Query Matching for Implicit Image Search Intent
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing search engines require users to input the word 'picture' to retrieve image search results, leading to inaccurate results when users intend to find pictures but do not include this term in their query.
Innovation Solution
A method and apparatus that match user queries with a query candidate set containing queries implying pictures, allowing for the output of corresponding pictures on the webpage without the need for the word 'picture' in the query, by preprocessing query history and calculating related scores based on query frequency and keyword presence.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If the search engine requires users to input the word 'picture' to retrieve image search results, then the search engine can filter queries explicitly requesting pictures, but it fails to provide accurate picture results when users do not include this term in their query
Solution Approach 1:
The system pre-processes query history data to build a query candidate set containing queries that imply picture search intent, even without the word 'picture'. This preliminary action enables the system to proactively identify picture search needs before users explicitly state them, resolving the contradiction between accuracy and convenience.
Solution Approach 2:
The system calculates related scores between user queries and query candidate set entries, using feedback from query frequency statistics and keyword presence to determine picture search intent. This feedback mechanism allows the system to accurately identify picture search needs without requiring explicit user input, achieving both precision and ease of use.
2Adaptability or versatility
If the search engine processes all queries without requiring the word 'picture', then it can potentially identify implicit picture search intent, but it increases the complexity of query processing and reduces search accuracy
Solution Approach 1:
The system performs query history pre-processing in advance to generate a query candidate set, which stores pre-analyzed queries with picture implications. This preliminary action reduces the complexity of real-time query processing while maintaining the ability to identify implicit picture search intent, resolving the contradiction between adaptability and system complexity.
Solution Approach 2:
The system calculates related scores based on query frequency and keyword presence, processing only the most relevant features from query history rather than analyzing every aspect of each query. This partial action approach reduces processing complexity while still effectively identifying implicit picture search intent.
Data Source
AI summary
The present invention provides a method and device for pushing information. The method comprises: obtaining a designated query input on a web page; matching the designated query with queries in a query candidate set including queries having a tendency of carrying an image; and when a query in the query candidate set matches the designated query, outputting a corresponding image on the web page according to the designated query.