Edge Learning via Federated Split Decision Transformers for Metaverse Resource Allocation

2026-02-17

AI Summary

This work addresses the challenge of delivering low-latency, high-visual-quality metaverse services in heterogeneous multi-radio access mobile edge computing (MEC) environments, where conventional federated learning struggles to balance performance and resource efficiency due to full-model transmission and simplistic aggregation. To overcome this limitation, the authors propose Federated Split Decision Transformer (FSDT), which vertically partitions the Transformer architecture: embedding and prediction layers are retained locally to adapt to individual user contexts, while shared global layers are offloaded to the cloud for cross-MEC collaborative training. Integrated with offline reinforcement learning for dynamic resource allocation, FSDT maintains stringent quality-of-service requirements while offloading approximately 98% of model parameters to the cloud, substantially reducing MEC computational overhead and improving user quality of experience by up to 10% in heterogeneous settings.

Technology Category

• Machine Learning: Distributed Machine Learning & Federated Learning
• Multiagent Systems: Mechanism Design
• Computer Vision: Multi-modal Vision

Application Category

• User Modeling, Personalization and Recommendation: Federated recommendation systems and personalization
• Systems and Infrastructure for Web, Mobile and WoT: Federated Web and WoT systems, including distributed, federated and edge-based data processing
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Mobile edge computing (MEC) based wireless metaverse services offer an untethered, immersive experience to users, where the superior quality of experience (QoE) needs to be achieved under stringent latency constraints and visual quality demands. To achieve this, MEC-based intelligent resource allocation for virtual reality users needs to be supported by coordination across MEC servers to harness distributed data. Federated learning (FL) is a promising solution, and can be combined with reinforcement learning (RL) to develop generalized policies across MEC-servers. However, conventional FL incurs transmitting the full model parameters across the MEC-servers and the cloud, and suffer performance degradation due to naive global aggregation, especially in heterogeneous multi-radio access technology environments. To address these challenges, this paper proposes Federated Split Decision Transformer (FSDT), an offline RL framework where the transformer model is partitioned between MEC servers and the cloud. Agent-specific components (e.g., MEC-based embedding and prediction layers) enable local adaptability, while shared global layers in the cloud facilitate cooperative training across MEC servers. Experimental results demonstrate that FSDT enhances QoE for up to 10% in heterogeneous environments compared to baselines, while offloadingnearly 98% of the transformer model parameters to the cloud, thereby reducing the computational burden on MEC servers.

Problem

• Metaverse
• Resource Allocation
• Federated Learning
• Mobile Edge Computing
• Quality of Experience

Innovation

• Federated Split Decision Transformer
• Mobile Edge Computing
• Offline Reinforcement Learning
• Model Partitioning
• Heterogeneous Multi-RAT