Stochastic Image Token Authentication Security
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing network access token systems face challenges in robust authentication, particularly against organized or repeated attacks, due to reliance on static methods.
Innovation Solution
The use of model-generated images, specifically stochastic images produced by artificial intelligence models, for secure token registration, access, and authentication, without requiring users to remember authentication credentials.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If static authentication methods are used, then the system is simple to implement, but the security robustness against organized or repeated attacks deteriorates
Solution Approach 1:
The patent transforms static authentication methods into dynamic ones by generating unique authentication images stochastically for each authentication event. These images are generated using machine learning models that create visually distinctive representations based on token data, making each authentication challenge unique and unpredictable, thereby enhancing security robustness while maintaining implementation feasibility
Solution Approach 2:
The system changes the parameter of authentication from static credentials to dynamically generated visual images. By varying the visual parameters of authentication images stochastically for each token and authentication event, the system creates a more robust security mechanism that adapts to different authentication contexts while remaining computationally manageable
2Measurement precision
If user images are stored for authentication, then authentication accuracy improves, but data storage requirements and privacy concerns increase
Solution Approach 1:
Instead of storing actual user images, the system generates stochastic visual copies or representations of token data using machine learning models. These generated images serve as authentication credentials without requiring storage of personal user images, thereby maintaining authentication accuracy while minimizing data storage requirements and privacy risks
Solution Approach 2:
The patent introduces machine learning-generated images as an intermediary between the original token data and the authentication process. These intermediary images convey the necessary authentication information visually without requiring direct storage or transmission of sensitive user data, thus reducing storage needs and privacy concerns while preserving authentication accuracy
3Reliability
If stochastic machine learning models are used to generate images, then security against replication improves, but computational complexity increases
Solution Approach 1:
The system employs periodic regeneration of authentication images using stochastic machine learning models for each authentication event. This periodic generation ensures that even if computational complexity increases for individual image generation, the overall security is enhanced because each authentication requires a new, unpredictable image rather than reusing the same computational output
Solution Approach 2:
The machine learning models are trained in advance on relevant data to enable rapid generation of authentication images during actual use. This preliminary training action reduces the computational burden during authentication events, balancing the increased complexity of stochastic generation with efficient real-time performance
Data Source
AI summary
Methods and systems are described herein for generating deep-linked stochastic image representations of access tokens that embed token access deep links on a mobile application interface. The system may obtain, in connection with a request to register an access token with an account, token data associated with the access token and event data associated with one or more events performed with the access token. The system may generate, for input to a stochastic machine learning model, input vectors using the token data and the event data. The system may obtain, via the stochastic machine learning model based on the input vectors, an image for the access token and may generate, for display on a user interface associated with the account, an image representation of the access token including the image and a deep link to functionality associated with the access token.