ASNR ASFNR
AI Workshop
Background
ASNR and ASFNR continue to partner to offer an AI education program to train neuroradiologists and make them effective leaders and collaborators in AI projects focusing on neuroimaging.
The goals of this education program are to:
-
Serve as a hands-on introductory workshop to AI and deep learning in neuroradiology
-
Allow attendees to understand and critically evaluate literature in AI and neuroradiology
-
Allow attendees to set up and conduct a basic research project using AI.
Details
This education program will consist of prerequisite work, two workshops (Saturday April 29 and Sunday April 30 at ASNR23 in Chicago and Friday October 6 at ASFNR 2023 in Boston), and a project to be conducted in between the two workshops. Hybrid access will be available for in-person and virtual attendance at both sessions.
Different datasets will be made available for the participants, including the 2019 ASFNR AI challenge dataset and the 2019 RSNA challenge dataset. Participants are free to use these datasets or to select other datasets that are publicly available or that they have access to.
A total award of $5,000 will be distributed to participants whose projects in the workshop are deemed to have the highest potential for future research in the field of AI applied to neuroradiology. These awards will be presented at the conclusion of the second part of the workshop.
The education program will accept 15 applicants a year, who will graduate at the end of the ASFNR workshop and be invited to join the ASNR/ASFNR AI study group for continued collaboration and networking.
In addition to clinical neuroradiologists and trainees, data scientists and PhD faculty engaged in neuroimaging research are encouraged to apply.
Selection
Applicants will be selected based on a competitive application process that will involve submission of a biosketch, letter of motivation, and potential research project topic, as well as the completion of a questionnaire and some pre-requisite work. The selection process will ensure that a diverse class is created each year, in terms of training levels (fellows, junior faculty, more advanced faculty), institution and geographical representation, as well as level of programming knowledge (a minimum level of programming knowledge will be required).
The application deadline is March 21, 2023, at 11:59 PM EDT.
Applicants will be informed of selection for the workshop by March 31, 2023.
There will be a fee of $350 per attendee. This fee will cover participation at BOTH ASNR23 and ASFNR 2023 workshop sessions.
Program
Organizing Committee
● Daniel Barboriak, Duke
● Peter Chang, UC Irvine
● Risto Filippi, Tufts
● Adam Flanders, Jefferson
● Christopher Hess, UCSF
● Yvonne Lui, NYU
● Max Wintermark, MD Anderson
● Greg Zaharchuk, Stanford
● Rajan Jain, NYU
Speakers/Mentors
We welcome experienced AI researchers to join our team of speakers who also serve as mentors for the accepted participants. Please apply below if you are interested in helping out with the AI workshop.
Sample Program (Schedule Subject to Change)
ASNR23
Block-1
Introduction
30 min
Historic Overview of AI and Machine Learning
30 min
The Role of Neuroradiologists in AI Research
30 min
Survey of Current Applications and Tools
Block-2
30 min
Designing a Neuroradiology AI Project
30 min
Data Curation and Annotation
30 min
Choosing Machine Learning Algorithms
Block-3
30 min
Getting Started: Hardware Considerations
30 min
Getting Started: Software Considerations
30 min
Getting Started: Personnel / Who to Hire
Block-4
30 min
Introduction to Python
30 min
Introduction to Machine Learning Libraries
30 min
Introduction to Artificial Neural Networks
Block-5
30 min
Overview of Convolutional Neural Networks
30 min
Optimization and Hyperparameters
30 min
Common CNN Architectures for Medical Imaging
Block-6
30 min
Towards a Scalable AI Research Infrastructure
30 min
Advanced Data Annotation Strategies
30 min
Ethics of Medical AI
Conclusion and Wrap Up
Sample Program (Schedule Subject to Change)
ASFNR 2023
Block-1
Introduction
30 min
Progress Report
all participants
30 min
Participant Presentations I
all participants
30 min
Participant Presentations II
all participants
Block-2
30 min
Collaborative Research
30 min
Customized CNN Architectures for Medical Imaging
30 min
Opportunities in Natural Language Processing
Block-3
30 min
Manuscripts/Grants: Putting Together A Paper
30 min
Manuscripts/Grants: Understanding Validation and CNN Statistics
30 min
Manuscripts/Grants: Clinical Trials/Funding Opportunities
Block-4
30 min
FDA and Regulatory Considerations
30 min
Fairness in AI
30 min
Clinical Deployment
30 min
Wrap-up