Graph Upscaling via Hash-Based Edge Generation
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Existing graph upscaling methods fail to maintain structural properties of original graphs when scaling, often requiring extensive memory and time, and are limited in scale due to decomposition and parameter extraction processes.
Innovation Solution
The method employs evolutionary hash-based and recursive preferential attachment to add edges to the original graph without sampling or parameterization, enabling parallel processing and maintaining structural properties, allowing for scalable graph upscaling to a trillion-unit scale.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Manufacturing precision
If existing graph upscaling methods use sampling and parameterization, then the upscaling process can be performed, but the structural properties of the original graph are not maintained
Solution Approach 1:
The patent extracts and eliminates the sampling and parameterization steps from the graph upscaling process. By removing these preprocessing operations, the method directly operates on the original graph structure, thereby maintaining structural properties while simplifying the overall process.
Solution Approach 2:
Instead of the conventional approach of sampling first then upscaling, the patent inverts the process by directly upsampling the original graph without prior sampling or parameterization. This reversal preserves the original graph's structural characteristics throughout the upscaling operation.
2Quantity of substance
If existing graph upscaling methods process large-scale graphs, then more graph data can be generated, but the run time becomes excessively long
Solution Approach 1:
The patent performs preliminary preparation by organizing the original graph data in a format optimized for direct upsampling operations. This preliminary structuring enables efficient parallel processing during the actual upscaling, significantly reducing the time required to generate large-scale graphs.
Solution Approach 2:
The patent implements dynamic parallel processing where multiple processing units operate simultaneously on different portions of the graph upscaling task. This dynamic approach allows the system to scale computation resources according to the graph size, maintaining efficient processing times even for trillion-unit scale graphs.
3Quantity of substance
If existing graph upscaling methods increase graph scale, then larger graphs are generated, but memory resources are exhausted
Solution Approach 1:
The patent segments the graph upscaling process into independent parallel tasks that can be distributed across multiple processing units. Each segment processes a portion of the graph independently, allowing memory resources to be distributed and reused across segments rather than requiring all graph data to reside in memory simultaneously.
Solution Approach 2:
The patent transitions from single-machine sequential processing to multi-dimensional parallel processing across multiple machines and processing units. This dimensional expansion distributes memory requirements across the distributed system, enabling the generation of trillion-unit scale graphs without exhausting memory resources on any single machine.
4Adaptability or versatility
If existing graph upscaling methods process graphs beyond limited scale, then larger graphs can be generated, but the system fails due to resource constraints
Solution Approach 1:
The patent creates a universal graph upscaling framework that can handle graphs of any scale from small to trillion-unit size using the same core algorithm. The system adapts its parallel processing configuration based on the input graph size, maintaining high success rates across the entire scale range without requiring different methods for different scales.
Solution Approach 2:
The patent implements feedback mechanisms that monitor resource utilization during the upscaling process and dynamically adjust processing parameters. This feedback control prevents resource exhaustion by scaling computation and memory usage appropriately, ensuring reliable operation across all graph scales from small to trillion-unit size.
Data Source
AI summary
A graph upscaling method and apparatus is disclosed. The graph upscaling apparatus may store, in a main memory, original graph data including topological data including a vertex and an edge, determine a parent edge using a hash function and the original graph data, and generate a new edge based on the determined parent edge. The graph upscaling apparatus may store, in a main memory, original graph data including topological data including a vertex and an edge, allocate identities (IDs) of edges to be generated to cores included in a central processing unit (CPU) based on the number of the cores, and generate new edges based on the IDs of the edges using a hash function and the original graph data.