17456169. ORCHESTRATING FEDERATED LEARNING IN MULTI-INFRASTRUCTURES AND HYBRID INFRASTRUCTURES simplified abstract (INTERNATIONAL BUSINESS MACHINES CORPORATION)

From WikiPatents
Jump to navigation Jump to search

ORCHESTRATING FEDERATED LEARNING IN MULTI-INFRASTRUCTURES AND HYBRID INFRASTRUCTURES

Organization Name

INTERNATIONAL BUSINESS MACHINES CORPORATION

Inventor(s)

HIFAZ Hassan of Singapore (SG)

Laura Wynter of Singapore (SG)

CHAITANYA Kumar of Singapore (SG)

Ali Anwar of San Jose CA (US)

ORCHESTRATING FEDERATED LEARNING IN MULTI-INFRASTRUCTURES AND HYBRID INFRASTRUCTURES - A simplified explanation of the abstract

This abstract first appeared for US patent application 17456169 titled 'ORCHESTRATING FEDERATED LEARNING IN MULTI-INFRASTRUCTURES AND HYBRID INFRASTRUCTURES

Simplified Explanation

The patent application describes a computer system and method for coordinating federated learning in multi-infrastructures and hybrid infrastructures. Here are the key points:

  • The system includes an infrastructure federated learning orchestrator that deploys an aggregator container and party containers to different infrastructures within a cluster.
  • The orchestrator creates aggregator and party processes for federated learning across these infrastructures.
  • Federated learning artifacts are moved to the aggregator and party containers by the orchestrator.
  • The orchestrator executes federated learning training commands in the aggregator and party processes.
  • It also monitors failure events and performance metrics in these processes.
  • The orchestrator provides automated recovery of the aggregator and party processes when functional failures or performance issues are detected.

Potential applications of this technology:

  • Federated learning in multi-infrastructures: This technology enables the coordination of federated learning across different infrastructures, allowing organizations to leverage resources from multiple environments.
  • Federated learning in hybrid infrastructures: It facilitates the orchestration of federated learning in hybrid infrastructures, where both on-premises and cloud-based resources are utilized.
  • Scalable federated learning: The system's ability to deploy containers and processes across infrastructures allows for scalable federated learning, accommodating large-scale datasets and distributed computing resources.

Problems solved by this technology:

  • Infrastructure coordination: The system addresses the challenge of coordinating federated learning across multiple infrastructures, ensuring that data and processes are properly managed and synchronized.
  • Failure recovery: By monitoring failure events and providing automated recovery, the system minimizes disruptions in federated learning processes, improving overall system reliability.
  • Performance optimization: The system's monitoring capabilities enable the identification of performance issues, allowing for timely intervention and optimization of federated learning processes.

Benefits of this technology:

  • Resource utilization: Organizations can make efficient use of resources from different infrastructures, maximizing the potential of federated learning.
  • Scalability: The system's ability to scale across infrastructures enables federated learning on large-scale datasets and distributed computing resources.
  • Reliability: Automated recovery and monitoring of failure events ensure that federated learning processes continue uninterrupted, enhancing system reliability.
  • Performance optimization: The system's monitoring capabilities help optimize the performance of federated learning processes, improving overall efficiency and effectiveness.


Original Abstract Submitted

A computer-implemented method and a computer system for orchestrating federated learning in multi-infrastructures and hybrid infrastructures. An infrastructure federated learning orchestrator deploys a container of an aggregator and containers of parties to respective infrastructures in an infrastructure cluster. The infrastructure federated learning orchestrator creates aggregator and party processes of federated learning across the respective infrastructures. The infrastructure federated learning orchestrator moves federated learning artifacts to the container of the aggregator and the containers of the parties. The infrastructure federated learning orchestrator executes federated learning training commands in the aggregator and party processes. The infrastructure federated learning orchestrator monitors failure events and performance metrics in the aggregator and party processes. The infrastructure federated learning orchestrator provides automated recovery of the aggregator and party processes, in response to detecting a functional failure or a performance issue.