Joint Optimization of Training and Inference in Federated Edge Learning via Constrained Multi-Objective Deep Reinforcement Learning
arXiv:2605.25916v1 Announce Type: new Abstract: Federated edge learning (FEEL) has recently emerged as a promising paradigm for achieving edge intelligence (EI) via enabling collaborative model training across edge devices while protecting data privacy. In this paper, we put forth an online optimization framework that jointly manages federated training and inference on resource-constrained edge devices. We introduce a tandem-queue-inspired conversion mechanism that bridges inference requests and training data, and further incorporate both data and model freshness into the accuracy formulation
The proliferation of edge devices and increasing demand for privacy-preserving AI models are driving research into optimizing federated learning at the edge.
This research provides a framework for more efficient and effective deployment of AI on resource-constrained edge devices, impacting data privacy and real-time intelligence capabilities.
The joint optimization of training and inference, combined with a novel 'tandem-queue-inspired conversion mechanism,' offers a more holistic approach to managing AI operations at the edge.
- · Edge device manufacturers
- · AI developers
- · Industries requiring real-time edge intelligence (e.g., IoT, robotics)
- · Privacy-focused data platforms
- · Centralized cloud AI services (for specific use cases)
- · Inefficient edge AI models
Improved performance and reduced resource consumption for federated learning applications on edge devices.
Accelerated adoption of AI in industries sensitive to data privacy and latency, leading to new service offerings.
Potential for a more distributed and resilient global AI infrastructure, less reliant on large central data centers.
This signal links to a primary source. Continuum Brief monitors and indexes it as part of the live intelligence stream — we do not republish source content.
Read at arXiv cs.LG