Scholarlink

In response to rising digital threats, we conduct a two-phased exploration of password security practices. Through a survey including 61 individuals of varying ages, we investigate password security practices and identify a gap between awareness and actual practices, particularly concerning the reuse of similar passwords and the lack of complexity. Next, we operate three machine learning (ML) algorithms on a dataset of 669,643 leaked passwords to determine the factors that contribute to password strength considered by the ML algorithms. Surprisingly, we find that ML algorithms achieve high accuracy in password strength classification by primarily focusing on password length only.

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

Password Security in Practice: An Appraisal Using Users’ Perception and Machine Learning

  • 会议论文

  • 发表日期：

    2025年

查看全文
  • Zackary Beck,
  • Adam Crooks,
  • Md. Fazle Rabbi,
  • Minhaz Zibran,
  • Farjana Eishita

摘要

In response to rising digital threats, we conduct a two-phased exploration of password security practices. Through a survey including 61 individuals of varying ages, we investigate password security practices and identify a gap between awareness and actual practices, particularly concerning the reuse of similar passwords and the lack of complexity. Next, we operate three machine learning (ML) algorithms on a dataset of 669,643 leaked passwords to determine the factors that contribute to password strength considered by the ML algorithms. Surprisingly, we find that ML algorithms achieve high accuracy in password strength classification by primarily focusing on password length only.