A classification-by-retrieval framework for few-shot anomaly detection to detect API injection

Udi Aharon, Ran Dubin, Amit Dvir, Chen Hajaj

Abstract: Highlights•A novel unsupervised few-shot anomaly detection framework to detect API Injection attacks.•Introduced a tokenizer designed to capture and emphasize language factors specific to APIs, addressing unique NLP challenges.•Leveraged FastText embedding combined with Approximate Nearest Neighbor search, employing a Classification-by-retrieval approach.•Validated the framework using public HTTP datasets against cutting-edge techniques.