Advancing Multimodal AI & Federated Healthcare.
Hello! I’m Arshad Al Galib, a Computer Science & Engineering researcher specializing in Cross-Attention Multimodal Fusion, Privacy-Preserving Federated Learning, and Explainable Healthcare AI.
Core Research Interests
Dedicated research directions bridging foundational machine learning theory with mission-critical applications in healthcare diagnostics and multimodal perception.
Multimodal Deep Learning
Developing representations that jointly model heterogeneous data streams—such as visual phenotypes, clinical tabular indices, and text—to achieve unified contextual understanding.
Healthcare AI
Designing predictive intelligence for early diagnosis, symptom-to-disease inference, readmission risk estimation, and clinical decision support under real-world hospital workflows.
Explainable AI (XAI)
Unpacking black-box neural architectures through feature attribution, calibration analysis, and saliency mappings to ensure accountability and trustworthiness in high-stakes decisions.
Privacy-Preserving Federated Learning
Decentralized model training across heterogeneous clinical nodes without centralizing raw patient records, incorporating differential privacy and non-IID client optimization.
Medical Image Analysis
Investigating spatial co-registration between multi-sequence MRI and CT volumes (BraTS 2020 & RIRE benchmarks), mutual information maximization, and automated lesion localization.
Cross-Attention Multimodal Fusion
Architecting attention-guided gating mechanisms to adaptively prioritize complementary perceptual cues across multiple inputs (such as co-evaluating fruit and leaf phenotypes).
B.Sc. in Computer Science & Engineering
Manarat International University, Dhaka • CGPA: 3.53 / 4.00 (2022 – 2026)
Research Publications
Published across IEEE conferences, open repositories, and international symposia.
The Impact of Generative AI Interaction on Human Metacognition and Critical Thinking
A. A. Galib, A. Uddowla, A. Islam, J. Abbas, A. Bashir, R. Seyam, et al.
Comprehensive human-subject experimental study examining cognitive reliance, bias verification, and critical metacognition under generative AI workflows.
Client-Level Privacy-Preserving Diabetes Classification Using Federated Learning Under Synthetic Small-Client Heterogeneity
A. A. Galib, A. Uddowla, A. Islam, J. Abbas, A. Bashir, R. Seyam, et al.
Addresses non-IID patient data distributions across simulated regional health nodes using differential privacy and federated aggregation.
Grassroots Leadership and Ethical Political Financing in Post-Authoritarian Bangladesh: A Case Study of Sharif Osman bin Hadi and the July 2024 People’s Uprising
A. A. Galib, A. Uddowla, A. Islam, J. Abbas, A. Bashir
Empirical case study investigating grassroots mobilization, open crowdsourced political financing mechanisms, and civic accountability in transitional governance.
Reimagining Justice: A Victim-Centered Framework for Mental Health and Rehabilitation in Post-Conflict Bangladesh
A. A. Galib, A. Uddowla, et al.
Psychosocial rehabilitation model centering mental healthcare delivery frameworks for post-conflict recovery.
Federated Diabetes Readmission Prediction Under Controlled Client Heterogeneity: A Privacy, Personalization, and Calibration Study
A. A. Galib, A. Uddowla, A. Islam, J. Abbas, A. Bashir, R. Seyam, et al.
Intelligent Symptom-to-Disease Prediction System Using Machine Learning for Early Healthcare Diagnosis
A. A. Galib, A. Uddowla, A. Islam, J. Abbas, A. Bashir, R. Seyam, et al.
Comparative Evaluation of Transfer-Learning CNNs for Olive Leaf Disease Classification Under Accuracy and Efficiency Constraints
A. A. Galib, A. Uddowla, A. Islam, J. Abbas, A. Bashir, R. Seyam, et al.
Curated Research Datasets
Contributing standardized, benchmarked datasets to the international scientific community.
Large-Scale Bangladesh Mango Leaf and Fruit Dataset
A multimodal agricultural computer vision dataset covering 20 distinct varieties with dual-angle leaf and fruit imagery across healthy and pathological phenotypes.
10.17632/67g6dmzvdf.1
Access on Mendeley
Mental Health Risk Prediction Dataset
Extensive psychological, academic, and behavioral metrics gathered across undergraduate university students (n = 2,000+).
Generative AI, Metacognition & Critical Thinking
Controlled experimental human-subject dataset assessing critical cognitive monitoring during LLM interaction tasks (n = 91).
July Revolution 2024 Victim and Witness Dataset
Primary source survey dataset recording post-authoritarian transitional justice, injury profiles, and trauma indices (n = 500).
Featured Research Projects
Rigorous experimental setups combining transfer learning, image registration, and cross-task evaluation.
Mango Variety Recognition & Multimodal Fusion
Engineered a dual-branch deep learning framework performing cross-task benchmarking and fusing fruit and leaf imagery. Overcomes seasonal single-phenotype limitations.
Multimodal Medical Image Registration: CT-MRI
Investigated rigid & non-rigid alignment between CT and MRI modalities on the BraTS 2020 and RIRE datasets. Benchmarked SIFT against Mutual Information (MI), boosting MI by 138%.
Deep Learning-Based Jute Pest Identification
Evaluated four modern CNN architectures across 6,460 agricultural pest images spanning 17 insect classes. EfficientNet-B0 yielded 96.3% top-1 accuracy and a 96.0% macro F1-score.
Technical & Research Toolkit
Systematic experimental acumen from mathematical modeling to deep neural network deployment.
Programming
AI & Deep Learning
Libraries & Tools
Research Methods
Professional Experience
Research Intern
Conducting empirical investigations with an independent non-profit research institute on algorithmic governance, human-centered technology, and interdisciplinary data analytics.
Editor
Editing scholarly feature articles, collaborating with academic writers, and curating public-interest analyses on science, education, and national development.
Honors & Scholarships
Leadership & Coordination
Organizer – National Research Seminar & Workshop 2026
Organized by Research and Integrated Thoughts (RIT), Manarat International University.
Program Coordinator – MIU IEEE Student Branch
Led student technical symposiums, research outreach, and IEEE membership activities across engineering cohorts.
Project Showcase Invigilator – Al-Khwarizmi Science Fest
Evaluated secondary and collegiate STEM science innovations and software models.
Workshop Organizer – AI, ML & Career Development
MIU CSE Club: “Future in Code: AI & Machine Learning with Python” and 25-Day Career Bootcamp.
Faculty Referees
Prof. Dr. Mijanur Rahman
Dean, SEST & Head, Dept. of CSE
Manarat International University
head_cse@manarat.ac.bdMd. Zahurul Haque
Assistant Professor, Dept. of CSE
Manarat International University
jahurulhaque@manarat.ac.bdProf. Md. Ershadul H. Choudhury
Professor, Dept. of EEE
Manarat International University
ershadul@manarat.ac.bdOpen for Research Collaboration & Graduate Inquiries
Actively exploring Master’s and Ph.D. research assistant opportunities in Multimodal Deep Learning, Federated Health AI, and Medical Vision.