A Survey on Intelligent Network Operations and Performance Optimization Based on Large Language Models

As Large Language Models (LLMs) have achieved significant success in handling multi-modal tasks such as text, images, videos, and sounds, particularly showcasing emergent capabilities in natural language tasks, they hold great potential for network operations that similarly involve vast amounts of text data, fault data, and log files. This paper focuses on the development of LLMs, detailing their fundamental principles and application scenarios across different domains. It highlights the remarkable capabilities of LLMs in tasks such as fault diagnosis, causal inference, and intelligent question answering, and applies these abilities to the field of network operations. Moreover, the paper reviews some of the…

• NN
  National Natural Science Foundation of China

  Award: 62302527
• NS
  Natural Science Foundation of Hunan Province

  Awards: 2023jj40774, 2024JJ9173
• PS
  Postdoctoral Science Foundation of Central South University