Three-Axis Graph Tensor Design for Hierarchical Attribute Retention
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing deep tensor (DT) methods for machine learning with graph-structured data fail to distinguish between the partial structure representing transaction networks and hierarchical information of account data, leading to inaccurate determinations due to loss of attribute information when expanding graph information to tensors with four or more axes.
Innovation Solution
A method that generates a tensor with a third axis combining hierarchical structure attribute information, allowing the use of three-axis tensors to retain attribute information, thereby distinguishing between transaction and hierarchical structures, enhancing determination precision.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If graph information is expanded to tensors with four or more axes to include hierarchical attribute information, then the quantity of information is improved, but the loss of information occurs due to inability to distinguish between transaction network structure and hierarchical attribute structure
Solution Approach 1:
The patent transforms the representation of hierarchical attribute information by mapping it onto a dedicated dimension (third axis) of the tensor, rather than attempting to represent it within the existing two-axis transaction network structure. This dimensional transformation allows the model to preserve both the transaction network structure (first and second axes) and hierarchical attribute information (third axis) simultaneously, preventing information loss while maintaining structural distinguishability.
2Ease of operation
If traditional deep tensor methods are used to process graph-structured data, then the ease of operation is maintained, but the measurement precision deteriorates due to inaccurate determination of transaction and hierarchical structures
Solution Approach 1:
The patent segments the tensor structure into distinct functional components: the first and second axes represent the transaction network structure, while the third axis represents hierarchical attribute information. This segmentation allows the model to process different types of information through appropriate computational operations for each component, thereby improving determination precision while maintaining ease of operation through a systematic and organized approach.
Data Source
AI summary
A non-transitory computer-readable recording medium having stored therein a machine learning program executable by one or more computers, the machine learning program including an instruction for generating a tensor comprising a first axis, a second axis, and a third axis, the first axis and the second axis representing relationships of a plurality of nodes included in graph information including data representing attributes of the plurality of nodes in the hierarchical structure, the third axis representing separately first data included in a first layer of the hierarchical structure and second data included in a second layer of the hierarchical structure, and an instruction for training a machine learning model by using the tensor as an input.