AI Chatbot Knowledge Graph Query Path Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current methods for using knowledge graphs with artificial intelligence chatbots are inflexible, limited to predefined queries, and face challenges in enumerating and maintaining query paths, leading to unexplainable results and dependency on high-quality training data.
Innovation Solution
A computer-implemented method that identifies query characteristics, maps them to base elements in a knowledge graph, generates and validates query paths based on scores, allowing for flexible and accurate query results without being limited to predefined queries.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Reliability
If predefined rules are used to call the knowledge graph, then the system can answer predefined questions, but the system becomes inflexible and cannot handle undefined queries
Solution Approach 1:
The system dynamically generates query paths based on the input question rather than following fixed predefined rules. The graph path generation module creates multiple possible query paths adaptively, allowing the system to handle both predefined and undefined queries effectively.
Solution Approach 2:
The system changes the approach from static rule-based querying to dynamic parameter-driven querying. By using question characteristics (such as entity types and relationships) as parameters to guide query path generation, the system achieves both reliability for predefined questions and adaptability for new queries.
2Productivity
If predefined question and answer templates are used, then the system can find matching templates, but the templates can only fit a small knowledge graph and require time-consuming preparation and maintenance
Solution Approach 1:
The system eliminates the need for manual template creation and maintenance by automatically generating query paths based on the knowledge graph structure and question characteristics. The graph path generation module self-adapts to the knowledge graph, removing the burden of template preparation and maintenance.
Solution Approach 2:
Instead of creating specific templates for different query types, the system uses a universal graph path generation approach that works across the entire knowledge graph. This single mechanism handles diverse queries without requiring separate templates for each scenario.
3Adaptability or versatility
If deep learning methods with graph neural networks are used, then the system can explore embedded knowledge graphs, but the query results become unexplainable and depend heavily on training data quality
Solution Approach 1:
The system introduces query paths as an intermediary between the question and the answer. These explicit query paths serve as a bridge that explains the reasoning process, making the results interpretable while still leveraging the power of graph exploration capabilities.
Solution Approach 2:
The system performs preliminary query path generation and validation before final answer extraction. By pre-establishing the query paths and validating them against the knowledge graph, the system ensures explainability is built into the process rather than added as a post-processing step.
4Measurement precision
If multiple query paths are generated in the knowledge graph, then the system can find accurate results, but enumerating all query paths becomes challenging and complex
Solution Approach 1:
The system segments the query path generation process into distinct modules: question analysis, base element mapping, query path generation, and path validation. This segmentation makes the complex task of enumerating query paths manageable by breaking it down into smaller, more tractable sub-tasks.
Solution Approach 2:
The system dynamically generates query paths based on the specific question and knowledge graph structure rather than attempting to pre-enumerate all possible paths. This dynamic approach reduces complexity by only generating relevant paths for each query while maintaining comprehensive coverage of possible answers.
Data Source
AI summary
Generating a query result utilizing a knowledge graph in an artificial intelligence chatbot is provided. Characteristics of a query are identified. The characteristics of the query are mapped to base elements of the knowledge graph in the artificial intelligence chatbot. A set of query paths are generated in the knowledge graph based on the mapping of the characteristics of the query to the base elements of the knowledge graph. One or more query paths in the set of query paths in the knowledge graph are validated based on a respective score of each query path. A query result corresponding to the query is generated based on the validated one or more query paths in the knowledge graph.