一句话：提出 Trident，利用基于 SQL 抽象语法树的神经网络建模上下文，以检测 SQL 注入攻击。
SQL 注入攻击数十年来持续威胁 Web 应用。攻击者将恶意代码混淆进看似自然的 SQL 语句中以窃取敏感数据，使其难以被检测。通常，SQL 语句的上下文信息可用于识别恶意信号，但这种上下文并不总是容易捕获。由于 SQL 作为形式语言具有高度结构化的特点，空间中相距较远的两个 token 在语义上可能非常接近，因此有效方法应在建模 SQL 语句上下文信息时考虑其结构特征。为此，论文提出 Trident，采用基于抽象语法树（AST）的神经网络来检测 SQL 注入攻击，通过引入 SQL 语句的结构信息提升对恶意信号的上下文建模能力。摘要未给出具体数据集、指标或对比结果。
SQL injection attacks have posed a significant threat to web applications for decades. They obfuscate malicious codes into natural SQL statements so as to steal sensitive data, making them difficult to detect. Generally, malicious signals can be identified by using the contextual information of SQL statements. Such contextual information, however, is not always easily captured. Due to the fact that SQL as a formal language is highly structured, two tokens that are spatially far away may be semantically very close. An effective approach thus should take the structural feature of SQL statements into account when modeling their contextual information.
@inproceedings{paperbot8333,
title = {Trident: Detecting SQL Injection Attacks via Abstract Syntax Tree-based Neural Network},
author = {Yuanlin Li and Zhiwei Xu and Min Zhou and Hai Wan and Xibin Zhao},
booktitle = {ASE},
year = {2024},
doi = {10.1145/3691620.3695289}
}