GXAI Graph Correlation Visualization for Interpretability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Graph explainable artificial intelligence (GXAI) results for graph-structured data are often presented in complex, unfamiliar formats, making it difficult for users to understand and interpret, as they require drilling down through multiple layers to extract relevant information, and existing methods lack effective visualization tools to connect GXAI results with graph analysis algorithm insights.
Innovation Solution
A method that computes correlation between GXAI classification analysis and graph analysis algorithm results using vector representations, employing Pearson correlation coefficient or cosine similarity, to generate a user-friendly graphical user interface (GUI) that visualizes similarities, enhancing usability and interpretability by framing GXAI results within the context of graph analysis algorithm outputs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If GXAI classification analysis is performed on graph-structured data, then comprehensive analysis results are obtained, but the results are presented in complex, unfamiliar formats that are difficult for users to understand and interpret
Solution Approach 1:
The patent introduces graph analysis algorithm results as an intermediary to bridge the gap between GXAI classification analysis and user understanding. The system computes correlation between GXAI results and graph analysis results, using the latter as a mediator that users can more easily interpret while still capturing the essence of the GXAI analysis.
Solution Approach 2:
The patent creates a simplified copy or representation of GXAI results through graph analysis algorithm outputs. Instead of presenting raw GXAI results directly, the system generates correlated graph analysis results that serve as an accessible representation, making the complex GXAI findings interpretable while preserving the underlying analytical insights.
2Loss of information
If users drill down through multiple layers of GXAI results to extract relevant information, then detailed insights are obtained, but time consumption and operational complexity increase
Solution Approach 1:
The patent extracts the essential and relevant information from multi-layered GXAI results by correlating them with graph analysis algorithm outputs. Instead of requiring users to navigate through all layers of GXAI results, the system extracts and presents the most relevant insights through the correlated graph analysis representation.
Solution Approach 2:
The patent segments the complex GXAI results into meaningful components that can be represented through graph analysis algorithms. By dividing the comprehensive GXAI output into correlated segments, the system enables users to access relevant information efficiently without needing to drill down through entire hierarchical structures.
3Loss of information
If existing visualization tools are used to display GXAI results, then some information is presented, but effective visualization to connect GXAI results with graph analysis algorithm insights is lacking
Solution Approach 1:
The patent merges GXAI classification analysis results with graph analysis algorithm results into a unified visualization framework. By combining these two sources of information and displaying their correlations together, the system creates an integrated visualization that connects previously disconnected insights while maintaining clarity through the use of familiar graph analysis representations.
Data Source
AI summary
A method may include obtaining a first result of a graph explainable artificial intelligence (GXAI) classification analysis of a dataset of graph-structured data and a second result of a graph analysis algorithm that represents relationships between elements of the dataset. The method may include determining a correlation between the first result and the second result and generating a display within a graphical user interface (GUI) that visualizes similarities between the first result and the second result based on the correlation. Determining the correlation between the first result and the second result may include generating a first vector of the first result of the classification analysis using GXAI techniques and a second vector of the second result of the graph analysis algorithm. A Pearson correlation coefficient or a cosine similarity coefficients may be computed based on the first vector and the second vector in which the computed coefficients are indicative of the correlation.