Dynamic Access Control via Behavioral ML Classification
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing access control methods for user devices are inadequate as they provide persistent access once gained, limiting recourse to security mechanisms and failing to effectively manage non-compliant user behavior in real-time.
Innovation Solution
A computer-implemented method using a machine learning classifier to determine user behavior, classifying it as compliant or non-compliant, and triggering credential-based authentication when non-compliant behavior exceeds a threshold, with the classifier being constructively trained on subsequent behaviors to improve authentication accuracy.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If persistent access is provided once gained, then ease of operation is improved, but security reliability deteriorates
Solution Approach 1:
The access control system transitions from static persistent access to dynamic continuous authentication. The system continuously monitors user behavior and adjusts access rights in real-time based on detected behavior patterns, allowing access to be revoked or modified without requiring user intervention while maintaining convenience during authorized usage periods.
Solution Approach 2:
The system implements continuous feedback loops where user behavior is constantly monitored and evaluated against established profiles. Authentication decisions are based on real-time feedback from behavior analysis, allowing the system to adapt access control decisions dynamically based on current user actions and contextual information.
2Reliability
If continuous behavior monitoring is implemented, then access control reliability is improved, but device complexity increases
Solution Approach 1:
The behavior monitoring system leverages existing multi-functional sensors and components already present in the device for various purposes such as location tracking, motion detection, and user interaction monitoring. These same components serve dual functions: their primary purposes plus continuous behavior analysis for access control, thereby avoiding additional hardware complexity.
Solution Approach 2:
The system utilizes the device's own existing sensors, processors, and software frameworks to perform behavior monitoring and authentication functions. Rather than requiring external monitoring systems or additional dedicated hardware, the device monitors and evaluates its own usage patterns using its inherent capabilities.
3Reliability
If credential-based authentication is required for non-compliant behavior, then access control reliability is improved, but ease of operation deteriorates
Solution Approach 1:
The authentication requirements dynamically adjust based on detected behavior compliance. During periods of compliant behavior, users enjoy uninterrupted access without frequent re-authentication. When non-compliant behavior is detected, the system temporarily imposes additional credential verification, then returns to permissive access once compliance is restored, creating a flexible adaptive authentication policy.
Solution Approach 2:
The system implements periodic behavior evaluation cycles where user actions are continuously assessed against compliance criteria. Authentication intensity varies periodically based on accumulated behavior scores, applying stricter controls only when necessary and maintaining ease of operation during normal compliant usage periods.
Data Source
Figure 1~2
Figure 3
Figure 4~5
AI summary
A computer implemented method of access control for a user device having at least one component for determining behaviours of the user, the method comprising: accessing a machine learning classifier trained based on at least one prior behaviour of the user using the device, the classifier classifying user behaviour as compliant or non-compliant such that compliant behaviour is determined by the classifier to be consistent with prior behaviour for permitting access to the device; initialising a measure of a period since a most recent compliant behaviour occurred; and responsive to a determination that a subsequent behaviour is classified as non-compliant and that the period meets a threshold, 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 re-initialising the measure of the period and constructively training the classifier based on the subsequent behaviour as a compliant behaviour by providing the subsequent behaviour as an additional training example.