Machine-Learned Graph Model Segregation for Distributed Processing
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current systems for autonomous vehicles struggle with efficient processing of machine-learned models, leading to underutilization of computation resources and increased processing times, which can impact safety and accuracy in navigation tasks.
Innovation Solution
A method for segregating machine-learned graph models into portions based on minimum processing capabilities and assigning these portions to computation resources with matching capabilities, optimizing distributed processing across various hardware resources such as GPUs and FPGAs.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Productivity
If machine-learned models are processed using current systems, then processing can be performed, but computation resources are underutilized and processing times increase
Solution Approach 1:
The patent segments the machine-learned model into multiple portions and distributes them across different computation resources. The system identifies computation resources with different capabilities and assigns model portions to matching resources, ensuring optimal utilization. This segmentation resolves the contradiction by enabling parallel processing across multiple resources, thereby improving productivity while reducing overall processing time through efficient resource allocation.
2Productivity
If computation resources are distributed across different hardware, then resource utilization can be optimized, but system complexity increases
Solution Approach 1:
The patent introduces an intermediary system that manages the distribution and coordination of model portions across different computation resources. This intermediary layer handles the complexity of resource matching, capability assessment, and assignment optimization, thereby enabling high resource utilization without requiring direct complex interactions between all system components. The intermediary abstracts the complexity, making the distributed system manageable while maintaining optimal productivity.
3Measurement precision
If model portions are assigned based on processing capabilities, then processing accuracy improves, but assignment complexity increases
Solution Approach 1:
The patent performs preliminary assessment of computation resource capabilities before assigning model portions. The system pre-identifies and catalogs the processing capabilities of available resources, creating a reference framework that guides subsequent assignment decisions. This preliminary action resolves the contradiction by establishing capability benchmarks in advance, enabling accurate model portion assignment without requiring complex real-time evaluation during the assignment process itself.
Data Source
AI summary
Systems and methods of the present disclosure are directed to a method for editing a machine-learned model to facilitate distributed processing. The method can include obtaining a machine-learned graph model comprising a plurality of connected nodes. The method can include determining a plurality of processing capabilities for a respective plurality of computation resources. The method can include determining a plurality of portions from the machine-learned graph model, wherein each of the plurality of portions comprises a respective subset of the plurality of nodes and a minimum processing capability. The method can include assigning each of the plurality of portions to a respective computation resource of the plurality of computation resources based at least in part on the minimum processing capability of a respective portion and the processing capability of the respective computation resource.