Knowledge-Based Authentication Using Synthetic Photos
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing knowledge-based authentication methods lack robustness in providing additional security layers, especially when users access sensitive information or resources from unknown devices or locations, and they do not effectively leverage user-specific mobile-device photos and assets for authentication.
Innovation Solution
A system that employs a machine learning model to identify and select mobile-device photos and assets, generating challenges that include synthetic photos consistent with user-specific images, thereby enhancing the accuracy of knowledge-based authentication through user recognition.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional knowledge-based authentication is used, then the authentication process is simple, but the security level is insufficient when accessing sensitive information from unknown devices
Solution Approach 1:
The system performs preliminary actions by collecting and analyzing user interaction data, device information, and authentication patterns before the authentication challenge. Machine learning models are pre-trained on this historical data to establish user profiles and behavioral baselines, enabling more accurate and secure authentication decisions when the actual authentication event occurs.
Solution Approach 2:
The system implements feedback mechanisms by continuously monitoring authentication outcomes and user responses, feeding this information back into the machine learning models to improve future authentication decisions. The system learns from successful and failed authentication attempts, adapting its challenge generation and verification processes to enhance security over time.
2Measurement precision
If user-specific mobile-device photos and assets are leveraged, then the memorability and recency of authentication challenges improve, but the complexity of selecting and verifying these resources increases
Solution Approach 1:
The system enables self-service by allowing users to pre-select and organize their own photos and assets as authentication resources. Users can curate their own photo albums and asset collections that they are comfortable using for authentication, reducing the system's burden of managing and verifying these resources while maintaining high authentication accuracy.
Solution Approach 2:
The system uses copying by creating synthetic or representative versions of user photos and assets. Instead of requiring direct access to and verification of original user files, the system generates copies or metadata representations that can be efficiently processed and verified, reducing complexity while preserving authentication accuracy.
3Productivity
If machine learning models are employed to identify authentication resources, then the efficacy of authentication challenges improves, but the computational resources and processing time increase
Solution Approach 1:
The system applies segmentation by dividing the machine learning processing into distinct stages: offline training phase where complex model development occurs, and online inference phase where lightweight prediction is performed during authentication. This segmentation allows heavy computational work to be done in advance when energy consumption is less critical, while maintaining fast, low-energy operation during actual authentication events.
Data Source
AI summary
Systems, methods, and computer program products disclosed herein relate to knowledge-based authentication leveraging mobile-device photos and assets. In one embodiment, the system can identify, by employing a machine learning model, a plurality of authentication resources associated with a user, wherein the machine learning model is trained using historical information efficacy of authentication challenges. In another embodiment, the system can select a mobile-device photo and a mobile-device asset associated with the user from the plurality of authentication resources. In another embodiment, the system can select a synthetic photo consistent with the mobile-device photo. In another embodiment, the system can generate a challenge that includes the mobile-device photo, the mobile-device asset and the synthetic photo. In another embodiment, the system can authenticate with knowledge-based authentication based upon accuracy of a reply received in response to the challenge.