Interest Taxonomy Mapping for Overloaded Visual Search Queries
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Online services struggle to effectively respond to text-based queries with visual content, as typical queries are brief and often overloaded, making it challenging to determine the intended context and relevance of visual content items.
Innovation Solution
A system that expands text-based queries by aggregating terms from frequently engaged visual content items, maps the expanded query to an interest taxonomy using a trained machine learning model, and selects relevant content items based on predicted scores for interest nodes.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the online service uses a simple text-based query matching approach, then the system complexity is low, but the relevance and accuracy of visual content retrieval deteriorates due to brief and overloaded queries
Solution Approach 1:
The patent segments the query processing into multiple components: query expansion module that generates additional terms, interest taxonomy module that organizes content categories, and scoring module that ranks results. This segmentation allows each component to specialize in one aspect of improving retrieval accuracy without overwhelming the entire system
Solution Approach 2:
The patent introduces an interest taxonomy as an intermediary layer between the text-based query and the visual content database. This taxonomy acts as a mediator that translates brief user queries into structured interest categories, enabling more accurate visual content retrieval without requiring direct complex matching between queries and all visual content items
2Measurement precision
If the online service expands text-based queries by aggregating terms from frequently engaged visual content items, then the query accuracy improves, but the processing time increases
Solution Approach 1:
The patent performs query expansion in advance by pre-aggregating terms from frequently engaged visual content items and pre-organizing the interest taxonomy structure. This preliminary preparation reduces the processing time during actual query execution, as the expansion rules and category mappings are already established
Solution Approach 2:
The patent dynamically adjusts parameters such as the number of expansion terms to add, the depth of taxonomy traversal, and the scoring thresholds based on query characteristics and performance metrics. This allows the system to optimize the balance between query accuracy and processing time for different query types and usage scenarios
3Measurement precision
If the online service maps expanded queries to interest taxonomy using machine learning models, then the content selection accuracy improves, but the computational resources required increase
Solution Approach 1:
The patent applies machine learning models selectively rather than to all queries. For simple queries with clear intent, traditional matching methods are used. ML models are applied to ambiguous or complex queries where they provide the most value, reducing overall computational resource consumption while maintaining high accuracy where needed
Solution Approach 2:
The patent adjusts ML model parameters such as model complexity, training data size, and inference confidence thresholds to optimize the balance between content selection accuracy and computational resource usage. Different models with varying resource requirements can be selected based on query characteristics and available resources
Data Source
AI summary
Systems and methods for identifying relevant content within a corpus of visual content items in response to a user's text-based query are presented. In response to a text-based query, the query is mapped to a most-engaged content item of the corpus of visual content items included in responses to the query from a plurality of users. At least one text-based term associated with the most-engaged content item is identified and combined with the query from an expanded query. The expanded query is mapped to an interest node of an interest taxonomy and content items associated with the mapped interest node are identified. At least some of the content items associated with the mapped interest node are selected and returned as response content to the received query.