Explore a 32-minute lecture from the Computational Genomics Summer Institute (CGSI) 2023 on dividing and conquering symbolic models from black box systems for more efficient domain adaptation. Delve into the importance of model explanation, general design approaches, pros and cons, desirable properties, and the general idea and setup. Examine applications in medical imaging through a toy example, comparing performance, transferability, and fine-tuning. Gain insights from related research papers to enhance understanding of the topic.
Divide and Conquer: Carving Out Symbolic Models from BlackBox for More Efficient Domain Adaptation
Computational Genomics Summer Institute CGSI via YouTube
Overview
Syllabus
Introduction
Why do we need model explanation
General design approaches
Pros and cons
desirable properties
general idea
general setup
Medical Imaging
Toy Example
Performance Comparison
Transferability
Fine Tuning
Performance
Conclusion
Taught by
Computational Genomics Summer Institute CGSI
