ML-Based Product Relevance and Offense Filtering
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online shopping concierge platforms face inefficiencies in matching customers with relevant products due to the lack of personalized filtering and relevance determination, which can lead to offensive or irrelevant search results.
Innovation Solution
The method involves receiving search parameters from customers, using machine learning models to determine product relevance and potential offense likelihood, and filtering results to generate ranked graphical user interfaces (GUIs) that showcase relevant and customer-friendly products.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If machine learning models are used to determine product relevance and filter search results, then the relevance and personalization of search results are improved, but the system complexity increases
Solution Approach 1:
The patent introduces machine learning models as intermediary components between the search query and product catalog. These models act as mediators that automatically determine product relevance and filter inappropriate content, replacing manual or rule-based filtering systems with intelligent algorithms that can process complex relevance determination tasks.
2Reliability
If comprehensive filtering based on multiple criteria is implemented, then the quality of search results is improved, but the processing time increases
Solution Approach 1:
The patent applies preliminary action by pre-computing and storing relevance scores and filtering criteria for products in advance. The machine learning models are trained beforehand to recognize relevant and inappropriate products, so that during actual search operations, the system can quickly retrieve and apply pre-determined filtering decisions rather than performing complex analysis in real-time.
3Adaptability or versatility
If personalized search results are provided to each customer, then the user experience is improved, but the computational resources required increase
Solution Approach 1:
The patent implements local quality by tailoring search results to individual customer preferences, histories, and characteristics. The machine learning models analyze specific customer data to determine relevance and appropriateness for each user, providing personalized filtering and ranking that adapts to local customer needs rather than applying uniform filtering to all users.
Data Source
AI summary
To determine search results for an online shopping concierge platform, the platform may receive, from a computing device associated with a customer of an online shopping concierge platform, data describing one or more search parameters input by the customer; identify, based at least in part on the data describing the search parameter(s), products offered by the online shopping concierge platform that are at least in part responsive to the search parameter(s); and determine, for each product and based at least in part on one or more machine learning (ML) models, a relevance of the product to one or more taxonomy levels of a product catalog associated with the online shopping concierge platform, a likelihood that the customer would be offended by inclusion of the product amongst displayed responsive search results, and/or the like.