一句话：本文探究神经Bug检测器在变量误用Bug的检测与定位上是否达到软件开发者的水平。
调试是软件开发的核心活动，开发者常需判断代码片段是否含Bug并定位Bug。近年来，数据驱动方法被用于学习Bug检测任务，产生了神经Bug检测器；这类检测器在数百万含Bug与正确代码片段上训练。本文针对变量误用Bug，提出并考察一个关键问题：神经Bug检测器在此类Bug上的表现是否可与软件开发者相媲美。根据标题，论文将神经Bug检测器与软件开发者在变量误用Bug任务上进行比较，以评估二者在Bug识别与定位能力上的差距或可比性。给出的摘要仅描述背景与训练方式，未报告具体实验设置、参与者规模及定量结果。总体而言，该工作属于神经程序分析与人类调试能力对比的交叉研究。
Debugging, that is, identifying and fixing bugs in software, is a central part of software development. Developers are therefore often confronted with the task of deciding whether a given code snippet contains a bug, and if yes, where. Recently, data-driven methods have been employed to learn this task of bug detection, resulting (amongst others) in so called neural bug detectors. Neural bug detectors are trained on millions of buggy and correct code snippets.
@inproceedings{paperbot7274,
title = {Are Neural Bug Detectors Comparable to Software Developers on Variable Misuse Bugs?},
author = {Cedric Richter and Jan Haltermann and Marie-Christine Jakobs and Felix Pauck and Stefan Schott and Heike Wehrheim},
booktitle = {ASE},
year = {2022},
doi = {10.1145/3551349.3561156}
}