Edge Large AI Models: Revolutionizing 6G Networks

TL;DR

Realizing AI-native 6G via edge LAMs is hampered by massive model sizes, data privacy concerns, and constrained edge resources. The paper proposes a holistic framework centered on model decomposition and resource management, including FedFT with split models, Looped Tensor Parallelism for full-parameter training, and a microservice-based multimodal inference stack, plus edge-aware air-interface applications. Key contributions include a practical architecture for collaborative training, privacy-preserving mechanisms, adaptive microservice orchestration and migration, and federated channel prediction with graph-based beamforming, demonstrated in a case study. This work enables scalable, low-latency, privacy-preserving edge AI services for 6G and informs practical directions for deploying large models at the network edge.

Abstract

Large artificial intelligence models (LAMs) possess human-like abilities to solve a wide range of real-world problems, exemplifying the potential of experts in various domains and modalities. By leveraging the communication and computation capabilities of geographically dispersed edge devices, edge LAM emerges as an enabling technology to empower the delivery of various real-time intelligent services in 6G. Unlike traditional edge artificial intelligence (AI) that primarily supports a single task using small models, edge LAM is featured by the need of the decomposition and distributed deployment of large models, and the ability to support highly generalized and diverse tasks. However, due to limited communication, computation, and storage resources over wireless networks, the vast number of trainable neurons and the substantial communication overhead pose a formidable hurdle to the practical deployment of edge LAMs. In this paper, we investigate the opportunities and challenges of edge LAMs from the perspectives of model decomposition and resource management. Specifically, we propose collaborative fine-tuning and full-parameter training frameworks, alongside a microservice-assisted inference architecture, to enhance the deployment of edge LAM over wireless networks. Additionally, we investigate the application of edge LAM in air-interface designs, focusing on channel prediction and beamforming. These innovative frameworks and applications offer valuable insights and solutions for advancing 6G technology.

Figures (4)

• Figure 1: Summary of the techniques and advantages in the proposed edge LAMs framework.
• Figure 2: Federated fine-tuning over wireless networks.
• Figure 3: An illustration of the proposed microservice architecture for edge LAM.
• Figure 4: Federated LAM for channel prediction.