ManyVuls4J
收藏资源简介:
ManyVuls4J数据集是由国防科技大学复杂系统软件工程重点实验室创建,旨在评估自动化漏洞修复工具的性能。该数据集包含103个真实世界中的漏洞实例,比前一版本增加了30%。数据集的创建过程包括从ProjectKB中收集漏洞信息,并进行手动验证以确保可重复性。ManyVuls4J主要用于评估自动化漏洞修复工具在不同类型漏洞上的修复效果,旨在提高软件安全性,减少手动修复的负担。
The ManyVuls4J dataset was created by the Key Laboratory of Software Engineering for Complex Systems, National University of Defense Technology, aiming to evaluate the performance of automated vulnerability repair tools. It contains 103 real-world vulnerability instances, marking a 30% increase compared to the previous version. The dataset construction process involves collecting vulnerability information from ProjectKB and conducting manual validation to ensure reproducibility. Primarily, ManyVuls4J is used to assess the repair efficacy of automated vulnerability repair tools across diverse vulnerability types, with the goal of enhancing software security and alleviating the workload of manual vulnerability repair.

- 1There are More Fish in the Sea: Automated Vulnerability Repair via Binary Templates复杂系统软件工程重点实验室, 国防科技大学计算机科学与技术学院, 长沙, 中国 · 2024年



