Unified Query Rewriting Framework for Search Relevance
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing query rewriting techniques are fragmented and lack a unified solution, failing to effectively bridge the language gap between user queries and web documents, leading to suboptimal search results due to differences in language styles and vocabularies.
Innovation Solution
A unified query rewriting framework that treats query rewriting as a machine translation problem, utilizing click-through graphs to generate query-title pairs for training a model, which then decodes to produce rewritten queries that better match search engine recognizable formats, incorporating features like translation scores, language models, and penalties for similarity, length, and stop-words.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If existing query rewriting techniques are used, then query rewriting can be performed, but the solution is fragmented and lacks unity, leading to suboptimal search results
Solution Approach 1:
The patent merges multiple query rewriting sub-problems into a single unified framework. Instead of developing separate rewriters for different rewriting tasks, the system integrates them into one cohesive model that handles various rewriting scenarios simultaneously, thereby improving search relevance while reducing overall system complexity
Solution Approach 2:
The query rewriting framework is designed as a universal system that can perform multiple rewriting functions through a single model. The framework handles diverse rewriting tasks (such as query expansion, query refinement, and query transformation) using the same underlying architecture, eliminating the need for multiple specialized rewriters and improving overall search effectiveness
2Reliability
If query rewriting is performed to bridge language gap, then search results improve, but the system becomes more complex
Solution Approach 1:
The patent introduces an intermediary mechanism that facilitates communication between user queries and web documents. The query rewriting model acts as a mediator that translates user queries into search engine recognizable formats, bridging the language gap between different styles and vocabularies without requiring complex direct matching mechanisms
Solution Approach 2:
The system changes parameters of the query (such as vocabulary, syntax, and structure) to transform it into a format that better matches search engine indexes. By modifying query parameters rather than changing the fundamental search architecture, the system improves search relevance while keeping the overall model structure manageable
Data Source
AI summary
The present teaching relates to rewriting a query and providing search results. In one example, a plurality of queries is obtained. For each of the plurality of queries, one or more search results are identified. The one or more search results have been obtained in response to the query and have been previously selected by a user submitting the query. A plurality of titles is obtained. Each of the titles corresponds to one of the one or more search results with respect to one of the plurality of queries. A model is generated based on the plurality of queries and the plurality of titles. The model is to be used for rewriting a query.