Semantic Keyword Mapping for Sparse Item Title Search Matching
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current search engines struggle with matching item data that lacks detail, particularly when generated by third parties, leading to poor performance in optimized search processes due to sparse data representations.
Innovation Solution
A system utilizing a trained semantic mapping model to generate item embeddings and compare them with platform-specific keyword embeddings, selecting recommended augmentation keywords based on similarity, and modifying item titles to enhance their relevance in search queries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If current search engines use basic text matching against source content, then the search process is simple and fast, but items with sparse or inaccurate data representations fail to be matched
Solution Approach 1:
The patent introduces an intermediary semantic mapping model that translates item titles into semantic embeddings, which then serve as the basis for search matching. This intermediary layer enables more accurate matching of items with sparse data by capturing semantic meaning rather than relying solely on exact text matches, while maintaining system tractability through efficient embedding comparisons.
2Productivity
If item data is generated by third parties with limited detail, then data collection is easier and more comprehensive, but the data lacks adequate detail for optimized search processes
Solution Approach 1:
The system performs preliminary semantic enrichment by generating semantic embeddings from item titles before the search process occurs. This preliminary action captures the semantic meaning of item titles in advance, compensating for the lack of detailed descriptions in third-party generated data, and makes the enriched representations available for subsequent search operations without requiring real-time processing.
3Reliability
If semantic mapping models are applied to all item titles, then search relevance is improved, but processing time and computational resources increase
Solution Approach 1:
The patent applies semantic mapping in advance to generate embeddings for item titles and stores them for future use. This preliminary action allows the system to avoid repeated computational processing during search operations, instead performing efficient comparisons between pre-computed embeddings and query embeddings, thereby reducing real-time processing time while maintaining improved search relevance.
Data Source
AI summary
Systems and methods of item-specific keyword recommendation are disclosed. An item data structure including an item title is received and at least one item embedding is generated by applying a first trained semantic mapping model to the item title. The first trained semantic mapping model includes a first semantic mapping framework. The at least one item embedding is compared to a set of keyword embeddings representative of a set of platform-relevant keywords and a set of item-specific recommended keywords is selected from the set of platform-relevant keywords based on a similarity between the at least one item embedding and each embedding in the set of keyword embeddings. The item title is modified to include at least one of the set of item-specific recommended keywords and an interface including the modified item title is generated.