Confidence-Based Authentication System Using Weighted Scoring
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Traditional authentication methods often result in either unauthorized access when credentials are compromised or legitimate users are locked out due to forgotten passwords, as they rely on binary verification rather than assessing user confidence.
Innovation Solution
Implementing a confidence-based authentication system that presents multiple questions with varying point values, incorporating face recognition, voice recognition, device identity, and behavioral data to generate a confidence score, which determines access to secured resources.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If traditional binary authentication (username/password) is used, then the system is simple to operate, but security is compromised when credentials are stolen or forgotten
Solution Approach 1:
The patent transforms binary authentication (yes/no) into a continuous confidence score (0-100%), allowing for nuanced authentication decisions. Multiple authentication factors are weighted and combined to produce a confidence percentage, enabling the system to distinguish between highly confident and less confident authentications rather than treating all as equal binary outcomes.
Solution Approach 2:
The authentication system is divided into multiple independent authentication factors (knowledge-based questions, device identity, behavioral data, biometric data) that can be evaluated separately and then combined. Each factor contributes a weighted confidence score, allowing the system to assess different aspects of user identity independently before making an overall authentication decision.
2Reliability
If multiple authentication factors with confidence scoring are implemented, then security is improved, but the authentication process becomes more complex
Solution Approach 1:
The system requires only enough authentication factors to reach a configurable confidence threshold (e.g., 70%), not all possible factors. Users can be authenticated with a subset of available factors depending on their confidence contribution, avoiding the burden of requiring every single authentication method while still achieving sufficient security through the confidence threshold mechanism.
Solution Approach 2:
The authentication requirements are dynamic rather than static. The system adjusts which authentication factors are requested based on the current confidence score and the threshold needed for the specific resource being accessed. If a user already has high confidence from device identity, the system may not require additional factors, adapting the authentication process to the specific context.
3Object-affected harmful factors
If confidence thresholds are set high to prevent unauthorized access, then security against malicious users is improved, but legitimate users may be locked out due to incomplete authentication data
Solution Approach 1:
The system performs preliminary authentication with available factors before determining the final confidence score. Users are authenticated incrementally as factors are evaluated, allowing the system to build confidence progressively rather than requiring all factors upfront. This preliminary authentication approach ensures legitimate users aren't blocked even if some factors are unavailable.
Solution Approach 2:
The system provides feedback through confidence scores that indicate how close an authentication attempt is to meeting the threshold. This feedback mechanism allows the system to adjust authentication requirements dynamically and inform users of their authentication status, enabling corrective actions if needed while preventing both unauthorized access and false rejections of legitimate users.
Data Source
AI summary
Disclosed are various embodiments that perform confidence-based authentication of a user. An identification of a user account is obtained from a user, and a minimum confidence threshold is determined. Multiple authentication challenges are presented to the user. Responses are obtained from the user to one or more of the challenges, with each response having a corresponding authentication point value. A confidence score is generated for the user, where the confidence score is increased by the respective authentication point values of the correct responses. The user is authenticated as being associated with the user account in response to determining that the confidence score meets the minimum confidence threshold.