Persona Classifier for Shared Account Recommendation Accuracy

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing digital-content-service systems fail to accurately detect user interests within shared accounts, often providing inconsistent or inaccurate digital-content recommendations due to the lack of individual profiles and inconsistent content-consumption data.

Innovation Solution

A persona identification system that analyzes content-consumption events to determine multiple personas within a user account, using a persona classifier to predict the correct persona for content requests and generate personalized recommendations based on contextual features.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If multiple profiles are created within a user account to distinguish different users, then digital-content recommendations can be more accurately delivered to individual users, but users may not consistently use the correct profile or may provide inaccurate personal information, leading to inconsistent or incorrect recommendations

Engineering Contradiction:

Improveaccuracy of user interest detectionVSAvoiduser profile management complexity

Core Design Contradiction:

Measurement precisionVSEase of operation

Solution Approach 1:

The system automatically detects and identifies different user personas by analyzing content-consumption behavior patterns without requiring manual profile creation or user intervention. The persona classifier autonomously processes consumption events to distinguish between multiple users sharing an account, eliminating the need for users to manually manage profiles while maintaining accurate recommendation personalization.

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the manual mechanical system of profile creation and selection with an automated machine-learning-based persona classification system. Instead of relying on users to create and select profiles, the system uses a trained persona classifier that automatically identifies which user persona is currently consuming content based on behavioral patterns, substituting automated intelligent processing for manual user actions.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If digital-content recommendations are based on content-consumption behavior from a shared account, then recommendations can be generated without requiring user profiles, but the system cannot adjust recommendations to different users who use the same computing device with the same digital content provider

Engineering Contradiction:

Improverecommendation personalization capabilityVSAvoiduser identity distinction

Core Design Contradiction:

Adaptability or versatilityVSLoss of information

Solution Approach 1:

The system segments the content-consumption behavior data into distinct patterns that correspond to different user personas. By analyzing variations in consumption patterns, the persona classifier divides the shared account's behavior into separable user-specific segments, enabling the system to distinguish between multiple users and provide personalized recommendations for each without requiring explicit user identification.

Inventive Principle:

Principle #1Segmentation

3Productivity

If device-specific detection or isolated context features are used to make digital-content recommendations, then recommendations can be generated quickly, but the system fails to distinguish between users of shared accounts

Engineering Contradiction:

Improverecommendation generation speedVSAvoiduser distinction accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The persona classifier serves multiple functions simultaneously: it identifies user personas, analyzes content-consumption patterns, and enables personalized recommendations all within a single unified system. This multi-functional approach allows the system to maintain fast recommendation generation while accurately distinguishing between users, combining the speed of automated processing with the precision of behavioral analysis.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS10904599B2Predicting digital personas for digital-content recommendations using a machine-learning-based persona classifier

Publication Date: 2021.01.26 ADOBE INC

AI summary

This disclosure relates to methods, non-transitory computer readable media, and systems that determine multiple personas corresponding to a user account for digital content and train a persona classifier to predict a given persona (from among the multiple personas) for content requests associated with the user account. By using the persona classifier, the disclosed methods, non-transitory computer readable media, and systems accurately detect a given persona for a content request upon initiation of the request. Based on determining the given persona, in some implementations, the methods, non-transitory computer readable media, and systems generate a digital-content recommendation for presentation on a client device associated with the user account.