Neural Array Knowledge Graph System for Parallel Query Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Knowledge graphs face performance challenges, particularly in large-scale applications, due to slower algorithm speeds when processing a large number of nodes and edges, despite existing hardware and software approaches that provide limited improvements.
Innovation Solution
An integrated knowledge graph system is developed, incorporating neural arrays with processing, communication, and learning capabilities, along with a query processing engine and policy/compliance engine, which enables efficient processing and management of complex relationships and data access control.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Quantity of substance
If conventional knowledge graph processing methods are used, then the system can handle large numbers of nodes and edges, but the algorithm speed and processing performance deteriorate
Solution Approach 1:
The knowledge graph system is segmented into multiple specialized components: graph processing units (GPUs) for parallel computation, vector processing units for dense linear algebra operations, and dedicated memory structures. This segmentation allows different parts of the system to handle specific tasks simultaneously, maintaining high speed even with large numbers of nodes and edges.
Solution Approach 2:
The patent transitions from traditional single-dimensional processing to multi-dimensional parallel processing by implementing three-dimensional memory architectures and multi-layer neural network structures. This dimensional expansion enables simultaneous processing of multiple data points across different layers and dimensions, dramatically increasing algorithm speed for large-scale knowledge graphs.
2Productivity
If hardware and software approaches are added to improve performance, then processing capability increases, but system complexity increases
Solution Approach 1:
The patent merges multiple processing functions into unified hardware structures. Graph processing units combine memory storage, computation, and communication capabilities into single integrated components. Vector processing units integrate multiple processing cores with shared memory resources, reducing overall system complexity while maintaining high processing capability.
Solution Approach 2:
The processing units are designed with universal functionality to handle diverse knowledge graph operations. The same graph processing units can perform both query processing and machine learning inference tasks. Vector processing units can handle both dense and sparse matrix operations, reducing the need for specialized hardware for each function and thereby controlling system complexity.
3Loss of time
If parallel processing is implemented to enhance performance, then response time improves, but system architecture complexity increases
Solution Approach 1:
The system performs preliminary actions by pre-computing and caching frequently accessed knowledge graph data in dedicated memory structures. Query processing is preliminary organized into parallel execution plans before actual computation begins. This preliminary preparation reduces the complexity of real-time parallel coordination and accelerates response time by having data and execution strategies ready in advance.
Solution Approach 2:
The patent introduces intermediary components that manage parallel processing coordination. A query optimization layer acts as an intermediary between user queries and parallel execution engines, translating high-level queries into coordinated parallel operations. This intermediary abstraction layer simplifies the overall architecture by hiding the complexity of parallel coordination from both users and underlying hardware.
Data Source
AI summary
An integrated knowledge graph system includes neural arrays, a query processing engine, and a policy/compliance engine. Each neural array contains circuits that represent neurons and synapses. Each neuron and synapse has a processing, communication, learning and storage capability. Entities and relationships of the knowledge graph are within the plurality of neural arrays. The query processing engine is coupled in parallel to the neural arrays and processes incoming queries and passes process queries to targeted entities and relationships representing the plurality of neural arrays. The policy/compliance engine is coupled to the plurality of neural arrays and sets policies for processing in the neurons and synapses of the plurality of neural arrays. A method implements the integrated knowledge graph system.