Real-time Query Suggestion for Troubleshooting
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Users face frustration when troubleshooting electromechanical devices due to a mismatch between their terminology and the knowledge base, leading to inefficient troubleshooting processes, especially on devices with limited input capabilities like multifunction devices (MFDs), resulting in costly service visits and suboptimal device performance.
Innovation Solution
A method that assists users in developing queries by storing logs of prior search sessions, ranking query suggestions based on frequency and performance, and presenting them as alternatives while typing, ensuring linguistically coherent expressions that match the knowledge base, including synonyms and corrected spellings, to bridge the terminology gap.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Ease of operation
If users enter queries using their own terminology, then the query entry is simple and quick, but the search results are inaccurate due to terminology mismatch with the knowledge base
Solution Approach 1:
The patent introduces an intermediary component (query suggestion system) that mediates between the user's natural language query and the structured knowledge base. The system analyzes the user's input and provides suggested queries that are reformulated to match the knowledge base's terminology, thus bridging the terminology gap without requiring users to learn specialized vocabulary
Solution Approach 2:
The patent replaces the mechanical approach of requiring users to manually learn and use specific terminology with an automated linguistic processing system. The system uses natural language processing and query suggestion algorithms to automatically translate user intent into appropriate search queries, eliminating the need for users to mechanically adapt to the knowledge base's terminology structure
2Measurement precision
If the knowledge base uses standardized terminology, then the search results are accurate and relevant, but users struggle to formulate queries without expertise in the terminology
Solution Approach 1:
The patent implements a self-service mechanism where the system automatically adapts to the user's input style and provides contextual query suggestions. The system learns from user interactions and automatically reformulates queries to match the knowledge base terminology, allowing users to benefit from standardized terminology without having to master it themselves
Solution Approach 2:
The patent performs preliminary action by pre-processing and analyzing the knowledge base structure and terminology before user interaction. The system pre-generates query suggestions based on common user inputs and knowledge base structures, so that when users enter queries, they immediately receive tailored suggestions that bridge their terminology with the knowledge base's standardized terms
3Measurement precision
If users iterate through multiple query versions to find effective searches, then comprehensive results can be found, but time is lost and users become frustrated
Solution Approach 1:
The patent performs preliminary action by analyzing the user's initial query intent and pre-generating multiple suggested query variations before the user needs to iterate. The system anticipates what queries the user will need and provides them in advance, eliminating the need for multiple manual query iterations while maintaining comprehensive search coverage
Solution Approach 2:
The patent implements a feedback mechanism where the system monitors user interactions with suggested queries and learns from which suggestions are selected or modified. This feedback loop allows the system to improve its query suggestion accuracy over time, providing increasingly better suggestions that reduce the need for user iteration and accelerate the troubleshooting process
Data Source
Figure 1
Figure 2
Figure 3
AI summary
A method for assisting a user to develop a query in a natural language includes receiving a user's query (S118) and, while the user's query is being entered, presenting (S114) ranked query suggestions from a collection to the user as candidates for user queries. Each query suggestion in the collection is formulated to retrieve at least one responsive instance from a knowledge base. The rankings of the query suggestions in the collection are based on stored logs of prior user sessions in the knowledge base was queried.