmith, a general framework for automatically learning probabilistic models of static analysis alarms. Several probabilistic reasoning techniques have recently been proposed which incorporate external feedback on semantic facts and thereby reduce the user's alarm inspection burden. However, these approaches are fundamentally limited to models with pre-defined structure, and are therefore unable to learn or transfer knowledge regarding an analysis from one program to another. Furthermore, these probabilistic models often aggressively generalize from external feedback and falsely suppress real bugs. To address these problems, we propose BayeSmith that learns the structure and weights of the probabilistic model. Starting from an initial model and a set of training programs with bug labels, BayeSmith refines the model to effectively prioritize real bugs based on feedback. We evaluate the approach with two static analyses on a suite of C programs. We demonstrate that the learned models significantly improve the performance of three state-of-the-art probabilistic reasoning systems.

链接与引用

DOI 原文 · DBLP

BibTeX 
@inproceedings{KimRH22,
  title = {Learning probabilistic models for static analysis alarms},
  author = {Hyunsu Kim and Mukund Raghothaman and Kihong Heo},
  booktitle = {ICSE},
  year = {2022},
  doi = {10.1145/3510003.3510098}
}