Multi-armed bandit approach to password guessing

TL;DR

This paper proposes a novel framework applying multi-armed bandit algorithms to optimize password guessing strategies by balancing exploration of different password sources and exploitation of the most promising ones.

Contribution

It introduces a new approach that models password guessing as a multi-armed bandit problem, enabling adaptive and potentially more effective guessing strategies.

Findings

Bandit-based approach improves guessing efficiency

Framework adapts to different password sources

Potential for higher success rates in password cracking

Abstract

The multi-armed bandit is a mathematical interpretation of the problem a gambler faces when confronted with a number of different machines (bandits). The gambler wants to explore different machines to discover which machine offers the best rewards, but simultaneously wants to exploit the most profitable machine. A password guesser is faced with a similar dilemma. They have lists of leaked password sets, dictionaries of words, and demographic information about the users, but they don't know which dictionary will reap the best rewards. In this paper we provide a framework for using the multi-armed bandit problem in the context of the password guesser and use some examples to show that it can be effective.