Research Overview

Our Four Pillars

CIMILab structures its research across four distinct pillars that span from theoretical foundation to clinical translation.

How the lab works

Research that is structured across methods, evaluation, and translation

Each pillar connects a scientific question to reproducible experimentation and practical deployment concerns.

We develop and refine machine learning methods that can support high-stakes prediction and analysis tasks.

Benchmarking, uncertainty analysis, and careful study design help keep our claims meaningful and credible.

We aim to move promising ideas toward useful tools, shared software, and more robust real-world workflows.

Deep learning architectures, ensemble methods, and attention mechanisms for high-stakes prediction and segmentation tasks.

Medical Image Segmentation Ensemble Methods Attention Networks Self-supervised Learning

Data infrastructure, benchmarking frameworks, and evaluation pipelines that make AI experiments reproducible and comparable.

Benchmarking Frameworks Data Engineering Evaluation Pipelines Reproducibility Standards

Uncertainty quantification, interpretability methods, and fairness auditing — building AI that clinicians and patients can trust.

Uncertainty Quantification Explainable AI Fairness & Bias Auditing Calibration

Bridging research and deployment — efficient inference, edge adaptation, and clinical decision support at real-world scale.

Clinical Decision Support Edge AI Foundation Models for Medicine Multimodal Fusion