Virtual Assistant Ambiguity Resolution via Hierarchical Knowledge Graph
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Conventional knowledge graph-based question-answering services face challenges such as extensive training and maintenance requirements, time-consuming annotation processes, and inefficiencies in ambiguity resolution, leading to user dissatisfaction.
Innovation Solution
A method for ambiguity resolution in virtual assistant servers using a hierarchical knowledge graph, where concepts from user queries are correlated with contextual, intermediate, and leaf nodes, prompting the user to resolve ambiguities progressing from contextual nodes towards leaf nodes until a matching path is identified, and content associated with the matched path is output.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Extent of automation
If machine learning models are used for knowledge graph-based question-answering, then the system can provide automated responses, but the training and maintenance requirements become extensive and time-consuming
Solution Approach 1:
The patent segments the knowledge graph into a hierarchical structure with contextual nodes, intermediate nodes, and leaf nodes. This segmentation allows the system to process queries by traversing specific paths rather than requiring comprehensive machine learning training for all possible query types, thereby reducing training and maintenance time while maintaining automated response capability.
Solution Approach 2:
The patent introduces an intermediary ambiguity resolution module that acts as a mediator between the knowledge graph search and the user. When multiple paths are found, this intermediary component resolves ambiguities by analyzing path characteristics and selecting the most appropriate path, eliminating the need for extensive machine learning training while maintaining high automation.
2Productivity
If machine learning models are used for knowledge graph construction, then automated processing is achieved, but annotation requirements increase significantly in terms of time and cost
Solution Approach 1:
The hierarchical segmentation of the knowledge graph into contextual, intermediate, and leaf nodes enables structured organization of information without requiring extensive annotations. Each node type serves a specific function, allowing the system to achieve high processing efficiency through rule-based path traversal rather than machine learning annotation.
Solution Approach 2:
The system employs self-service mechanisms where the ambiguity resolution module automatically analyzes and resolves ambiguities by examining path characteristics, node types, and query context. This eliminates the need for manual machine learning annotations while maintaining high processing efficiency through automated path selection.
3Speed
If conventional knowledge graph search is used, then quick responses are provided, but ambiguity resolution capabilities are insufficient leading to user dissatisfaction
Solution Approach 1:
The patent introduces an intermediary ambiguity resolution module that quickly analyzes multiple paths and resolves ambiguities by examining path characteristics, node types, and query context. This intermediary component maintains fast response speeds while significantly improving ambiguity resolution accuracy through structured analysis rather than relying on fire-and-forget search mechanisms.
Solution Approach 2:
The system implements feedback mechanisms where the ambiguity resolution module continuously evaluates path characteristics and adjusts path selection based on analysis results. This feedback loop ensures reliable ambiguity resolution while maintaining quick response times by avoiding exhaustive search procedures.
4Reliability
If hierarchical knowledge graph with ambiguity resolution is implemented, then ambiguity resolution accuracy is improved, but system complexity increases
Solution Approach 1:
The hierarchical segmentation into contextual nodes, intermediate nodes, and leaf nodes provides a structured framework that organizes complexity in a manageable way. Each node type has defined characteristics and functions, making the system's complexity systematic rather than chaotic, and enabling reliable ambiguity resolution through structured path analysis.
Solution Approach 2:
The patent applies local quality by assigning specific characteristics and functions to different node types within the hierarchy. Contextual nodes handle high-level context, intermediate nodes manage specific domains, and leaf nodes provide detailed information. This local differentiation enables precise ambiguity resolution without requiring the entire system to be equally complex.
Data Source
AI summary
A virtual assistant server extracts concept(s) from a received user query and correlates the extracted concept(s) to a plurality of nodes in a knowledge graph to identify qualifying path(s) in the knowledge graph. Further, when the two or more qualifying paths are identified, the virtual assistant server addresses one or more ambiguities in the knowledge graph by: prompting the user device to resolve a current one of the one or more ambiguities progressing from contextual nodes towards leaf nodes in the knowledge graph; receiving a response to the prompt from the user device; and repeating the prompting and the receiving until each of the one or more ambiguities is resolved and a matching path to the received query is identified. Finally, the virtual assistant server outputs content associated with a leaf node of the matched path to the user device.