Behavioral Biometric Authentication via Segmented User Models

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing systems for behavioral biometrics in user authentication struggle to accurately verify identities when users switch between manually entering login information and using password managers, or when using different computing devices and physical positions, leading to inconsistencies in behavioral data analysis.

Innovation Solution

The technology monitors and analyzes behavioral biometric data across various devices and environments, using multiple entry methods and physical positions, by comparing received data to personalized user models and initiating secondary authentication methods when similarity thresholds are not met, and creating multiple user models to accommodate different scenarios.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If a single user model is used for behavioral biometric authentication, then the authentication process is simple, but the accuracy decreases when users switch between manual entry and password manager entry or use different devices

Engineering Contradiction:

Improveauthentication accuracyVSAvoiduser model structure

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent divides a single user model into multiple device-specific user models. Each device model captures behavioral patterns specific to that device, allowing accurate authentication whether the user enters credentials manually or via password manager. The system segments the behavioral biometric data by device type and maintains separate models for each, resolving the contradiction between simplicity and accuracy.

Inventive Principle:

Principle #1Segmentation

Solution Approach 2:

The patent implements dynamic user models that adapt to different entry methods (manual vs. password manager) and different physical positions. The system dynamically selects or creates appropriate user models based on the detected device and entry pattern, making the authentication system flexible and accurate across varying conditions without requiring a single complex static model.

Inventive Principle:

Principle #15Dynamics

2Adaptability or versatility

If behavioral biometric data is collected from multiple devices and positions, then the system adapts to user variations, but the data analysis consistency deteriorates

Engineering Contradiction:

Improveaccommodation of entry methodsVSAvoidbehavioral data consistency

Core Design Contradiction:

Adaptability or versatilityVSStability of the object's composition

Solution Approach 1:

The patent applies local quality by creating device-specific user models that capture the unique behavioral characteristics of each device and position. Instead of forcing all data into a single uniform model, the system tailors each model to its specific context (device type, entry method, physical position), maintaining data consistency within each local context while accommodating global diversity.

Inventive Principle:

Principle #3Local quality

Solution Approach 2:

The system creates copies of user models for different devices and contexts. Each device-specific model is essentially a copy adapted to that device's characteristics, allowing the system to maintain consistent authentication standards across multiple devices while accounting for device-specific behavioral variations.

Inventive Principle:

Principle #26Copying

3Reliability

If secondary authentication is required frequently, then security is improved, but the user experience and productivity deteriorate

Engineering Contradiction:

Improvesecurity levelVSAvoidlogin efficiency

Core Design Contradiction:

ReliabilityVSProductivity

Solution Approach 1:

The patent implements feedback mechanisms where the system continuously monitors behavioral biometric data and adjusts authentication requirements based on confidence levels. When behavioral patterns match established user models with high confidence, the system provides feedback that secondary authentication is not needed. When confidence is low or patterns deviate, feedback triggers appropriate authentication challenges, balancing security with user convenience.

Inventive Principle:

Principle #23Feedback

Data Source

PatentUS12155651B2Systems and methods for automatically performing secondary authentication of primary authentication credentials

Publication Date: 2024.11.26 CAPITAL ONE SERVICES LLC

AI summary

The disclosed technology includes systems and methods for determining secondary authentication of a user's log-in attempts by comparing received behavioral biometric data and/or received scenario-specific data to saved behavioral biometric data and/or saved scenario-specific data, respectively. Responsive to determining that the received behavioral biometric data and/or received scenario-specific data is above a predetermined threshold of similarity with respect to the saved behavioral biometric data and/or saved scenario-specific data, respectively, the systems and methods can determine that the corresponding log-in attempt is secondarily authenticated. of a user device via behavioral biometric data. Responsive to determining that the level of similarity is not above the predetermined threshold, the systems and methods can initiate a secondary authentication method and can associate the received behavioral biometric data with a second user model.