Tanzim Mahfuz, Swarup Bhunia & Prabuddha Chakraborty

This work introduces an Explainable AI (XAI)-guided Design-for-Security framework, where we demonstrate the applicability of transfer learning in countering specifically reverse engineering and route for a broad spectrum of microelectronics threats.

Collaborators: University of Florida

Tanzim Mahfuz, Pravin Gaikwad, Tasneem Suha, Swarup Bhunia & Prabuddha Chakraborty

This framework integrates Explainable AI (XAI) with Graph Neural Networks to enable dynamic post-processing for hardware Trojan detection in gate-level circuits.

Collaborators: University of Florida

Niroop Sugunaraj, Shree Ram Abayankar Balaji, Barathwaja Subash Chanda, Prashanth Rajagopalan,  Utku Kose, David Charles Loper, Tanzim Mahfuz, Prabuddha Chakraborty, Seerin Ahmad, Taesic Kim, Giovanni Apruzzese, 

Anamika Dubey , Luka Strezoski , Benjamin Blakely , Subhojit Ghosh , Maddikara Jaya Bharata Reddy, Harsha Vardhan Padullaparti & Prakash Ranganathan

The survey systematizes DERMS cybersecurity by mapping realistic threat scenarios such as: false-data injection, DoS/DDoS, malware, and identity/privilege misuse onto the IEEE 1547.3 multi-level hierarchical framework, spanning architectures, assets, and communication protocols end-to-end. 

Collaborators: University of North Dakota, Argonne National Laboratory, National Renewable Energy Laboratory, Suleyman Demirel University, Texas A&M University-Kingsville, University of Missouri, University of Liechtenstein, Washington State University, University of Novi Sad, National Institute of Technology Raipur, National Institute of Technology Tiruchirapalli

Md Hafizur Rahman, Zafaryab Haider, Md Mashfiq Rizvee, Sumaiya Shomaji & Prabuddha Chakraborty 

ILASH is a Neural Architecture Search framework designed to search network for multi-task applications. Its predictive approach reduces the overall searching time.

Collaborators: University of Kansas

Md Mashfiq Rizvee, Md Hafizur Rahman, Prabuddha Chakraborty & Sumaiya Shomaji 

This paper investigates steps necessary to create AI systems that are both environmentally sustainable and cybersecurity-aware, spanning neural architecture, energy-efficient edge processing, and lifecycle-based security strategies.

Collaborators: University of Kansas

Prabuddha Chakraborty,  Reiner N. Dizon-Paradis & Swarup Bhunia

IoT‑based dynamic charging framework that enables on‑the‑move replenishment of battery electric vehicles (BEVs) using unmanned aerial vehicles (UAVs) and mobile charging stations (MoCS), eliminating the need for prolonged charging stops.

Collaborators: University of Florida

Prabuddha Chakraborty,  Jonathan Cruz, Rasheed Almawzan, Tanzim Mahfuz & Swarup Bhunia

This paper demonstrates how learning-based (e.g., ML or GNN) attacks can exploit structural artifacts in logic locking schemes to recover the secret key, highlighting a critical gap in state of the art defense techniques.

Collaborators: University of Florida

Zafaryab Haider, Md Hafizur Rahman, Vijay Devabhaktuni, Shane Moeykens & Prabuddha Chakraborty 

COBRA is a security-focused framework designed to protect AI models during training, specifically when the models are being taught using RLHF.

Collaborators: Illinois State University

Tasneem Suha, Md Hafizur Rahman, Ashley E. Rice, Colin Smith, Rima Asmar Awad & Prabuddha Chakraborty

TRIM is a generative AI-based, robust random number generator solution that falls between PRNG and TRNG.
Collaborators: Oak Ridge National Laboratory

Nickolas Millett, Katherine Arsenault, Nick Tozier, Benjamin Bailey, Gregory Studer, and Prabuddha Chakraborty

CAST-Map is framework designed to track digital systems in real-time, with clustering based scoring system based on similar systems.

Collaborators: Advanced Structures & Composites Center, Maine

Tanzim Mahfuz, Sudipta Paria, Tasneem Suha, Swarup Bhunia & Prabuddha Chakraborty

An Explainable AI (XAI)-driven power side-channel mitigation framework, offering high parameterization and full customization to meet user-specific requirements.

Collaborators: University of Florida

Prabuddha Chakraborty, Tasneem Suha & Swarup Bhunia 

This work presents a timing side-channel security analysis that incorporates awareness of hardware specifications, e.g. microarchitectural behaviors to more accurately detect vulnerabilities that software-only analyses may fail to capture.

Collaborators: University of Florida

Abdullah Al Nomaan Nafi, Habibur Rahaman, Zafaryab Haider, Tanzim Mahfuz, Fnu Suya, Swarup Bhunia & Prabuddha Chakraborty

DAASH is a meta-attack framework for synthesizing effective and stealthy adversarial examples.

Collaborators: University of Florida, University of Tennessee (Knoxville)

Md Hafizur Rahman, Zafaryab Haider & Prabuddha Chakraborty 

LEMONADE is a Neural Architecture Search framework designed to automatically generate architecture with user preferences.

Sarah Glatter & Prabuddha Chakraborty 

DINGO is a distributed and graph based memory database system for Internet of Things and edge devices. 

Prabuddha Chakraborty & Swarup Bhunia 

A self-aware digital memory for IoT/edge devices that monitors stored data and query patterns, using continual reinforcement learning to adapt memory organization (associations/granularity) on the fly.

Collaborators: University of Florida

Adeeb Ibne Alam, Md Hafizur Rahman, Akhter Zia, Nate Lowry, Prabuddha Chakraborty, Md Rafiul Hassan & Bashir Khoda

The paper proposes a modular in-situ pipeline for heterogeneous particle–substrate images that performs image enhancement and sharpening, uses a MobileNet-based transfer-learning selector to choose the best detector.

Collaborators: University of Maine at Presque Isle

Aritra Dasgupta, Sudipta Paria, Prabuddha Chakraborty, & Swarup Bhunia

The paper proposes a cooperative trust model for SoCs with untrusted third-party IPs that “splits” sensitive secrets across multiple independent IPs.

Collaborators: University of Florida

Prabuddha Chakraborty & Swarup Bhunia

The paper presents an intelligent, application-aware memory framework for resource-constrained IoT devices that dynamically reorganizes stored content.

Collaborators: University of Florida