Product Catalog Error Mitigation via Vector Grouping

Resolve Bottlenecks,
Find Innovative SolutionsGenerate Solutions

Solution Overview

Problem

Product catalogs often contain erroneous or outdated product types due to manual or automated errors, leading to the display of non-compliant information in graphical user interfaces, which affects the accuracy and reliability of user search experiences.

Innovation Solution

A system and method that utilize vector representations and proximity searches to group product items based on metadata, identifying and correcting errors by modifying metadata to prevent the display of non-compliant material, and employing computer vision to detect misclassifications in product types.

Engineering Contradictions & Design Principles

VSEngineering Contradiction Analysis

1Productivity

If product types are assigned manually or automatically in a product catalog, then the product catalog can be populated with product information, but errors and outdated classifications occur leading to non-compliant information display

Engineering Contradiction:

Improveproduct catalog population efficiencyVSAvoidproduct type classification accuracy

Core Design Contradiction:

ProductivityVSReliability

Solution Approach 1:

The system performs preliminary actions by generating vector representations for product items and establishing their relationships before actual classification. The graph embedding model pre-computes similarity metrics and relationships, allowing the system to proactively identify and correct classification errors before they manifest as non-compliant information displays.

Inventive Principle:

Principle #10Preliminary action

Solution Approach 2:

The patent replaces manual classification mechanisms with an automated computer vision-based system. Instead of relying on manual assignment or simple automated tagging, the system uses graph neural networks and vector space modeling to automatically detect and correct product type misclassifications, substituting human labor and basic automation with advanced AI-based classification.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

2Adaptability or versatility

If the product catalog taxonomy evolves over time, then the catalog can adapt to new product types and attributes, but existing classifications become outdated and erroneous

Engineering Contradiction:

Improvetaxonomy evolution capabilityVSAvoidproduct type classification accuracy

Core Design Contradiction:

Adaptability or versatilityVSLoss of information

Solution Approach 1:

The system implements dynamic classification by continuously updating vector representations and re-evaluating product-item relationships as the taxonomy evolves. The graph embedding model can be retrained and updated to reflect new product types and attributes, allowing the classification system to adapt dynamically rather than remaining static, thus preventing information loss during taxonomy evolution.

Inventive Principle:

Principle #15Dynamics

Solution Approach 2:

The system incorporates feedback mechanisms where the graph neural network continuously evaluates product classifications and provides corrections. The vector-based similarity metrics provide feedback on classification quality, allowing the system to identify and correct outdated or erroneous classifications as the taxonomy evolves, maintaining information accuracy through iterative refinement.

Inventive Principle:

Principle #23Feedback

3Productivity

If manual or automated product type assignment is used, then product catalogs can be populated efficiently, but the accuracy and coverage of product types significantly affect user search experience

Engineering Contradiction:

Improveproduct catalog population speedVSAvoidproduct type classification accuracy

Core Design Contradiction:

ProductivityVSMeasurement precision

Solution Approach 1:

The patent replaces manual and basic automated classification systems with an advanced graph neural network-based computer vision system. This substitution maintains the productivity benefits of automated population while dramatically improving measurement precision through AI-driven classification that can accurately distinguish between similar product types and detect misclassifications.

Inventive Principle:

Principle #28Mechanics substitution (Replace mechanical system)

Solution Approach 2:

The system changes the parameters of classification by moving from traditional rule-based or keyword-matching approaches to vector-based semantic similarity measurement. By representing products and items in a continuous vector space, the system achieves higher classification accuracy while maintaining automated efficiency, as the vector representations capture nuanced relationships between product types.

Inventive Principle:

Principle #35Parameter changes

Data Source

PatentUS20250022033A1Systems and methods for mitigating display of non-compliant information

Publication Date: 2025.01.16 WALMART APOLLO LLC

AI summary

According to an embodiment of the present disclosure, a system is provided. The system includes a processor and a non-transitory computer-readable medium storing computing instructions. The instructions, when executed on the processor, perform: receiving user search queries and product items, the product items including metadata corresponding to product types, the product types comprising an error, and the error corresponding to one of the product items having a non-compliant product type; modifying the metadata corresponding to the product types based on a product type group to mitigate a computing system from displaying non-compliant material to a user; and in response to a product type of the product type group being the non-compliant product type, performing a computer vision prediction on an image of a product corresponding to the product type to detect a misclassification in the product type group. Other embodiments are disclosed.