Welcome to the Causal AI Lab @ UMBC
The Causal AI Lab conducts research at the intersection of causal inference and artificial intelligence to address real-world problems in healthcare, climate science, mobility, and other fields. Led by Dr. Md Osman Gani, we aim to develop trustworthy and explainable AI systems grounded in rigorous causal principles. Our lab is dedicated to developing next-generation artificial intelligence systems that reason about cause and effect. We believe that unlocking causal understanding is critical for creating AI that is not only accurate but also trustworthy, equitable, and actionable in real-world settings.
Our research is guided by two core principles:
-
From Data to Decisions: We go beyond traditional machine learning by uncovering the underlying mechanisms that drive observable phenomena, enabling stronger generalization, better decision-making, and more effective interventions.
-
Real-World Impact: We apply causal AI to domains where better decisions can save lives, improve well-being, and support vulnerable populations, including healthcare, climate adaptation, and accessible mobility for people with disabilities.
We combine data-driven approaches with expert knowledge, drawing from fields like statistics, computer science, epidemiology, and human-centered computing to develop tools that are as rigorous as they are relevant.
Whether it’s modeling the impact of health policies, forecasting climate risks, or navigating the physical world through accessible routing systems, our work bridges fundamental causal theory with practical deployment.
Research Focus Areas
Our Focus
-
Causal Discovery & Inference: Scalable, knowledge-guided methods for learning causal relationships from observational and time-series data.
-
Human-Centered AI: Designing AI systems that are interpretable and aligned with societal needs.
-
Real-World Applications: Deploying causal AI in domains such as healthcare, climate adaptation, and accessible transportation.
Current Highlights
-
MyPath: AI for Accessible Navigation – A personalized routing system for wheelchair users that integrates multimodal data and AI models.
-
Causal Modeling for Climate Resilience – Modeling supraglacial lake evolution in Greenland using time-series causal analysis.
-
Policy Evaluation in Healthcare – Estimating treatment effects from observational data to support evidence-based healthcare decisions.
Collaborate With Us
We collaborate with interdisciplinary researchers, disability rights organizations, healthcare providers, and policymakers to ensure our work is impactful and inclusive.
Collaborators
Sponsors