About
I am a computer scientist specialising in cyber security and artificial intelligence, with expertise in deep learning and computer vision. I received my PhD in Computer Science from The University of Queensland (UQ), Australia, and have held postdoctoral research positions at the Australian National University (ANU) and The University of New South Wales (UNSW).
I currently serve as a Sessional Lecturer at the Australian Institute of Higher Education (AIH) and a Sessional Academic at RMIT University, Melbourne. My current research focuses on AI-driven cyber security systems — including threat detection, adversarial decision modelling, and the application of machine learning, deep learning, and federated learning to develop intelligent and privacy-preserving security solutions.
My broader interests include bio-inspired algorithms for autonomous systems and intelligent vision systems. I have extensive teaching experience across undergraduate and postgraduate programs, delivering courses in cyber security, deep learning, computer vision, data science, and software development, with a strong emphasis on hands-on, industry-relevant computing education.
News & Announcements
- Feb 2026 Joined RMIT University as Sessional Academic, School of Computing Technologies, Melbourne — teaching Cyber Security Attack Analysis and Incident Response. New
- Feb 2026 Joined Australian Institute of Higher Education (AIH) as Sessional Lecturer, delivering courses in Data Mining, Database Management, and Programming. New
- Jan 2026 Paper accepted: "A vision transformer-based fine-tuned DINOv2 model for Bangla sign language recognition" — Multimedia Tools and Applications, 85(2), Springer. Journal
- Jan 2026 Paper accepted: "Hybrid adversarial distilled defense for deep learning-based breast cancer detection" — Engineering Research Express, IOP Publishing. Journal
- Dec 2025 Three papers submitted and under review: secure GANs framework (Information Systems), blockchain-secured federated learning (Defence Technology), and moving-obstacle avoidance in budgerigars (Animal Behaviour). Under Review
- Aug 2025 Paper published: "Optimizing federated learning efficiency: FedCCO" — Neural Computing and Applications, 37(20), 15369–15387, Springer. Journal
- Oct 2024 Invited as Resource Person at the International Faculty Development Program (IFDP 2024), Chandigarh University, India. Invited Talk
- Jan 2023 Promoted to Associate Professor at American International University–Bangladesh (AIUB) — went on to supervise 7 Master's and 15 Bachelor's research theses over three years.
Research & Projects
AI-Driven Intelligent Systems for Security, Perception & Autonomous Systems
Intelligent Systemsfor Security, Perception and AutonomyPerceive · Learn · Decide · Act · Protect
My research integrates cyber security, federated learning, computer vision and bio-inspired autonomous systems into intelligent AI frameworks for real-world security, perception and decision-making.
AI-Driven Cyber Security
My work in cyber security centres on the application of artificial intelligence and machine learning to build intelligent, autonomous defence systems. I investigate how AI can be used to model attacker behaviour, identify vulnerabilities, and support proactive threat response — moving beyond rule-based detection towards systems that reason under uncertainty and adapt to evolving adversarial conditions. A key focus is the use of probabilistic and game-theoretic decision frameworks to simulate complex attack scenarios, enabling more realistic penetration testing and strategic cyber defence. This research has been conducted in collaboration with national defence and research organisations, contributing to both theoretical foundations and applied security tooling.
Federated & Privacy-Preserving Learning
This research stream addresses the fundamental tension between collaborative machine learning and individual data privacy. I develop federated learning architectures that enable multiple parties — institutions, organisations, or devices — to train shared models without exposing their underlying data. My work spans both efficiency and security dimensions: designing aggregation strategies that accelerate convergence across heterogeneous participants, and building defence mechanisms that protect federated systems from adversarial manipulation and data poisoning. Underpinning this is an interest in responsible AI — ensuring that intelligent systems can be deployed at scale without compromising the rights and privacy of the individuals whose data shapes them.
Computer Vision & Deep Learning
I develop deep learning systems for real-world visual understanding — from object detection and tracking to semantic segmentation and fine-grained recognition. My research spans the full pipeline: designing novel architectures and loss functions, building scalable inference systems, and deploying vision models in operationally demanding environments such as surveillance, health monitoring, and assistive technology. I work extensively with modern deep learning frameworks and state-of-the-art architectures, including convolutional networks, vision transformers, and self-supervised pre-training paradigms. A recurring theme across this work is robustness — building systems that perform reliably under noisy conditions, class imbalance, and domain shift, with outputs that are interpretable and actionable.
