April 11, 2025
CS Colloquium Announcement
| Priority: | No |
|---|---|
| School or Program: | Electrical and Computer Engineering |
| College Calendar: | Show |
CS Colloquium Announcement
Colloq announceent for this coming Wednesday. If you have any questions please let me know. Below is the information.
D. Manivannan
Wednesday, April 16th
3pm-4pm
LWSN 3102AB
Host: Bharat Bhargava
Title: Machine Learning-Based Intrusion Detection Systems: Enhancing Performance with Novel Approaches
Abstract: Machine Learning (ML)-based Intrusion Detection Systems (IDS) have gained significant attention in recent research due to their potential to enhance cybersecurity. However, improving the performance of ML classifiers for intrusion detection remains a challenging task, particularly when dealing with imbalanced datasets, optimizing loss functions, and handling limited labeled data. In this talk, I will present three key methods we explored to enhance the effectiveness of ML classifiers in detecting intrusions in computer networks: (i) Balancing Data with Synthetically Generated Samples – Many intrusion detection datasets suffer from class imbalance, where malicious activities are underrepresented compared to normal traffic. To address this, we leverage synthetic data generation techniques to create a more balanced training set, ultimately improving model generalization and detection accuracy. (ii) Optimizing the Loss Function – Selecting an appropriate loss function is crucial for training robust classifiers. We investigate the impact of different loss functions tailored for intrusion detection, ensuring the model effectively learns from both common and rare attack patterns. (ii) Few-Shot Learning for Enhanced Detection – Traditional ML models require large amounts of labeled data, which is often scarce in cybersecurity applications. Few-shot learning techniques enable models to recognize new intrusion patterns with minimal labeled examples, making them more adaptable to emerging threats.
Biography Dr. D. Manivannan is currently an Associate Professor in the Computer Science Department at University of Kentucky, Lexington, Kentucky, USA. He earned his Ph.D in computer and information science, from The Ohio State University, Columbus, Ohio, USA, under Prof. Mukesh Singhal. He published his research work in the following areas: fault-tolerance and synchronization in distributed systems, mobile ad hoc networks, vehicular ad hoc networks, channel allocation in cellular networks, machine learning based approaches for intrusion detection in computer networks and Internet of Things. Dr. Manivannan has more than 95 publications in refereed International Journals (a vast majority of which were published by IEEE, ACM, Elsevier, and Springer) and Proceedings of International Conferences. Dr. Manivannan served as an Associate Editor/Editorial board member of IEEE Transactions on Parallel and Distributed Systems, IEEE Communications Magazine and Wireless Personal Communications journal. Currently, he is on the Editorial Board of Information Sciences journal and Internet of Things Journal. Dr. Manivannan is a recipient of the Faculty Early Career Development award (CAREER award) from US National Science Foundation. He has been a senior member of the IEEE and ACM.
Victoria Renn
Associate administrative assistant, Department of Computer Science
The Richard and Patricia Lawson Computer Science Building , 305 N. University Street , West Lafayette, IN 47906
o: 765-464-6003