Attribute Machine Learning System for Scalable Product Description Generation

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Existing ecommerce platforms face challenges in generating accurate and scalable product descriptions for item images, requiring significant human resources for review and drafting.

Innovation Solution

A product description generating system that includes an attribute machine learning system and an attribute quality control (QC) tool, dynamically generating descriptions based on item images and allowing for user review through a novel graphical user interface.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Measurement precision

If manual review and drafting of product descriptions is used, then accuracy of descriptions is improved, but productivity and scalability deteriorate due to significant human resources required

Engineering Contradiction:

Improveaccuracy of product descriptionsVSAvoidproductivity of product listing

Core Design Contradiction:

Measurement precisionVSProductivity

Solution Approach 1:

The system enables automated self-service generation of product descriptions through machine learning models that automatically extract attributes from images and generate descriptions without requiring manual human intervention for each product, thereby improving productivity while maintaining accuracy through intelligent algorithms

Inventive Principle:

Principle #25Self-service

Solution Approach 2:

The patent replaces the mechanical human review and drafting process with an automated machine learning system that uses computer vision and natural language generation to create product descriptions, eliminating the need for manual labor while maintaining or improving description quality

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Productivity

If automated machine learning generation is used, then productivity is improved, but measurement precision deteriorates due to potential inaccuracies in generated descriptions

Engineering Contradiction:

Improveproductivity of product listingVSAvoidaccuracy of product descriptions

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The system incorporates feedback mechanisms where generated descriptions and attribute extractions are evaluated and refined based on performance metrics, allowing the machine learning models to continuously improve accuracy while maintaining high productivity through automated processing

Inventive Principle:

Principle #23Feedback

Solution Approach 2:

The system performs preliminary attribute extraction and description generation before final review, allowing for automated pre-processing that catches most errors early and enables focused human review only on edge cases, thereby maintaining both productivity and precision

Inventive Principle:

Principle #10Preliminary action

3Measurement precision

If comprehensive user review of all generated descriptions is implemented, then measurement precision is improved, but loss of time increases due to extensive human review required

Engineering Contradiction:

Improveaccuracy of attribute valuesVSAvoidtime for quality control review

Core Design Contradiction:

Measurement precisionVSLoss of time

Solution Approach 1:

The system implements partial review by focusing human quality control efforts only on attribute values with low confidence scores or edge cases, while automatically accepting high-confidence predictions without human review, thereby reducing time loss while maintaining precision for critical attributes

Inventive Principle:

Principle #16Partial or excessive action

Solution Approach 2:

The system applies different review strategies to different attribute types and confidence levels, with high-confidence attributes requiring no review, medium-confidence attributes receiving automated validation, and low-confidence attributes receiving human review, optimizing the balance between precision and time efficiency

Inventive Principle:

Principle #3Local quality

4Productivity

If scalable automated generation is implemented, then productivity is improved, but device complexity increases due to machine learning infrastructure requirements

Engineering Contradiction:

Improvescalability of product listingVSAvoidcomplexity of generation system

Core Design Contradiction:

ProductivityVSDevice complexity

Solution Approach 1:

The system segments the product description generation task into distinct modules including image processing, attribute extraction, description generation, and quality control components, allowing each module to be independently optimized and maintained, thereby managing complexity while enabling scalable deployment

Inventive Principle:

Principle #1Segmentation

Data Source

PatentUS12266159B1Generating descriptions of items from images and providing interface for scalable review of generated descriptions

Publication Date: 2025.04.01 REALREAL INC

AI summary

A product description generating system dynamically generates a description of an item based on one or more images. The product description generating system includes an attribute machine learning system that receives images of the item and determines at least one attribute value and corresponding prediction confidence value based on the image data. The user review interface provides an interface for reviewing attribute values generated by the attribute machine learning system. If the prediction confidence value is outside a threshold, then the attribute value is reviewed by a user via the user review interface. If, on the other hand, the prediction confidence value is within a threshold, then the attribute value is used to generate the description without user review. The user review interface groups items to be reviewed by taxon and facilitates user review of many items at scale.