Recurrent Neural Network Authentication Key Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing user authentication systems are inefficient and vulnerable to adversarial attacks, as they rely on digital passwords and tokens that can be compromised, leading to security risks and productivity losses.
Innovation Solution
A computer-implemented method using a combination of user recurrent neural networks and system recurrent neural networks to generate unique combined keys, which are used for authentication, providing a physically unclonable function that is resistant to adversarial attacks and offers flexible multi-factor authentication.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of manufacture
If digital passwords and tokens are used for authentication, then implementation is simple and cost-effective, but security is vulnerable to adversarial attacks and key compromise
Solution Approach 1:
The patent replaces traditional digital authentication mechanisms (passwords, tokens) with a hardware-based neural network system. The neural network hardware generates cryptographic keys through its physical structure, substituting software-based authentication with hardware-based authentication that leverages the physical unclonability of neural network circuits.
Solution Approach 2:
The patent combines multiple authentication factors into a composite authentication system. It integrates hardware authentication (neural network keys), software authentication (passwords), and biometric authentication (fingerprint, facial recognition) to create a multi-factor authentication system that provides both simplicity and robust security.
2Reliability
If hardware authentication is implemented, then security is improved, but device complexity increases
Solution Approach 1:
The patent divides the authentication system into separate functional modules: a neural network hardware unit for generating cryptographic keys, a software authentication module for processing passwords, and a biometric authentication module for verifying physical characteristics. This segmentation allows each component to be optimized independently and simplifies the overall system architecture.
Solution Approach 2:
The patent creates a universal authentication system where a single neural network hardware unit can serve multiple authentication purposes. The same neural network can generate keys for different authentication methods (password-based, biometric-based) and can be used across multiple devices and systems, reducing overall complexity through reusability.
3Ease of operation
If passwords and tokens are used, then authentication is straightforward, but productivity is reduced due to inefficiency and user circumvention
Solution Approach 1:
The patent implements self-service authentication where the neural network hardware automatically generates and manages cryptographic keys without user intervention. The system performs authentication decisions autonomously based on the generated keys, eliminating the need for users to manually manage passwords or tokens, thereby improving both simplicity and productivity.
Data Source
AI summary
A computer-implemented method of user authentication is provided. The method comprises combining, by a computer system, a user recurrent neural network with a system recurrent neural network to form a unique combined recurrent neural network. The user recurrent neural network is configured to generate a unique user key, and the system recurrent neural network is configured to generate a system key. The computer system inputs a predetermined input into the combined recurrent neural network, and the combined recurrent neural network generates a unique combined key from the input, wherein the combined key differs from both the user key and system key. The computer system then associates the combined key with a unique access authorization to authenticate a user.