SBE: Small: Continuous Human-User Authentication by Induced Procedural Visual-Motor Biometrics

验证用户身份是网络空间的基本安全要求之一。当前的身份验证方法要求人们创建和记住诸如复杂密码之类的秘密凭证，或者拥有特殊的硬件身份验证令牌。两者都很容易被泄露或非法共享。更糟糕的是，通常只在会话开始时支持身份验证。生物特征认证的最新发展旨在克服这些挑战，但要么容易绕过（例如，冒名顶替者攻击），不能令人满意的假阳性-假阴性权衡，或者需要昂贵的专用设备。为了解决这些弱点，本项目的目标是开发一种非传统的生物特征认证方法，该方法利用运动学习和视觉-运动适应来提供安全，有效和连续的人类用户认证。本项目专注于开发网络安全的新范式-基于专业知识的连续生物特征认证。新范式的独特之处在于通过将特定的专业知识引入人类主体与机器的交互中，积极训练主体以实现持续的认证。专业知识被定义为一种人工诱导的交互模式，它是人类主体所独有的。一旦受试者暴露于特定的刺激并学习了他们的专业知识，只要刺激保持不变，他们的行为就很容易区分。为了增强系统的安全性，以及保护用户隐私（通过促进刺激停止后的专业知识淡出），诱导的专业知识不是静态的，而是在随机选择的目标之间不断发展。该项目开发基于键盘、鼠标和触摸屏设备的基于专业知识的主动生物识别认证模式。本研究成果将对我国高等教育、高科技产业和政府产生更广泛的影响。这项研究加强了PI的课程开发，有助于吸引，招聘和留住工程专业的学生，并扩大代表性不足的少数民族和妇女在工程项目中的参与。它还支持PI向有才华的大学和高中学生推广，并激励他们学习并最终攻读计算机工程领域的博士学位。

Validating a user's identity is one of the fundamental security requirements in cyberspace. Current authentication approaches require people to create and remember secret credentials such as complex passwords, or to possess special hardware authentication tokens. Both are vulnerable to being compromised, or illegally shared. Even worse, authentication is typically supported solely at the start of a session. Recent developments in biometric authentication aim to overcome these challenges, but are either easy to bypass (e.g., imposter attacks), incapable of satisfactory false-positive-false-negative tradeoffs, or require expensive dedicated equipment. To address these weaknesses, the objective of this project is to develop a non-traditional approach to biometric authentication, which exploits motor learning and visual-motor adaptation to provide secure, effective and continuous human-user authentication.This project focuses on developing a new paradigm in cyber security - expertise-based continuous biometric authentication. The uniqueness of the new paradigm is to actively train the subject to achieve continuous authentication, by inducing a specific expertise into a human subject's interaction with a machine. An expertise is defined as an artificially-induced interaction pattern that is unique to the human subject. Once the subjects have been exposed to specific stimuli and learned their expertise, their behaviors are easily differentiable as long as the stimuli are maintained. To enhance the security of the system, as well as to preserve user privacy (by facilitating expertise fade-out after the cessation of the stimuli), the induced expertise are not static, but rather continuously evolving between randomly-chosen targets. The project develops expertise-based active biometric authentication modalities based on keyboard, mouse, and touchscreen devices. The outcome of this research shall have broader impacts on the nation's higher education, high-tech industries, and government. The research enhances the PIs' curriculum developments and helps attract, recruit and retain engineering students and broaden the participation of under-represented minorities and women in engineering programs. It also supports the PIs to outreach to talented college and high-school students and motivate them to study and eventually pursue Doctoral degrees in computer engineering areas.