Ontological Machine Learning for Product Catalog Labeling
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems require manual labeling of products in digital catalogs, which is time-consuming and inefficient, especially for large product catalogs, and do not effectively utilize machine learning for dynamic retraining and batch labeling.
Innovation Solution
Implementing a machine learning model for product ontology that automatically labels products by vectorizing text and image data, performing similarity searches, and dynamically retraining as data is labeled, allowing for quick and intuitive labeling of large batches of similar products.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If manual labeling is used for products in digital catalogs, then labeling accuracy can be maintained through human judgment, but the time and labor required increases significantly
Solution Approach 1:
The patent replaces manual human labeling (mechanical process) with an automated machine learning system that uses vectorization and similarity searching to label products. This substitution dramatically reduces labeling time while maintaining accuracy through algorithmic consistency and dynamic retraining on newly labeled data.
Solution Approach 2:
The system enables self-service labeling where the machine learning model automatically labels products without human intervention. The model improves itself over time by learning from newly labeled products, making the system progressively more accurate and autonomous without requiring continuous manual oversight.
2Productivity
If automated machine learning labeling is implemented, then labeling speed and efficiency improve, but system complexity increases
Solution Approach 1:
The patent breaks down the complex labeling task into distinct modular components: text vectorization, image vectorization, similarity searching, and dynamic retraining. Each component handles a specific aspect of the labeling process, making the overall system more manageable and easier to implement despite the automated nature of the solution.
3Measurement precision
If dynamic retraining of the machine learning model is performed, then labeling accuracy improves over time, but computational resources and time are consumed
Solution Approach 1:
The system performs dynamic retraining periodically or incrementally as new products are labeled, rather than continuously. This periodic action allows the model to improve accuracy over time through learning from new data while avoiding the excessive computational burden of constant retraining, balancing accuracy improvement with resource conservation.
Data Source
AI summary
Systems and methods are provided for receiving, at a server, a selection of an anchor product from an electronic catalog stored in at least one storage device communicatively coupled to the server, and vectorizing at least one of text and images associated with the selected anchor product and other products in the catalog. At least one of key words may be determined from text data and key images from image data for each product of the catalog. Vectors may be formed from at least one of the keywords and key images, and concatenating the separate vectors together to form final vectors for the products. A similarity search may be performed using the final vectors to determine a group of similar products from the vectorized products of the catalog. Selected products that are within a same slot as the anchor product may be labelled in batch.