Bio-inspired Algorithms & UAV Systems
Nature has spent millions of years solving problems of navigation, collision avoidance, and autonomous decision-making in complex, dynamic environments. This research programme draws on those solutions — studying the flight behaviour of birds and insects to extract principles that can be translated into computational guidance systems for unmanned aerial vehicles. By combining experimental observation with mathematical modelling and computer vision, I develop algorithms that capture the elegance of biological navigation: lightweight, reactive, and robust without the computational overhead of traditional path planning. The broader goal is to bridge the gap between biological intelligence and engineered autonomy, informing the design of next-generation UAV systems capable of operating safely in unstructured environments.
I have had the privilege of collaborating with leading national and international organisations across defence, aerospace, and mental health research — conducting applied AI and computational research with real-world impact.
Boeing Research & Technology Australia
Investigated bio-inspired collision avoidance strategies for unmanned aerial systems, translating principles from avian flight behaviour into guidance algorithms for UAV navigation. Research conducted in collaboration with the Queensland Brain Institute, University of Queensland.
Defence Science & Technology Group
Conducted cyber security research applying probabilistic decision-theoretic models to penetration testing and strategic threat analysis. Developed adversarial simulation frameworks in collaboration with the Australian National University, with direct application to national cyber defence.
Black Dog Institute, UNSW Sydney
Developed real-time computer vision and behavioural AI systems for detecting indicators of psychological distress in public environments. Worked closely with clinical psychologists to bridge computational methods and evidence-based mental health intervention, contributing to early-warning safety systems.
Publications
-
InformationGainLoss: An edge-aware and class-balanced loss function for robust semantic segmentationImtiaz, A., Nishi, N. I., Hossain, S. K. M., Bhowmik, A., & Karmaker, D.Neurocomputing — in press 2026
-
Harnessing deep learning for plant disease analysis: Current trends, challenges, and future prospectsRahman, H. N., Mamun, R., Nasir, T., Netu, K. N. S., Bhowmik, A., & Karmaker, D.*Agriculture Communications — in press 2026
-
Hybrid adversarial distilled defense for deep learning-based breast cancer detectionDas, S., Noor, F., Raim, T. Y., Das, B., Bhowmik, A., Nandi, D., Rahman, M., & Karmaker, D.Engineering Research Express, doi:10.1088/2631-8695/ae60c9 2026
-
Optimizing federated learning efficiency: Accelerating model averaging with cluster-based approach using class occurrences-FedCCOElahi, A. S. E., Khanam, S., Rahman, N., Rabbi, R. I., Bhowmik, A., & Karmaker, D.Neural Computing and Applications, 37(20), 15369–15387 2025
-
A vision transformer-based fine-tuned DINOv2 model for Bangla sign language recognitionTasnim, S. A., Mahmud, R., Ahmed, T., & Karmaker, D.Multimedia Tools and Applications, 85(2), 163 2026
-
Geospatial dataset on deforestation and urban sprawl in Dhaka, BangladeshKhan, M. F., Islam, M. R., Basak, S. K., Imtiaz, A., Bhowmik, A., Nandi, D., Rahman, M., & Karmaker, D.Data in Brief, 61, 111786 2025
-
Tomato leaf dataset: A dataset for multiclass disease detection and classificationImtiaz, A., Islam Swapnil, F. B., Masud, S. R., & Karmaker, D.Data in Brief, 60, 111520 2025
-
Unlocking educational excellence: Leveraging federated learning for enhanced instructor evaluation and student successIslam, M. A., Karmaker, D., Bhowmik, A., Billah, M. M., Mobin, M. I., & Noor, N. M.International Journal of Modern Education and Computer Science, 2, 87–110 2025
-
Budgerigars adopt robust, but idiosyncratic flight pathsKarmaker, D., Groening, J., Wilson, M., Schiffner, I., & Srinivasan, M. V.Scientific Reports, 10(1), 2535 2020
-
