Edge Machine Learning Model Synchronization
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Cloud computing systems face increased load and latency issues due to the large amount of training data being processed, which hinders efficient machine learning operations.
Innovation Solution
Implementing a local machine learning system on a mobile computing device that can access and update a global machine learning system hosted in the cloud, allowing for periodic data synchronization and processing of changes only, thereby reducing traffic and latency.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Power
If machine learning is performed using a cloud computing system, then processing power and storage capacity are improved, but system load and latency increase
Solution Approach 1:
The patent segments the machine learning system into two parts: a global machine learning system in the cloud and a local machine learning system on the mobile device. The local system handles immediate processing needs while the global system provides comprehensive processing power, thereby reducing latency for time-sensitive operations while maintaining access to cloud-based computational resources.
Solution Approach 2:
The patent implements preliminary action by periodically updating the local machine learning system with changes from the global system before they are needed. This proactive synchronization ensures that the local system has up-to-date models and parameters ready for immediate use, reducing the time required to respond to local processing requests.
2Measurement precision
If training data is processed in the cloud computing system, then machine learning model accuracy is improved, but system load increases
Solution Approach 1:
The patent extracts the machine learning processing function from the cloud system and places it locally on the mobile device. By taking out the inference processing from the cloud, the system maintains model accuracy through periodic updates while significantly reducing the continuous load on the cloud computing system.
Solution Approach 2:
The patent creates a local copy of the machine learning model on the mobile device. This copy is periodically updated with changes from the global system, allowing the device to perform local processing with near-cloud accuracy while reducing cloud system load by eliminating the need for continuous data transmission and processing.
3Extent of automation
If training data traverses a long route to the cloud computing system, then centralized processing is improved, but data transmission time increases
Solution Approach 1:
The patent segments the data processing architecture into local and cloud components. Local processing handles immediate data needs at the edge, eliminating long transmission routes for routine operations. Only model updates and aggregated insights need to traverse the network, significantly improving effective data transmission speed while maintaining centralized coordination.
Solution Approach 2:
The local machine learning system acts as an intermediary between the mobile device and the global cloud system. It processes data locally without requiring direct communication with the cloud for each operation, thereby eliminating long transmission routes while maintaining the benefits of centralized model management through periodic updates.
4Speed
If a local machine learning system is implemented on a mobile device, then processing speed is improved, but device resource consumption increases
Solution Approach 1:
The patent applies partial action by implementing a local machine learning system that handles only local inference operations rather than full training and processing. This partial localization provides speed benefits for time-sensitive operations while avoiding the excessive energy consumption that would result from running complete machine learning workloads locally.
Solution Approach 2:
The system uses periodic action by synchronizing the local machine learning system with the global system at scheduled intervals rather than continuously. This periodic update mechanism reduces device energy consumption by minimizing active communication and computation cycles while maintaining processing speed benefits when the local system is active.
Data Source
AI summary
A system and method for performing machine learning in a mobile computing device which is configured to be coupled with a cloud computing system is disclosed. The method may include activating, on the mobile computing device, a machine learning application, which accesses a local machine learning system including a local machine learning model, periodically updating the local machine learning system based upon updates for the local machine learning system received from a global machine learning system hosted by the cloud computing system, performing machine learning based on received training data, and periodically transmitting changes to the local machine learning system from the mobile computing device to the global machine learning system hosted by the cloud computing system.