Breaking Out of the Black Box
Whatever decisions are being taken should be explainable and trustworthy
Focusing on healthcare applications where explainability is not just important - it's essential for saving lives
Find my research profiles:
Vision & Direction
Research Vision
Breaking out of the black box to make AI understandable, interpretable, and beneficial for humanity
Why It Matters
In healthcare, AI decisions can mean life or death. When a model predicts stroke risk or diagnoses a disease, doctors and patients need to understand WHY—not just trust a black box. My research ensures AI systems are transparent, trustworthy, and accountable.
Primary Research Areas
My work spans three interconnected domains, all focused on making AI understandable and beneficial for humanity
Healthcare AI
Multimodal AI
Explainable AI (XAI)
Current Research Focus
Primary Focus
Stroke-related problems using AI/ML approaches
Research Domains
Methodology: Methodological research with emphasis on explainability
Research Timeline
A compact view of how the research agenda has evolved across grants, projects, and published outputs.
VC's Research Fund
Awarded grant for "Unveiling the Linguistic Diversity of Bangla" (No. VCRF-SETS:24-013).
Related outputs
- Conceptual Design and Evaluation Plan of a Mobile Relational Agent for Dengue Management in Bangladesh — 19th International Conference on Design Science Research in Information Systems and Technology (DESRIST), University West, Trollhättan, Sweden (Paper 66)
- Relational Agent-Enabled mHealth Platform for Addressing Dengue Crisis in Bangladesh — 2024 IEEE Engineering in Medicine & Biology Society (EMBC), Annual International Conference (Accepted for poster presentation)
Multimodal Healthcare Diagnostics
Initiated research combining diverse data modalities for complete clinical analysis.
Related outputs
- Multiclass Classification for GvHD Prognosis Prior to Allogeneic Stem Cell Transplantation — 36th Australasian Joint Conference on Artificial Intelligence (AJCAI), AI 2023: Advances in Artificial Intelligence, Perth, WA, Australia. Lecture Notes in Computer Science, vol 14430. Springer, Cham
- Liver Disease Classification by Pruning Data Dependency Utilizing Ensemble Learning Based Feature Selection — 36th Australasian Joint Conference on Artificial Intelligence (AJCAI), AI 2023: Advances in Artificial Intelligence, Perth, WA, Australia. Lecture Notes in Computer Science, vol 14430. Springer, Cham
Stroke Analysis using Explainable AI
Commenced predictive analytics research on transparent, interpretable machine learning models for early stroke detection.
Featured Research Projects
In-depth explorations of AI applications in healthcare and beyond
AI-Powered Stroke Analysis and Prediction
Explainable AI for critical healthcare decisions
Developing interpretable machine learning models for stroke-related problems, focusing on early detection and outcome prediction.
Detailed Information Coming Soon
Comprehensive project details including methodologies, results, and code repositories will be added here.
Impact: Improving stroke diagnosis and treatment through explainable AI
Multimodal AI for Healthcare Diagnostics
Integrating diverse data sources for complete analysis
Research on combining multiple data modalities for enhanced healthcare diagnostics.
Detailed Information Coming Soon
Comprehensive project details including methodologies, results, and code repositories will be added here.
Impact: To be detailed
Remote Sensing and Signal Processing Applications
Advanced signal analysis for real-world applications
Applying signal processing and remote sensing techniques to solve practical problems.
Detailed Information Coming Soon
Comprehensive project details including methodologies, results, and code repositories will be added here.
Impact: To be detailed
Open Source Contributions
Building tools for the research community
SortyPy
SearchyPy
Looking Ahead
Long-term Vision
Become an expert in the field working towards innovations that benefit humanity
Impact Goal
Making AI decisions in critical domains transparent and trustworthy