Lightning Talk 2: Using Natural Language Processing Techniques to Crack Passwords

No ratings

Presented at SecuriTayDundee 2019 by

Robin Vickery

A custom dictionary that exploits the shared social experience of a userbase can be interactively built by making multiple cracking passes through a hash dump, and on each pass adding other similar words to the dictionary. We might crack one user's password that is based on a local football team and another based on an anime character but if we can add all the other regional football teams and other anime characters to the dictionary for the next cracking pass, we are likely to discover that other users share similar interests. Here we explore the use of Natural Language Processing models for automatically discovering candidate words for a custom password cracking dictionary.