ML Orchestrator for Automated Distributed Learning Deployment
Find Innovative SolutionsGenerate Solutions
Solution Overview
Problem
Current solutions for network data analytics in 5G communication systems lack flexibility in deploying machine learning models that meet diverse consumer requirements, including privacy, data distribution, and energy consumption constraints.
Innovation Solution
A machine learning orchestrator entity that automates the selection and configuration of distributed learning procedures, such as Federated Learning, Split Learning, and Hierarchical FL, by defining a ML profile and determining job information for local learning agents to compute analytics outputs, considering constraints and requirements.
Engineering Contradictions & Design Principles
Engineering Contradiction Analysis
1Adaptability or versatility
If distributed ML techniques are used to meet diverse consumer requirements, then adaptability is improved, but device complexity increases
Solution Approach 1:
The patent introduces an ML orchestrator entity as an intermediary between consumers and local learning agents. This orchestrator manages the complexity of deploying distributed ML techniques by coordinating between different components, selecting appropriate techniques (Federated Learning, Split Learning, etc.), and handling configuration details, thereby enabling adaptability without exposing complexity to end users.
Solution Approach 2:
The ML orchestrator entity is designed as a universal component that can handle multiple distributed ML techniques (Federated Learning, Split Learning, Federated Distillation, Hierarchical FL) through a unified interface. This multi-functionality allows the system to adapt to diverse consumer requirements while maintaining a single deployment mechanism, reducing the complexity of managing multiple separate systems.
2Ease of operation
If automated selection and configuration of learning procedures is implemented, then ease of operation is improved, but device complexity increases
Solution Approach 1:
The ML orchestrator entity implements automated self-service capabilities by autonomously selecting appropriate distributed ML techniques and configuring learning procedures based on consumer requirements without manual intervention. The orchestrator automatically analyzes requirements, selects techniques (Federated Learning, Split Learning, etc.), and configures local learning agents, thereby improving ease of operation while encapsulating complexity within the orchestrator itself.
Data Source
AI summary
The present disclosure relates to a machine learning (ML) orchestrator entity for a ML system. The ML system comprises one or more local learning agents (LLAs) and is configured to compute an analytics output for an analytics service. The ML orchestrator entity comprises first processing circuitry configured to: receive an analytics service request for the analytics service from a consumer entity; define a ML profile for the analytics service based on the analytics service request; and determine ML job information based on the ML profile, wherein the ML job information indicates, for each LLA of the one or more LLAs, a computation operation to be performed by that LLA to compute the analytics output.