The role of the AI/ML Task Force will be to provide expertise to the DPA in the following ways:
- The AI/ML TF will develop AI/ML initiatives to cement the DPA as the “go-to” resource for AI/ML in pathology, including regulatory insight, best practices, scholarly activity, vendor relationships, and ethics.
- The AI/ML TF will provide website content oversight, including vetting blog submissions and AI/ML-related content. Given the external facing nature of the website, a member of the AI/ML committee should be designated to attend Website Committee meetings and ideally should be encouraged to become a member of the Website Committee and Blog Vetting group.
- The AI/ML TF will be asked to review concept papers developed by the Education Committee that are relevant to AI/ML.
- The AI/ML TF will contribute expertise to the Regulatory and Standards Task Force in initiatives pertaining to AI/ML.
- The AI/ML TF will be available to assist with other DPA efforts including Pathology Visions planning and speaker recruitment, webinar recruitment, and educational initiatives such as DAPA and PathologyOutlines content.
The AI/ML Task Force's initial charge will be to expand the DPA’s footprint beyond whole-slide imaging into computational pathology. In addition to the above support activities, this will include the following initiatives:
- Lead efforts alongside the Education Committee to develop a white paper on the ethics of AI/ML in digital pathology
- Add AI/ML content to the web site’s FAQ pages, which is currently lacking AI
- Create an AI/ML section of the DPA web site. This may include:
- Educational content that synthesizes information such as the PV2019 “AI 101” lecture into an easy-to-access web-based format
- Begin an AI/ML tools repository, similar in concept to the WSI Repository, in which basic tools necessary for processing whole-slide images are made available to researchers (e.g. WSI anonymization, patch extraction, data augmentation)
- Pursue fee-based educational opportunities (similar to API’s recent Introduction to R Workshop).
- Interact with AI/ML and image management system vendors to encourage interoperability, perhaps with plans to pursue a Connectathon-style demonstration.
If you are interested in joining this Committee, please email the DPA staff, email@example.com
Mark Zarella, Johns Hopkins University (Chair)
Stan Cohen, Rutgers-NJMS
Toby Cornish, University of Colorado
Steven Hart, Mayo Clinic
Andrew Janowczyk, Case Western Reserve University
George Lee, Bristol Myers Squibb
Richard Levenson, UC Davis Health
Hooman Rashidi, University of California Davis
Hamid Tizhoosh, Kimia Lab, University of Waterloo
Jeroen van der Laak, Radboud University Medical Center Nijmegen
Joe Yeh, aetherAI