Neural Network Password Generation Using Cognitive User Data
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional password generation methods fail to effectively incorporate user information, making generated passwords vulnerable to targeted attacks, as they do not adequately consider personal details that can be easily mined from social networks or other sources.
Innovation Solution
A deep learning-based approach that utilizes a neural network to encode and transform user information into secure passwords, leveraging sequence-to-sequence models and semantic transformations to generate passwords that are stronger and less predictable.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional dictionary-based password generation is used, then the generation process is simple and fast, but the generated passwords are vulnerable to targeted attacks and lack security
Solution Approach 1:
The patent replaces traditional mechanical dictionary-based password generation with a neural network-based deep learning system. The neural network encodes user information and generates passwords through learned transformations, substituting the mechanical lookup and mangling process with an intelligent system that adapts to user patterns while maintaining security against targeted attacks
Solution Approach 2:
The patent transforms the password generation approach by changing key parameters: instead of using fixed dictionaries and predefined mangling rules, the system dynamically adjusts password construction based on encoded user information features. The neural network learns optimal transformations of user data into secure passwords, adapting parameters like character selection, positioning, and transformation based on the input user information
2Ease of manufacture
If user personal information is directly encoded into passwords, then passwords are easier to generate, but they become more vulnerable to attacks from mined social network information
Solution Approach 1:
The patent introduces an intermediary neural network encoding layer between user information and password generation. Instead of directly using raw user information, the system encodes it through the neural network which learns meaningful representations and transformations. This intermediary process maintains the connection to user information for generation ease while protecting against direct attacks on mined social network data
Solution Approach 2:
The patent creates transformed copies of user information through neural network encoding rather than using the original data directly. The network learns to generate password-like structures that capture essential user information patterns while obscuring the direct connection to sensitive personal data, effectively copying the informational structure without exposing the original vulnerable data
Data Source
AI summary
An approach is provided that receives a set of user information pertaining to a user. The received set of information is encoded into a neural network and the neural network is trained using the encoded user information. As an output of the trained neural network, passwords corresponding to the user are generated.