The APAN AI driven Networks Working Group (AINWG) has started at the APAN 50 meeting held in 2020 by having BoF (Birds of Feather) sessions. The aim of AINWG is to share experience to design, manage, maintain, and protect the network using artificial intelligence (AI) and escalate collaboration for AI driven networks in the APAN community.
Combining human intellect and creativity with the massive computing power of AI will create situations in which new design and management techniques may be created that humans could not build on their own, but self-improving intelligent algorithms will harness over time.
All activity of AINWG is to encourage the collaboration of technical experiences and knowledge regarding AI-enabled networks. This will lead to the development of an intelligent, scalable, sustainable, and easy-to-deploy technical platform for each APAN member country to manage its respective national research and education networks.
- Chin Guok <chin@es.net> from ESnet (USA)
Title:ESnet Data and AI Workshop Outcomes
Synopsis: The objective of the Data and AI workshop is to identify challenges within ESnet that can be addressed through data-driven methods. This will help define ESnet’s data analysis requirements and shape its AI strategy, guiding data stewardship efforts, and the direction of AI research and AIOps exploration for ESnet7. This presentation will provide a summary of the outcomes from the workshop.
- Syed Saqib Jamal <saqib132413@gmail.com> from Jeju National University (Korea)
Title: LLM-Based IBN For Resource Management in SAGIN: Controller Placement, Slicing, and Mobility Solutions
Summary: A novel Large Language Model (LLM)-based Intent-Based Networking (IBN) framework is introduced to streamline resource orchestration and automate network slicing. By translating high-level intents into precise network configurations, this framework enhances the efficiency and agility of resource management. The controller placement and the satellite mobility issues are also introduced.
- Sirawich Pongraweewongsa <sirawich.pp2002@gmail.com> from Chulalongkorn University (Thailand) -virtual
Title: Development of mobile wireless software-defined network security for abnormal data detection in cyber-attack analysis using machine learning
Synopsis: This talk highlights ongoing activities on an innovative approach to network security testing and DDoS mitigation by integrating Mininet-WiFi, SDN, Docker, and Machine Learning into a cybersecurity testbed. The ability to simulate attacks, detect anomalies, and mitigate threats in real time enhances proactive defense mechanisms. The outcomes of this research are expected to contribute to the development of next-generation security solutions for modern mobile wireless network infrastructures.
- Prof. Muhammad Ibrahim <ibrahimislamian@gmail.com> from University of Haripur (Pakistan) -virtual
Title: Privacy in LLMs: Advances in Differential Privacy, Federated Learning, and Beyond
Synopsis: Large Language Models (LLMs) have revolutionized natural language processing but raise significant privacy concerns due to their reliance on vast amounts of user data. We will explore recent advancements in privacy-preserving techniques, focusing on Differential Privacy (DP) and Federated Learning (FL) as key solutions. DP ensures formal privacy guarantees by introducing controlled noise, while FL enables decentralized model training without exposing raw data. We review recent studies addressing data leakage risks, model inversion attacks, and adversarial threats, highlighting emerging hybrid approaches that combine DP and FL for enhanced security. We will further discusses regulatory compliance, challenges, and future directions for ensuring robust privacy in LLMs.
Open Zoom
