Talk

Breaking the Static Barrier: Mastering Continuous Learning in AI Research

3 min read
Poster for the IEEE Computer Society Malaysia Chapter webinar, Breaking the Static Barrier: Mastering Continuous Learning in AI Research, 17 December 2025

The IEEE Computer Society Malaysia Chapter, with the Centre of Research for Cyber Security & Network (CSNET), Universiti Malaya, held this session of the 2025 IEEE Computer Society Malaysia Chapter Webinar Series & Professional Workshops, attended by 44 participants.

Session Details

Synopsis

The limitations of static, train-then-deploy models are increasingly apparent in modern AI research. Frozen after training, conventional models struggle with concept drift and fail to adapt to real-world data.

This technical workshop addressed Continuous Learning (CL) in research applications, with a practical demonstration of CL methods on live network traffic data. It covered the stability-plasticity dilemma, balancing the ability to learn new information without overwriting earlier knowledge, and architectures that overcome catastrophic forgetting to enable truly incremental learning.

Speaker

Afif Haziq bin Haris, Researcher at CSNET, Universiti Malaya. He develops lightweight deep learning models for network traffic classification on edge hardware and secure continual learning intrusion detection systems, drawing on industry experience as a Network Operation Centre engineer.

Moderator

Prof. Ts. Dr. Nor Badrul Anuar Bin Juma'at, Director, Centre of Research for Cyber Security and Network, Universiti Malaya.

Learning Objectives

  1. Understand the differences between static training and Continuous Learning workflows.
  2. Analyse the stability-plasticity dilemma and catastrophic forgetting in research datasets.
  3. Apply strategies and architectures that enable lifelong learning in experimental AI models.
  4. Integrate adaptive AI techniques into ongoing network and security research.

← Back to all news