Dynamic Authentication Trust Scoring to Reduce False Positives
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional authentication systems are inefficient and prone to false positives, leading to increased transactional costs and decreased user satisfaction due to fraudulent authentication action requests.
Innovation Solution
A method for dynamic trust score determination using machine learning models to evaluate authentication action requests, incorporating device trust scores and authentication action request metadata to generate trust scores for successful authentication responses, and updating device trust scores based on historical data and metadata.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If conventional authentication systems are used to evaluate authentication requests, then security checks are performed, but false positives occur and user satisfaction decreases
Solution Approach 1:
The patent changes the parameter of authentication evaluation from static rule-based checks to dynamic trust score calculations. The trust score is computed based on multiple parameters including device characteristics, user behavior patterns, request metadata, and historical data. This dynamic parameter adjustment allows the system to differentiate between legitimate and fraudulent requests more accurately, reducing false positives while maintaining security.
Solution Approach 2:
The patent replaces the mechanical system of conventional authentication (fixed rules and thresholds) with an intelligent system using machine learning models. The action trust score model and device trust score model automatically learn from historical data and adapt to new authentication patterns, substituting rigid mechanical evaluation with flexible intelligent assessment that improves both accuracy and user experience.
2Reliability
If conventional authentication systems perform security evaluations, then fraudulent requests are detected, but transactional costs increase
Solution Approach 1:
The patent performs preliminary trust score calculations during the authentication request processing. By pre-computing device trust scores based on historical data and device characteristics, and pre-evaluating request metadata against established patterns, the system prepares authentication assessments in advance. This preliminary action reduces the computational burden during actual authentication transactions, lowering processing costs while maintaining high fraud detection accuracy.
Solution Approach 2:
The machine learning models continuously self-improve by learning from authentication outcomes. The system automatically updates its understanding of fraudulent patterns and legitimate behavior without requiring manual intervention for each transaction. This self-service capability reduces the need for expensive manual security reviews and optimizes authentication decisions automatically, decreasing transactional costs while improving detection accuracy.
Data Source
AI summary
Methods, apparatuses, and computer program products are provided for dynamically determining a trust score for an authentication action request. An example method includes receiving an authentication action request from a user device. The method further includes determining a device trust score associated with the user device and generating an action trust score for the authentication action request based at least in part on the device trust score. The method further includes providing an authentication action response to the user device based at least in part on the trust score for the authentication request. The authentication action request metadata may include one or more of event data, user device information, location data, user biometric information, user device interaction information.