Multi-Modal Product Classification Network
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
In e-commerce, precise and efficient product classification is challenging due to the high volume of new products and dynamic categories, requiring machine learning solutions to reduce the reliance on human editors and crowd sourcing, while existing methods struggle with multi-class, multi-label, and multi-modal classification tasks.
Innovation Solution
A multi-modal computer classification network system that combines text and image inputs using state-of-the-art deep neural networks, implementing a decision-level fusion approach to improve classification accuracy by learning from both data types and selecting the best predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If human editors and crowd sourcing platforms are used to classify products, then classification accuracy can be maintained, but time consumption and economic costs increase significantly
Solution Approach 1:
The patent replaces the mechanical human editing process with an automated deep neural network system that processes product images and metadata to generate classifications. The system substitutes human cognitive work with computational algorithms, achieving both high accuracy and efficiency simultaneously.
Solution Approach 2:
The classification system performs self-service by automatically generating product classifications without human intervention. The deep neural network processes incoming product data, generates predictions, and outputs classifications autonomously, eliminating the need for human editors while maintaining high accuracy.
2Measurement precision
If human editors and crowd sourcing platforms are used to classify products, then classification accuracy can be maintained, but economic costs increase significantly
Solution Approach 1:
The patent replaces the mechanical human editing process with an automated deep neural network system that processes product images and metadata to generate classifications. The system substitutes human cognitive work with computational algorithms, achieving both high accuracy and efficiency simultaneously.
Solution Approach 2:
The classification system performs self-service by automatically generating product classifications without human intervention. The deep neural network processes incoming product data, generates predictions, and outputs classifications autonomously, eliminating the need for human editors while maintaining high accuracy.
3Productivity
If existing machine learning methods are used for product classification, then time and costs are reduced, but classification accuracy deteriorates due to inability to handle multi-class, multi-label, and multi-modal tasks effectively
Solution Approach 1:
The patent merges multiple classification tasks (multi-class, multi-label) and data modalities (images, metadata) into a unified deep neural network framework. This integration allows the system to handle complex e-commerce classification requirements simultaneously, achieving both high productivity and accuracy that existing separate methods cannot accomplish.
Solution Approach 2:
The deep neural network system is designed with universal capability to handle multiple classification tasks and data types simultaneously. The architecture processes product images, titles, descriptions, and attributes through integrated processing pathways, providing multi-functional classification that existing specialized methods cannot achieve.
4Device complexity
If single-modal classification approaches are used, then system complexity is reduced, but classification accuracy is insufficient for complex e-commerce products
Solution Approach 1:
The patent merges multiple classification tasks (multi-class, multi-label) and data modalities (images, metadata) into a unified deep neural network framework. This integration allows the system to handle complex e-commerce classification requirements simultaneously, achieving both high productivity and accuracy that existing separate methods cannot accomplish.
Solution Approach 2:
The deep neural network system is designed with universal capability to handle multiple classification tasks and data types simultaneously. The architecture processes product images, titles, descriptions, and attributes through integrated processing pathways, providing multi-functional classification that existing specialized methods cannot achieve.
Data Source
AI summary
A multi-modal computer classification network system for use in classifying data records is described herein. The system includes a memory device, a first classification computer server, a second classification computer server, and a policy computer server. The memory device includes an item records database and a labeling database. The first classification computer server includes a first classifier program that is configured to select an item record from the item database and generate a first classification record including a first ranked list of class labels. The second classification computer server includes a second classifier program that is configured to generate a second classification record including a second ranked list of class labels. The policy computer server includes a policy network that is programmed to determine a predicted class label based on the first and second ranked lists of class labels.