User Representation Model for Personalized Content Recommendations

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing solutions for content customization in online services fail to effectively leverage unstructured and noisy application usage records from content manipulation applications, leading to insufficient personalization due to the inability to model user characteristics accurately.

Innovation Solution

A user representation model is trained iteratively using action histories from content manipulation applications, generating user representation vectors that are adjusted to predict user actions, allowing for personalized content recommendations across different platforms.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Adaptability or versatility

If application usage records are used for content customization, then user characteristics can be modeled more comprehensively, but the noisy and unstructured nature of the data makes it difficult to derive useful information

Engineering Contradiction:

Improvecontent customization capabilityVSAvoiduser characteristic accuracy

Core Design Contradiction:

Adaptability or versatilityVSMeasurement precision

Solution Approach 1:

The patent introduces an intermediary processing layer that transforms unstructured application usage records into structured user representation vectors. This intermediary process includes parsing raw log data, extracting meaningful features, and encoding them into standardized vector representations that can be reliably used for content customization while filtering out noise.

Inventive Principle:

Principle #24Intermediary (Mediator)

Solution Approach 2:

The patent replaces manual or rule-based user characteristic modeling with machine learning models that automatically learn patterns from application usage data. The system uses trained models to transform raw usage records into user representation vectors, substituting mechanical processing with intelligent algorithms that can handle noise and unstructured data effectively.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Measurement precision

If user representation models are trained on action histories, then recommendation accuracy improves, but the complexity of the training process increases

Engineering Contradiction:

Improverecommendation accuracyVSAvoidmodel training complexity

Core Design Contradiction:

Measurement precisionVSDevice complexity

Solution Approach 1:

The patent performs preliminary actions by pre-processing application usage records and pre-training user representation models offline. User representation vectors are computed in advance from action histories and stored for later use, avoiding the need for complex real-time training during recommendation generation.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent segments the complex training process into distinct modules: data collection from action histories, feature extraction, model training, and vector generation. Each module handles a specific aspect of the process, making the overall system more manageable and easier to implement despite the inherent complexity.

Inventive Principle:

Principle #1Segmentation

3Adaptability or versatility

If user representation vectors are transformed across different domains, then content personalization can be achieved even without direct data, but the transformation process becomes more complex

Engineering Contradiction:

Improvecross-platform personalizationVSAvoidvector transformation complexity

Core Design Contradiction:

Adaptability or versatilityVSDevice complexity

Solution Approach 1:

The patent creates a universal user representation vector format that can be applied across different domains and platforms. The same vector structure and transformation methodology work for both application usage data and online service data, enabling cross-platform personalization without requiring domain-specific customization of the transformation process.

Inventive Principle:

Principle #6Universality (Multi-functionality)

Data Source

PatentUS10614381B2Personalizing user experiences with electronic content based on user representations learned from application usage data

Publication Date: 2020.04.07 ADOBE INC

AI summary

This disclosure involves personalizing user experiences with electronic content based on application usage data. For example, a user representation model that facilitates content recommendations is iteratively trained with action histories from a content manipulation application. Each iteration involves selecting, from an action history for a particular user, an action sequence including a target action. An initial output is computed in each iteration by applying a probability function to the selected action sequence and a user representation vector for the particular user. The user representation vector is adjusted to maximize an output that is generated by applying the probability function to the action sequence and the user representation vector. This iterative training process generates a user representation model, which includes a set of adjusted user representation vectors, that facilitates content recommendations corresponding to users' usage pattern in the content manipulation application.