Mimicking nature: Analysis of dragonfly pursuit strategies using LSTM and Kalman filterZidan, M., Ahmed, R., Adnan, K., Mumu, T., Rahman, M., & Karmaker, D.International Journal of Information Technology and Computer Science, 16(4) 2024
-
From nature to UAV: A study on collision avoidance in bee congregationUday, N. H., Hasan, M. Z., Ahmed, R., Rahma, M. M., Bhowmik, A., & Karmaker, D.International Journal of Intelligent Systems and Applications (IJISA), 16(3) 2024
-
Aspect-based sentiment analysis model for evaluating teachers’ performance from students’ feedbackBhowmik, A., Noor, N. M., Miah, M. S. U., & Karmaker, D.AIUB Journal of Science and Engineering (AJSE), 22(3), 8 2023
-
Evaluating teachers’ performance through aspect-based sentiment analysisBhowmik, A., Noor, N. M., Haque, M. Z., Miah, M. S. U., & Karmaker, D.Proceedings of the International Conference for Convergence in Technology, IEEE 2024
-
A comprehensive study of camouflaged object detection using deep learningKhair, K. B., Jahir, S., Ibrahim, M., & Karmaker, D.Proceedings of the International Conference on Robotics, Electrical and Signal Processing Techniques, IEEE 2023
-
Evaluation of deep learning models on UV ink: a fake money detection scheme with RPNChowdhury, A. A., Das, A., Karmaker, D., & Hoque, K. K. S.Proceedings of the International Conference on Computing Advancements, ACM 2022
-
A comparative study of hyperparameter optimization techniques for deep learningChowdhury, A. A., Das, A., Hoque, K. K. S., & Karmaker, D.Proceedings of the International Joint Conference on Advances in Computational Intelligence, Springer 2022
-
An inverse differential game approach to modelling bird mid-air collision avoidance behavioursMolloy, T. L., Garden, G. S., Perez, T., Schiffner, I., Karmaker, D., & Srinivasan, M. V.7th IFAC Workshop on Distributed Estimation and Control in Networked Systems, IFAC-PapersOnLine, 51(15), 754–759 2018
-
Image denoising with weighted orientation-matched filtersKarmaker, D., Schiffner, I., Wilson, M., & Srinivasan, M. V.Proceedings of the IEEE International Conference on Robotics and Biomimetics, IEEE 2018
-
Tracking multiple deformable objects in noisy environments using WORMKarmaker, D., Schiffner, I., Wilson, M., & Srinivasan, M. V.Proceedings of the International Symposium on Visual Computing, Springer 2018
-
WHoG: A weighted HOG-based scheme for bird detection and pose identificationKarmaker, D., Schiffner, I., Strydom, R., & Srinivasan, M. V.Proceedings of the International Conference on Control, Automation, Robotics and Vision, IEEE 2016
-
Cricket shot classification using motion vectorKarmaker, D., Chowdhury, A. Z. M. E., Miah, M. S. U., Imran, M. A., & Rahman, M. H.Proceedings of the International Conference on Computing Technology and Information Management, IEEE 2015
-
Optimized tracking using VAIT for moving objectsKarmaker, D., Miah, M. S. U., Imran, M. A., Rahman, M. H., & Alam, M. Z.Proceedings of the International Conference on Computing Technology and Information Management, IEEE 2015
-
SHIMPG: Simple human interaction with machine using physical gestureRahman, M. A. U., Miah, M. S. U., Fahad, M. A., & Karmaker, D.Proceedings of the International Conference on Control Automation Robotics and Vision, IEEE 2014
-
Stretching time and space: How flying budgerigars evade a moving obstacleKarmaker, D., Groening, J., Schiffner, I., & Srinivasan, M. V.Submitted to Animal Behaviour 2026Under Review
-
Secure generative AI: A multi-player adversarial learning framework for robust GANsBokhtiar, A., Towsif, M. A., Tishad, M. I. A., Paul, B., & Karmaker, D.Submitted to Information Systems 2026Under Review
-
Secure federated learning: Blockchain-coordinated adversarial defense with gradient purificationSadi, M. T. H., Islam, M. R., Sarker, D., Hasan, M. Z., & Karmaker, D.Submitted to Defence Technology 2026Under Review
Teaching
Academic Background
Contact
I welcome enquiries from researchers interested in collaboration on AI, cyber security, computer vision, or federated learning, as well as prospective students, industry partners, and academic colleagues. Please reach out via email — I aim to respond within a few working days.