Large scale machine learning projects for medical imaging and natural language processing in Pathology
Iain Carmichael, Professor
Statistics
Closed. This professor is continuing with Fall 2024 apprentices on this project; no new apprentices needed for Spring 2025.
Work with UCSF on large scale machine learning projects for medical imaging and text processing in Pathology. The ultimate aim of this collaboration is to develop clinically impactful deep learning algorithms for disease diagnosis/prognosis using massive (e.g. 100,000x100,000 pixel) whole slide images (https://www.pixelscientia.com/article-finding-prognostic-patterns-in-gigapixel-images.html) that have only recently become available at scale through UCSF's impressive digital pathology operation. These projects will involve close collaboration with an interdisciplinary team of statisticians, computer scientists, and clinicians.
Role: Possible tasks may include working with our team to
Build deep learning pipelines for massive scale supervised learning model for clinical imaging data
Train “foundational models” for histopathology data with self-supervised learning approaches
Construct new deep learning architectures that handle the size and complexity of histopathology images
Develop deep learning approaches for semantic segmentation of cancerous tissue
Develop interpretability algorithms to understand how our deep learning models are working and uncover new insights into cancer biology
Contribute to software packages
Write manuscripts
Qualifications: Previous deep learning experience is necessary for most (though not necessarily all) projects.
Experience in computer vision is appreciated.
Proficiency in Python (including standard scientific python libraries like numpy, matplotlib, etc) is required. Previous software engineering experience in industry is desired (e.g. internships).
Working knowledge of machine learning at the advanced undergraduate course level is required.
Hours: to be negotiated
Off-Campus Research Site: Remote
Related website: https://idc9.github.io/group.html
Digital Humanities and Data Science Mathematical and Physical Sciences