Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

LLM Agents as 6G Orchestrator: A Paradigm for Task-Oriented Physical-Layer Automation

Zhuoran Xiao, Chenhui Ye, Yunbo Hu, Honggang Yuan, Yihang Huang, Yijia Feng, Liyu Cai, Jiang Chang

TL;DR

A two-stage continual pre-training and fine-tuning scheme to build the field basic model and diversities of specialized expert models for meeting the requirements of various application scenarios is proposed and a novel inference framework based on semantic retrieval for leveraging the existing communication-related functions is proposed.

Abstract

The rapid advancement in generative pre-training models is propelling a paradigm shift in technological progression from basic applications such as chatbots towards more sophisticated agent-based systems. It is with huge potential and necessity that the 6G system be combined with the copilot of large language model (LLM) agents and digital twins (DT) to manage the highly complicated communication system with new emerging features such as native AI service and sensing. With the 6G-oriented agent, the base station could understand the transmission requirements of various dynamic upper-layer tasks, automatically orchestrate the optimal system workflow. Through continuously get feedback from the 6G DT for reinforcement, the agents can finally raise the performance of practical system accordingly. Differing from existing LLM agents designed for general application, the 6G-oriented agent aims to make highly rigorous and precise planning with a vast amount of extra expert knowledge, which inevitably requires a specific system design from model training to implementation. This paper proposes a novel comprehensive approach for building task-oriented 6G LLM agents. We first propose a two-stage continual pre-training and fine-tuning scheme to build the field basic model and diversities of specialized expert models for meeting the requirements of various application scenarios. Further, a novel inference framework based on semantic retrieval for leveraging the existing communication-related functions is proposed. Experiment results of exemplary tasks, such as physical-layer task decomposition, show the proposed paradigm's feasibility and effectiveness.

LLM Agents as 6G Orchestrator: A Paradigm for Task-Oriented Physical-Layer Automation

TL;DR

A two-stage continual pre-training and fine-tuning scheme to build the field basic model and diversities of specialized expert models for meeting the requirements of various application scenarios is proposed and a novel inference framework based on semantic retrieval for leveraging the existing communication-related functions is proposed.

Abstract

The rapid advancement in generative pre-training models is propelling a paradigm shift in technological progression from basic applications such as chatbots towards more sophisticated agent-based systems. It is with huge potential and necessity that the 6G system be combined with the copilot of large language model (LLM) agents and digital twins (DT) to manage the highly complicated communication system with new emerging features such as native AI service and sensing. With the 6G-oriented agent, the base station could understand the transmission requirements of various dynamic upper-layer tasks, automatically orchestrate the optimal system workflow. Through continuously get feedback from the 6G DT for reinforcement, the agents can finally raise the performance of practical system accordingly. Differing from existing LLM agents designed for general application, the 6G-oriented agent aims to make highly rigorous and precise planning with a vast amount of extra expert knowledge, which inevitably requires a specific system design from model training to implementation. This paper proposes a novel comprehensive approach for building task-oriented 6G LLM agents. We first propose a two-stage continual pre-training and fine-tuning scheme to build the field basic model and diversities of specialized expert models for meeting the requirements of various application scenarios. Further, a novel inference framework based on semantic retrieval for leveraging the existing communication-related functions is proposed. Experiment results of exemplary tasks, such as physical-layer task decomposition, show the proposed paradigm's feasibility and effectiveness.
Paper Structure (14 sections, 1 equation, 8 figures)

This paper contains 14 sections, 1 equation, 8 figures.

Table of Contents

  1. Introduction
  2. System Framework
  3. Two-Stage Model training
  4. Stage One: Knowledge Injection Through Continual Learning
  5. Stage Two: Role-Specific Fine-Tuning
  6. Workflow Design for Task-Oriented Physical-Layer Automation
  7. Role-Playing and Collaboration of Agents
  8. API Retrieve Through Vector-Based Semantic Similarity Search
  9. Tool Execution Based on Self-Organization
  10. Experiments
  11. Parameters Setting and Datasets Generation
  12. Dataset Generation
  13. Experimental Results
  14. Conclusions

Figures (8)

  • Figure 1: System framework of the proposed 6G LLM agent.
  • Figure 2: The proposed three model levels and the corresponding functions. The L1 model is trained from the pre-training L0 model and serves as the domain basic model. L2 models are fine-tuned based on the L1 model tailored for specific target scenarios or agent roles.
  • Figure 3: Learning structure of two-stage training process.
  • Figure 4: Example of the workflow of a communication system orchestrated by LLM agents. LLM agents play different roles in different parts of the system. The agents could understand UE's upper-layer task requirements and orchestrate the system step by step in an optimal way accordingly.
  • Figure 5: The winning ratio comparison of answering communication related questions between three LLMs.
  • ...and 3 more figures