Behavioral Biometric Authentication Using Keystroke and Mouse Dynamics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional password-based authentication techniques are vulnerable to attacks, such as brute-forcing and data breaches, and require a large number of training sessions for accurate machine learning models, which increases security risks.
Innovation Solution
A biometrics-based user login authentication system that uses machine learning models trained with a small number of user login sessions to recognize legitimate users, incorporating mouse movement and keyboard stroke data, and applies a bias towards false negatives and tolerance to false positives, allowing for quick authentication with secondary verification measures.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional password-based authentication is used, then implementation is simple, but security is vulnerable to attacks
Solution Approach 1:
The patent combines traditional password-based authentication with behavioral biometrics (keyboard typing patterns and mouse movement patterns) into a unified authentication system. The machine learning model integrates multiple data sources including keystroke dynamics, mouse click patterns, and navigation behavior to create a composite authentication mechanism that leverages both password security and behavioral uniqueness, thereby enhancing overall authentication security while maintaining reasonable system complexity
Solution Approach 2:
The patent introduces a machine learning model as an intermediary component that analyzes behavioral patterns and generates authentication decisions. This intermediary layer processes behavioral biometric data and interacts with the traditional authentication system, enabling enhanced security through behavioral analysis without requiring complete system redesign, thus balancing security improvement with acceptable complexity
2Measurement precision
If machine learning models are trained with more sessions, then authentication accuracy is improved, but security risks increase due to prolonged exposure
Solution Approach 1:
The patent applies partial action by training the machine learning model with a limited number of login sessions (e.g., 3-7 sessions) rather than requiring extensive training data. This approach achieves sufficient authentication accuracy while minimizing the time window for security risks during training. The system accepts that perfect accuracy may not be achieved but obtains adequate performance with reduced exposure risk
Solution Approach 2:
The patent performs preliminary authentication using traditional password verification before applying behavioral biometric analysis. This preliminary action allows the system to quickly filter out obviously incorrect attempts and only apply the more resource-intensive behavioral analysis to plausible attempts, thereby reducing overall training requirements and security exposure while maintaining accuracy
3Reliability
If behavioral biometrics are incorporated, then authentication security is enhanced, but processing complexity increases
Solution Approach 1:
The patent segments the authentication process into distinct components: password verification, keyboard typing pattern analysis, and mouse movement pattern analysis. Each component processes specific aspects of user behavior independently, allowing the system to handle complexity in a modular fashion. The machine learning model receives segmented behavioral features from each component and integrates them for final authentication decisions, making the overall system more manageable
4Reliability
If fewer training sessions are used, then security risk is reduced, but model training speed must be increased
Solution Approach 1:
The patent changes key parameters of the machine learning model including feature selection (focusing on discriminative behavioral features like keystroke intervals and mouse movement vectors), model architecture (using efficient neural network structures), and training optimization (applying techniques like data augmentation and transfer learning). These parameter changes enable the model to achieve adequate accuracy with fewer training sessions, thereby reducing security risk while maintaining acceptable training speed
Data Source
AI summary
In one approach, a method includes: receiving a login event input from a user, the login event input being associated with a session of the user logging into an account; accessing a machine learning model; and authenticating, with the machine learning model, the user for the account, based at least in part on the login event input. In examples, the login event input comprises one or more items of biometric data associated with the user, an item of the one or more items of biometric data associated being generated by interaction of the user with an input device for logging into the account, and the interaction communicating a login credential of the user. In examples, an item of the one or more items of biometric data associated with the user is keyboard event-related biometric data, or mouse event-related biometric data.