Query Rewriting for Speech Assistant Understanding
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current speech assistant products often fail to understand diverse user queries, leading to default or rule-based replies that degrade user experience by appearing stiff and unintelligent.
Innovation Solution
A query rewriting method and apparatus that extracts context and intention information from original queries, using a machine vocabulary collection to determine a new query that improves machine understanding, enhancing retrieval efficiency and user experience.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If rule-based replies are used for unknown queries, then the system can provide a response, but the user experience deteriorates due to stiff and unintelligent replies
Solution Approach 1:
The patent introduces a query rewriting module as an intermediary component between the user's original query and the search system. This mediator translates diverse, unstructured user queries into standardized machine vocabulary queries, enabling the system to handle unknown queries intelligently rather than relying solely on pre-configured rules. The rewriting module acts as a bridge that converts human language variations into machine-understandable formats.
Solution Approach 2:
The patent changes the parameter of query representation from raw user input to rewritten machine vocabulary-based queries. By transforming the query parameters through the rewriting module, the system can maintain response reliability while improving user experience. The parameter transformation enables the system to understand and respond to diverse user expressions accurately.
2Device complexity
If the system uses default or rule-based replies, then implementation is simple, but the product appears stiff and not smart enough
Solution Approach 1:
The patent implements preliminary action by pre-building a machine vocabulary collection and training the query rewriting module before actual user interactions. The system pre-processes and stores standardized vocabulary items, enabling it to automatically rewrite user queries without requiring complex real-time decision-making. This preliminary preparation allows the system to appear intelligent while maintaining manageable implementation complexity.
3Productivity
If diverse speech queries are not rewritten, then the system processes queries directly, but the machine understanding degree is insufficient
Solution Approach 1:
The query rewriting module serves as an intermediary that processes between raw user queries and machine understanding. It translates diverse speech queries into standardized machine vocabulary representations, significantly improving the machine's understanding degree. The mediator architecture allows for efficient processing by leveraging pre-trained models and vocabulary collections.
Solution Approach 2:
The patent replaces mechanical rule-based query matching with an AI-based rewriting mechanism. Instead of using simple keyword matching or pre-configured rules, the system employs machine learning models to rewrite queries, substantially improving understanding accuracy. This substitution of mechanical systems with intelligent algorithms resolves the contradiction between processing efficiency and understanding precision.
Data Source
AI summary
Provided are a query rewriting method and apparatus, a device and a storage medium, relating to the technical field of data processing and, in particular, to technical fields including artificial intelligence, speech technology, intelligent search and deep learning. The solution includes, in response to a query rewriting request, extracting at least one of context information of an original query and intention information of the original query; and determining a new query based on a machine vocabulary collection and the at least one of the context information and the intention information. The understanding degree to the new query by a machine is greater than the understanding degree to the original query by the machine.