Evaluation of Recurrent Neural Networks for Detecting Injections in API Requests

摘要

Application programming interfaces (APIs) are a vital part of every online business. APIs are responsible for transferring data across systems within a company or to the users through the web or mobile applications. Security is a concern for any public-facing application. The objective of this study is to analyze incoming requests to a target API and flag any malicious activity. This paper proposes a solution using sequence models to identify whether or not an API request has SQL, XML, JSON, and other types of malicious injections. We also propose a novel heuristic procedure that minimizes the number of false positives. False positives are the valid API requests that are misclassified as malicious by the model.