Graph Sequence Networks for Preserving Sequence Characteristics
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Graph neural networks (GNN) lose sequence characteristics when processing sequence data, leading to reduced accuracy in classification predictions due to summarizing vectors into preset dimensions, which fails to accurately reflect inherent characteristics of input objects.
Innovation Solution
A graph sequence network is used to represent nodes with feature matrices, updating these matrices using adjacent node matrices through cooperative attention coding, preserving sequence characteristics and mining correlations between nodes, and using machine learning models for classification predictions.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If vectors are summarized into preset dimensions in GNN, then processing efficiency is improved, but sequence characteristics are lost leading to reduced classification accuracy
Solution Approach 1:
The patent changes the parameter representation from fixed preset dimensions to dynamic feature matrices that adapt to sequence characteristics. Each node maintains a feature matrix with dimensions determined by the sequence data rather than predetermined values, allowing the model to preserve sequence information while maintaining computational efficiency through matrix operations.
Solution Approach 2:
The patent transitions from traditional vector summarization to feature matrix representation, adding an additional dimension to the data structure. This feature matrix structure allows simultaneous preservation of sequence characteristics and efficient processing through matrix-based operations in the graph neural network.
2Measurement precision
If feature matrices are updated using adjacent node matrices through cooperative attention coding, then correlation mining is improved, but computational complexity increases
Solution Approach 1:
The patent segments the computational process into distinct layers with specific functions: cooperative attention coding layer for correlation extraction, feature matrix update layer for information integration, and classification layer for prediction. This segmentation allows complex correlation mining to be broken down into manageable steps, improving accuracy while controlling computational complexity through structured processing.
Solution Approach 2:
The patent introduces cooperative attention coding as an intermediary mechanism between adjacent nodes. This intermediary process extracts correlations through attention mechanisms before updating feature matrices, enabling precise correlation mining while managing computational complexity through the structured intermediate representation.
Data Source
AI summary
A method includes constructing a graph including a plurality of nodes for a set of sequences, wherein each node corresponds to a sequence in the set of sequences; for each node, determining an initial feature matrix of the node, wherein the initial feature matrix of the node includes initial vectors of various elements in a sequence corresponding to the node; and, inputting the initial feature matrix of the node of the graph into a graph sequence network to enable the graph sequence network to update the feature matrix of the node using the feature matrix(es) of adjacent node(s) of the node; and obtaining a feature matrix output by the graph sequence network of each node to perform a sequence-based classification prediction using output feature matrixes, wherein the feature matrix output for each node includes updated vectors corresponding to the various elements in the sequence corresponding to the node.