Research Activities
Our research addressed an important challenge in pancreatic cancer diagnosis and treatment planning, where precise tumor segmentation near vascular structures significantly impacts surgical decision-making. The framework we developed provides more reliable uncertainty estimates near critical vessel interfaces while maintaining efficiency in less critical regions.
I am pleased to report that our paper, "Spatially-Aware Conformal Prediction for Medical Image Segmentation," has been accepted for publication at the Medical Imaging with Deep Learning (MIDL) conference. This publication represents a direct outcome of the research conducted during my visit and would not have been possible without the expertise and mentorship of Prof. Samavi and Hamed Karimi.
Christmas drinks with my supervisor and the students from the lecture “Secure Machine Learning”
Educational Development
In addition to my research activities, I had the valuable opportunity to attend Prof. Samavi's course on Secure Machine Learning at TMU. This experience complemented my research work and provided broader context for trustworthy AI methodologies.
Participating in the course from a student perspective offered fresh insights and allowed me to engage with concepts in a different way than I typically do as a researcher. The course material strengthened my understanding of the theoretical foundations underlying my research project.
Benefits and Impact
This research visit has significantly contributed to my professional development in several ways:
- Acquisition of new technical skills in conformal prediction and uncertainty quantification
- Publication success with a paper accepted at a recognized conference in the field
- Establishment of international research connections that may lead to future collaborations
- Development of expertise at the intersection of statistical methods and medical imaging
- Practical implementation experience with statistical guarantees for AI models
The knowledge gained during this visit has already begun to influence my approach to uncertainty quantification in ongoing projects at Amsterdam UMC, providing valuable new methodological perspectives.
I am grateful to the Cancer Center Amsterdam for supporting this research exchange through the CCA travel grant. The experience has been professionally enriching and has opened new research directions that will benefit my work going forward.
The Department of Electrical, Computer, and Biomedical Engineering and Toronto Metropolitan University (formerly Ryerson).
Research Visit: Trustworthy AI Lab, Toronto Metropolitan University
September - December 2024