About DeepGI Lab
Since 2022, under the leadership of Partho Ghose, the DeepGI: Deep General Intelligence
Lab has been pursuing cutting-edge research in deep learning and blockchain to build
secure, scalable systems for democratic participation and critical infrastructure. We
develop hybrid blockchain architectures, AI-driven voting protected by post-quantum
security, and machine-learning tools for fraud detection and resource allocation.
Core Research Areas
- Post-Quantum Cryptography
- Blockchain E-Voting Systems
- Medical AI and Computer Vision
- Deep Learning for Healthcare
- Federated Learning
- Multi-Agent Systems
- Decentralized Networks
The Lab Team
Partho Ghose - Lab Head
Graduate Research Assistant at Texas A&M University (Biological & Agricultural
Engineering). Former Lecturer at BUBT. Focuses on AI in Healthcare and Agriculture.
Affiliation: Texas A&M University (Ph.D. Candidate)
Skills: Deep Learning, Medical Imaging, Computer Vision, Machine Learning
Sohel Ahmed Joni - Research Associate
Specializes in secure, decentralized systems. Lead author on HAC-Bchain and Grainbee.
Researcher in post-quantum cryptography and consensus mechanisms.
Affiliation: Bangladesh University of Business and Technology
Skills: Blockchain, Post-Quantum Cryptography, Decentralized Systems, LLM,
Reinforcement Learning
Nishat Tasnin - Research Associate
Award-winning researcher (Best Paper, IEEE COMPAS 2024). Focuses on data sharding
strategies and deep learning optimization for voting and healthcare.
Affiliation: Bangladesh University of Business and Technology
Skills: Data Analysis, Machine Learning, Python, Feature Engineering
Rabiul Rahat - Research Associate
Researcher in blockchain security and post-quantum cryptography. Key contributor to
HAC-Bchain, Grainbee, and post-quantum e-voting systems.
Affiliation: Bangladesh University of Business and Technology
Skills: Blockchain, Cryptography, Quantum Security, NLP, Federated Learning
Selected Publications
A blockchain model for ensuring privacy, trust, and dependability of electronic voting
systems
Authors: Sohel Ahmed Joni, Rabiul Rahat, Nishat Tasnin, Partho Ghose, Hasan Jamil
Published in: Knowledge and Information Systems, 2026
DOI: 10.1007/s10115-026-02693-6
Towards secure democracy: a hybrid blockchain-enabled secure and scalable e-voting
system with sharding and post-quantum cryptography
Authors: Sohel Ahmed Joni, Rabiul Rahat, Nishat Tasnin, Partho Ghose
Published in: International Journal of System Assurance Engineering and Management,
2025
DOI: 10.1007/s13198-025-02927-w
Grainbee: A Quantum-Resistant Blockchain-Based Ration Distribution System with
Hardware Security Modules
Authors: Sohel Ahmed Joni, Rabiul Rahat, Nishat Tasnin, Partho Ghose, M. Rashid
Published in: IEEE COMPAS 2024 (Best Paper Award)
DOI: 10.1109/COMPAS60761.2024.10796594
Hybrid-Blockchain-Based Electronic Voting Machine System Embedded with Deepface,
Sharding, and Post-Quantum Techniques
Authors: Sohel Ahmed Joni, Rabiul Rahat, Nishat Tasnin, Partho Ghose, MA Uddin, J
Ayoade
Published in: MDPI Blockchains, 2024
DOI: 10.3390/blockchains2040017
HAC-Bchain: A Secure and Scalable Blockchain-Shard Based E-Voting System
Authors: Sohel Ahmed Joni, Rabiul Rahat, Nishat Tasnin, Partho Ghose, L Gaur
Published in: IEEE TEMSCON-ASPAC, 2023
DOI: 10.1109/TEMSCON-ASPAC59527.2023.10531344
An Effective Combination of Deep and Machine Learning Models for Monkeypox Detection
Authors: Partho Ghose, M Biswas, Sohel Ahmed Joni, Rabiul Rahat, Nishat Tasnin
Published in: TEHI 2024
DOI: 10.1007/978-981-97-3937-0_3
MultKAN-Nash: Strategic Multi-Agent Disaster Response using MultKAN and Nash
Equilibriums
Authors: Sohel Ahmed Joni, Nishat Tasnin, Sorna Das, Ahmed Shafkat, Samin Yasar, Bijon
Mallik
Published in: Engrxiv Preprint, 2025
DOI: 10.31224/6079
Research Projects
HAC-Bchain E-Voting
A secure and scalable blockchain-shard based e-voting system. Implements Hierarchical
Authoritative Consensus (HAC) and ZK-proofs to balance transparency with voter
privacy.
Technologies: Go, Flutter, LevelDB, ZK-Proofs, Gin
Status: Published
Grainbee
A Quantum-Resistant Blockchain-Based Ration Distribution System with Hardware Security
Modules. Designed to prevent corruption in food subsidy programs.
Technologies: Hyperledger Fabric, Flutter, Go, HSM
Status: Published
Multi-Spectral RTR
Hybrid ConvNeXt–Swin Transformer for Multi-Band Radar Cross-Section Classification
with 92% validation accuracy.
Technologies: PyTorch, Swin Transformer, ConvNeXt, HFSS
Status: Ongoing
Monkeypox Detection System
Annotated 2,000+ dermatographic images. Implemented transfer learning pipeline with
ResNet-50, VGG-16, and MobileNet for high-accuracy disease classification.
Technologies: PyTorch, CNN, Transfer Learning
Status: Published
MultKAN-Nash
A strategic multi-agent disaster response system leveraging MultKAN architecture and
Nash Equilibrium game theory for optimal resource allocation during emergencies.
Technologies: Python, Multi-Agent Systems, Game Theory, KAN
Status: Published
MediKAN
AI framework integrating LLMs with Kolmogorov-Arnold Networks (KANs) for medical
guidance and personalized physician recommendations. Achieves 96% accuracy in
physician quality prediction.
Technologies: Python, KAN, LLM, RAG, GEMMA
Status: Ongoing