Targeted Code Inspection based on Human Errors

人类与检查（审计）软件对人类数据表错误的检测

The Detection of Human Spreadsheet Errors by Humans versus Inspection (Auditing) Software

以前的电子表格检查实验让人类受试者在电子表格中寻找种子错误。在这项研究中，受试者试图在人类开发的电子表格中找到错误，以避免错误播种造成的潜在伪影。将人类受试者的成功率与应用于相同电子表格的电子表格静态分析工具（SSAT）的错误标记成功率进行比较。人为错误检测结果与使用错误播种的研究结果相当。然而，在这项研究中，Excel错误检查和电子表格专业版几乎无法正确标记自然（人为）错误。

代码审查的终结：编码代理取代人工审查

The End of Code Review: Coding Agents Supersede Human Inspection

本文论证基于大型语言模型的编码代理已超越传统人工代码审查的能力阈值，能以更低成本、更高吞吐量实现审查目标，并指出人工审查与AI协作的路径不可持续。

How Readable is Model-generated Code? Examining Readability and Visual Inspection of GitHub Copilot

Background: Recent advancements in large language models have motivated the practical use of such models in code generation and program synthesis. However, little is known about the effects of such tools on code readability and visual attention in practice. Objective: In this paper, we focus on GitHub Copilot to address the issues of readability and visual inspection of model generated code. Readability and low complexity are vital aspects of good source code, and visual inspection of generated code is important in light of automation bias. Method: Through a human experiment (n=21) we compare …

需求工程中的人机协作：大语言模型（LLM）对需求检查的负面影响证据

Human-AI Collaboration in Requirements Engineering: Evidence of the Negative Effect of LLMs on Requirements Inspection

该研究通过控制交叉实验发现，大语言模型（LLM）支持会降低新手需求检查员的气味检测准确率，且先使用LLM学习需求检查会减缓技能获取。

基于溯源跟踪的Human-in-the-Loop合成文本数据审查

Human-in-the-Loop Synthetic Text Data Inspection with Provenance Tracking

为剔除合成数据中的错误标签，本文开发了结合溯源跟踪与辅助标注的human-in-the-loop数据审查技术INSPECTOR，显著提升了人工识别正确标签文本的效率，验证了组合技术减少人工审查的必要性。

通过检索增强生成提升大型语言模型用于软件测试和检查自动化

Enhancing Large Language Models with Retrieval Augmented Generation for Software Testing and Inspection Automation

本文通过检索增强生成方法提升大型语言模型在软件测试和检查自动化中的应用，有效减少项目成本并提高验证与确认活动的效果和效率。