Arun Kumar Dey

About

My research is centered on extracting implicit knowledge from IoT-enabled network traffic and applying Machine Learning (ML) and Deep Learning (DL) techniques to address challenges in feature selection, cyber threat detection, and mitigation. My current research interests span ML, DL, IoT Analytics, Federated Learning, Explainable AI, GNS3, and Quantum Computing, all within the context of cybersecurity.

From my perspective, cybersecurity is an inherently interdisciplinary field that draws upon computer science, information technology, artificial intelligence, and network systems. This convergence of disciplines is what makes it both demanding and intellectually stimulating. A cyber security researcher often serves as a bridge between professionals from different domains, collectively working to resolve complex security challenges in modern digital ecosystems, particularly those involving the Internet of Things (IoT) and distributed computing environments. My aspiration is to operate at this multidisciplinary crossroads, engaging with researchers from varied technical backgrounds to design intelligent and secure cyber systems.

A significant strand of my current research addresses privacy-preserving collaborative learning for cybersecurity applications. I have built federated learning-based frameworks tailored for distributed cyber threat detection, integrating secure aggregation protocols, quantum-safe mechanisms, and communication-efficient architectures suited to IoT environments. Additionally, I have developed federated training and evaluation pipelines using the Flower (flwr) framework to facilitate distributed threat detection across heterogeneous IoT datasets. To complement this, I have used VMware, GNS3, and Kali Linux to build multi-VM attack simulation environments to generate labeled cyber-attack datasets. A core objective of this work is to deliver practical, real-time solutions deployable on resource-constrained edge devices, such as the Raspberry Pi 4/5 and NVIDIA Jetson Nano, while upholding strong detection performance, computational efficiency, and privacy guarantees.

Alongside these efforts, my doctoral research at the Department of Information Technology, National Institute of Technology Raipur, Chhattisgarh, India, focuses on developing advanced cyber threat detection models through hybrid deep learning, ensemble learning, and metaheuristic-driven feature optimization. Among my key contributions, I have proposed Explainable AI-based detection models, Bayesian reinforcement learning frameworks, and ensemble feature selection strategies to improve both detection accuracy and model interpretability. These research outputs have been accepted for publication in high-impact SCI-indexed journals, including ACM Transactions on Internet Technology and other Q1-ranked venues.

Employment

• National Institute of Technology Raipur PhD Student
  2021 - Present

Education

Education history is unavailable.

Projects & Funding

Projects & funding information is unavailable.

Publications

Publications are unavailable.