SePass: Semantic Password Guessing Using k-nn Similarity Search in Word Embeddings

Hünemörder, Maximilian ; Schäfer, Levin; Schüler, Nadine-Sarah; Eichberg, Michael und Kröger, Peer (2022): SePass: Semantic Password Guessing Using k-nn Similarity Search in Word Embeddings. 18th International Conference on Advanced Data Mining and Applications, ADMA 2022, Brisbane, QLD, Australia, November 28–30, 2022. In: Advanced Data Mining and Applications. 18th International Conference, ADMA 2022, Brisbane, QLD, Australia, November 28–30, 2022, Proceedings, Part II, Lecture Notes in Computer Science Bd. 13726 Cham: Springer. S. 28-42

Abstract

Password guessing describes the process of finding a password for a secured system. Use cases include password recovery, IT forensics and measuring password strength. Commonly used tools for password guessing work with passwords leaks and use these lists for candidate generation based on handcrafted or inferred rules. These methods are often limited in their capability of producing entirely novel passwords, based on vocabulary not included in the given password lists. However, there are often semantic similarities between words and phrases of the given lists that are highly relevant for guessing the actual used passwords. In this paper, we propose SePass, a novel method that utilizes word embeddings to discover and exploit these semantic similarities. We compare SePass to a number of competitors and illustrate that our method not only is on par with these competitors, but also generates a significant higher amount of entirely novel password candidates. Using SePass in combination with existing methods, such as PCFG, improves the number of correctly guessed passwords considerably.

• Dokumententyp: Konferenzbeitrag (Paper)
• Fakultät: Mathematik, Informatik und Statistik > Informatik
• Themengebiete: 000 Informatik, Informationswissenschaft, allgemeine Werke > 000 Informatik, Wissen, Systeme
• ISSN: 0302-9743
• Ort: Cham
• Sprache: Englisch