Graph Dataset Visualization for Machine Learning Suitability
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current graph machine learning techniques face challenges in accurately determining the suitability of graph datasets for training due to knowledge gaps and high computational resource requirements, especially when dealing with large subgraphs and non-Euclidean data, leading to potential brittleness and inefficiency in model training.
Innovation Solution
A method involving a computing system that uses task queues to parallelize the computation of graph attributes, transmitting these attributes to a web service, and rendering graph views on a client browser for interactive analysis, enabling stepwise exploration and visualization of graph datasets to assess their suitability for machine learning tasks.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Measurement precision
If subgraphs are extracted and analyzed to determine variations amongst them, then the accuracy of learning of the graph machine learning model is improved, but the computational resources required increase significantly
Solution Approach 1:
The patent segments the graph dataset into multiple subgraphs extracted around specific nodes, allowing parallel processing of these subgraphs through task queues. This segmentation enables the system to analyze variations amongst subgraphs independently, improving measurement precision while distributing computational load across multiple parallel tasks rather than processing everything sequentially.
Solution Approach 2:
The patent implements a sampling approach where not all possible subgraphs are analyzed in full detail. Instead, a representative subset of subgraphs is selected and analyzed to determine variations. This partial action approach provides sufficient information to assess dataset suitability without requiring exhaustive analysis of every possible subgraph, thereby reducing overall computational resource consumption.
2Measurement precision
If large subgraphs with many properties are used to determine dataset suitability, then the accuracy of determination is improved, but the feasibility of the process deteriorates due to computational complexity
Solution Approach 1:
The patent divides large subgraphs with many properties into smaller, more manageable units for processing. Task queues are generated to process different aspects of the subgraphs in parallel, breaking down the complex analysis into discrete, feasible tasks that can be executed independently and aggregated to form the overall determination.
Solution Approach 2:
The patent introduces a web service as an intermediary between the task queues and the client browser. This intermediary manages the complex data flow and coordinate the processing of large subgraphs, abstracting the computational complexity from the user interface and providing a simplified interaction model that maintains feasibility despite the underlying complexity.
3Ease of operation
If visualization techniques are used to bridge knowledge gap between graph datasets and machine learning, then the ease of operation is improved, but the device complexity increases
Solution Approach 1:
The patent transforms complex graph data into visual representations that add a new dimension to the data presentation. By rendering graph views that display subgraph variations and properties in visual form, the system makes the data more accessible and easier to interpret, bridging the knowledge gap between raw graph datasets and machine learning concepts through visual intuition.
Solution Approach 2:
The web service acts as an intermediary that handles the complexity of generating and managing visualizations. It receives processed subgraph data from task queues, coordinates the generation of appropriate visual representations, and delivers them to the client browser. This intermediary approach encapsulates the complexity within the service layer while presenting a simplified interface to users.
Data Source
AI summary
In an embodiment, operations include receiving a graph dataset including a set of graphs. The operations further include generating, by a task scheduler, a set of task queues configured to process the received graph dataset in parallel. The operations further include determining, by the set of task queues, a set of graph attributes for each graph of the set of graphs. The operations further include receiving, by a web service, one or more graph attributes of the determined set of graph attributes. The operations further include transmitting, by the web service, attribute information including the received one or more graph attributes to a client browser. The client browser is configured to determine a set of graph views based on the transmitted attribute information. The operations further include controlling rendering of the determined set of graph views on the client browser.