This course aims to teach learners how to develop generalist medical imaging AI models using multimodal fusion techniques and self-supervised learning. The course focuses on overcoming the limitations of supervised learning, such as the need for large-scale labeled datasets and the lack of consideration for clinical context in medical imaging models. The teaching method includes discussing prototyping methods using a cohort of pulmonary embolism patients, exploring different fusion types, and explaining major types of self-supervised methods for images. The intended audience for this course includes individuals interested in medical imaging, AI, and self-supervised learning techniques.
Towards Generalist Imaging Using Multimodal Self-Supervised Learning - Mars Huang
Stanford University via YouTube
Overview
Syllabus
Intro
Increase of Medical Imaging Utilization Can Hurt Patient
Limitation 1: Supervised learning requires large sc labeled datasets
Limitation 2: Few Medical Imaging Models Consider Clinical Context
Prototyping Methods Using Cohort of Pulmonary Embolism Patients
Specific Aims
Challenges For Pulmonary Embolism Detection
PENet
Fusion Types
Major types of self-supervised method for images
Learning global representations can be limiting
Global & Local Representations for Images using Attention G
Representation Learning Objective
Retrieval Results
Fine-tune Classification
Strategies for Generating Class Prompts
Zero-shot Classification Results
Next Steps
Generalizability to Other Downstream Tasks
Demonstrate feasibility of applying the propose to other imaging modalities and patient cohort
Taught by
Stanford MedAI
