What LLMs Actually Do in Network Automation

#AI #ML #operations #automation #intent #device #config #generative #trust #communications #networking #futurenetworks #ieee #artificial-intelligence #machine-learning

Special Presentation by Fan Liu (Washington U. in St. Louis, USA)

Hosted by the Future Networks Artificial Intelligence & Machine Learning (AIML) Working Group

Date/Time: Thursday, 5 June 2025 @ 6 PM EDT

Topic:

Beyond the Buzz: What Large Language Models Actually Do (and Don’t Do) in Network
Automation

Abstract:

Large Language Models (LLMs) are increasingly explored for automating network operations and management — from translating intents to generating device configurations. But how well do they actually perform? In this talk, I share insights from a recent survey at the intersection of LLMs and computer networking. Rather than covering every application, I focus on one of the most active areas: network automation. I will walk through representative use cases, highlight current limitations, and discuss what is still missing. The talk concludes with a short research roadmap and open questions for making LLM-based automation more trustworthy and practical in real-world networks.

Speaker:

Fan Liu is a second-year Ph.D. student in Computer Science at Washington University in St. Louis. Her research focuses on applying artificial intelligence to computer network operations and management, with an emphasis on large language models (LLMs). She is the first author of a recent survey on LLM applications in network operations, currently available as a preprint. Fan is also exploring LLM-driven approaches to automate incident response workflows. Her broader interests include using AI to enhance the infrastructure and applications of next-generation networks.