Event schedules / ICONIP 2026

SS-03: Scaling Intelligence: From Federated Learning to Agentic AI

Monday, November 23, 2026, – (Australia/Sydney)
Clarendon Room B · Special Session

Scaling Intelligence: From Federated Learning to Agentic AI

This session focuses on the convergence of distributed and agentic AI systems, addressing the challenges of designing and coordinating intelligent systems at scale. It encompasses architectures for federated learning, multi-agent systems, and edge computing, alongside agentic systems capable of autonomous goal-directed behavior and long-horizon planning. The session aims to bridge theoretical advances in neural information processing with practical deployment in real-world distributed pipelines, exploring critical issues such as communication efficiency, privacy-preservation, and AI safety. Featured invited speakers include Prof. Hai Dong (RMIT University) and Prof. Amin Beheshti (Macquarie University).

Paper 1207 - From Biomarkers to Reports: Evaluating Open Large Language Models for Structured Neuroimaging Report Generation
Authors: Ghada Abdulsalam, Moataz Ahmed and Mufti Mahmud

Paper 1217 - A Weakly Supervised Knowledge Distillation Technique For Efficient Lung Nodule Segmentation
Authors: Mohammad Asifur Rahim, Muhammad Awais and Mufti Mahmud

Paper 1237 - NeuroSoma: A Brain-Inspired XAI Paradigm for Interpretable Classification and Regression Analysis Beyond SHAP and LIME
Authors: Shahriar Siddique Ayon, Md. Saef Ullah Miah, M. Mostafizur Rahman and Mufti Mahmud