Why explainability matters more once a model reaches patients
Notes on what changed once PolyGuard moved from prototype to something clinicians could question.
Sign in to read →Research Initiative — Dr. Anthony Ojo
Advancing research in computing, artificial intelligence, digital education, and process automation.
Marvel Lab is an initiative by Dr. Anthony Ojo dedicated to conducting impactful research, developing innovative digital solutions, promoting academic collaboration, and advancing technology that solves real-world challenges.
About
Marvel Lab bridges academic research with practical solutions across computing, artificial intelligence, digital education, and process automation.
Foundational work in systems, algorithms, and applied computer science.
Explainable, human‑centred AI applied to real‑world decision problems.
Tools and frameworks that make learning more adaptive and accessible.
Automation systems that reduce friction across research and industry workflows.
Focus Areas & Research Works
A look at the areas we work in and a sample of finished projects.
Recently Finished
An AI-driven system that flags risky drug combinations and explains why, per patient.
View DetailsA recommendation engine that reorders coursework based on live learner performance.
View DetailsPublications
Journal of Health Informatics · 2025 · DOI: 10.0000/mlab.2025.001
Intl. Conf. on Digital Learning · 2024 · DOI: 10.0000/mlab.2024.014
Blog
Notes on what changed once PolyGuard moved from prototype to something clinicians could question.
Sign in to read →What we learned building for intermittent internet instead of designing around it.
Sign in to read →Events
University Campus, Main Auditorium
A day of talks from the lab and invited collaborators across AI and digital education.
DetailsComputing Building, Room 214
A hands-on session on SHAP-based explanations for tree-based models.
DetailsTools
A per-patient explainability viewer for polypharmacy risk scores.
Sign in to launchResearchers, institutions, industry partners, and students are welcome to collaborate with Marvel Lab.
Collaborate With Marvel Lab