Behavior-Based Access Control With Continuous User Verification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing access control systems for user devices, such as smartphones and tablets, provide persistent access once unlocked, lacking effective mechanisms to dynamically adapt to user behavior changes and ensuring secure access control.
Innovation Solution
Implement a machine learning classifier to analyze user behavior, classifying it as compliant or non-compliant, and utilize credential-based authentication to confirm authorization, with constructive training to improve the classifier's accuracy over time.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If persistent access is provided once unlocked, then ease of operation is improved, but security deteriorates
Solution Approach 1:
The access control system transitions from a static persistent access model to a dynamic continuous verification model. The machine learning classifier continuously monitors user behavior and dynamically adjusts access permissions in real-time, preventing unauthorized access while maintaining convenience for legitimate users.
Solution Approach 2:
The system implements feedback through continuous behavior monitoring and classification. The machine learning model receives ongoing feedback from sensor data about user behavior patterns and adjusts its classification of compliant versus non-compliant behavior, enabling adaptive security that responds to changing conditions without requiring frequent manual authentication.
2Reliability
If machine learning classifier is used to classify user behavior, then security is improved, but device complexity increases
Solution Approach 1:
The machine learning classifier serves multiple functions: it classifies user behavior for security verification, adapts to changing usage patterns over time, and works across different device contexts (locked versus unlocked states). This multi-functionality justifies the complexity by providing comprehensive security enhancement rather than a single specialized function.
Solution Approach 2:
The system employs self-service through automatic adaptive training of the machine learning classifier. The classifier continuously learns from new behavior data without requiring manual reconfiguration or intervention, automatically improving its security classification capabilities while adapting to the specific user's patterns, thereby managing complexity through automation rather than manual processes.
3Measurement precision
If constructive training is performed continuously, then measurement precision is improved, but loss of time occurs
Solution Approach 1:
The constructive training operates continuously in the background without interrupting device usage. The machine learning classifier learns from new behavior data as it accumulates, maintaining continuous improvement of classification accuracy while the device remains fully functional, thus avoiding time loss through seamless parallel operation.
Solution Approach 2:
The system performs preliminary classification using existing training data before new behavior data arrives, enabling immediate security decisions. Subsequent constructive training then refines the classification in preparation for future decisions, allowing the system to maintain high measurement precision without requiring time-consuming training cycles to complete before each classification event.
Data Source
AI summary
A computer implemented method of access control for a user device having at least one component for determining behaviors of the user. The method including accessing a machine learning classifier trained based on at least one prior behavior of the user using the device, the classifier classifying user behavior as compliant or non-compliant such that compliant behavior is determined by the classifier to be consistent with prior behavior for permitting access to the device The method further includes, in response to a determination that a subsequent behavior is classified as non-compliant, requesting a credential-based authentication of the user and permitting access to the device in response to the credential-based authentication, wherein permitting access to the device further includes constructively training the classifier based on the subsequent behavior as a compliant behavior by providing the subsequent behavior as an additional training example.