Session 1: Focus Group on AI for Health
The session will frame the AI4H workshop by providing an overview of the activities of the FG and the topics and aspects that the FG and workshop would like to address including opportunities for standardization of AI solutions for health. It will highlight how AI and machine learning can revolutionize healthcare in the areas of clinical decision support, personalized medicine, augmented diagnosis capabilities, medical imaging interpretation, therapeutic purposes and disease prevention. The session will provide an overview of the process and methodology of benchmarking AI algorithms for health, along with approaches and challenges to algorithms assessment. It will discuss how standardization process can lead to responsible adoption of algorithmic decision-making tools and to improving the accuracy of the algorithms. It will describe the proposal of a highly available system/ infrastructure for benchmarking the algorithms.   

Key Topics:

• Identify standardization opportunities for a benchmarking framework that will enable the broad adoption of AI for health (AI4H).
• Process to apply the benchmark dataset for testing new algorithms.
• What are different types of benchmarking we want to do.
• How to set-up an  open source testing infrastructure that will allow 24h testing system to operate

Moderator: Reinhard Scholl, Deputy Director, TSB, ITU [ Biography ]

• Thomas Wiegand, FG-AI4H Chair, Fraunhofer HHI, Germany [ Biography I Presentation ]
• "Consideration on the Development of AI", Min Dong, FG-AI4H Vice-Chair, CAICT, China
  [ Biography I Presentation ]
• "Framing the AI for Health to advance national health-related Sustainable Development Goals and targets", Ramesh Krishnamurthy, FG-AI4H Vice-Chair, WHO [ Biography I Presentation ]
• "On Benchmarking", Marcel Salathé, FG-AI4H Vice-Chair, EPFL, Switzerland [ Biography I Presentation ]

Session 2: Applications and Use Cases for AI in Health
This session will provide best practices for the applications of AI in Health

Key Topics:  Potential business case for AI for Health solutions for:

• AI enabled decision support systems role in self-assessment and professional setting
• Interpretation of laboratory results or imaging data (CT, MRI, X-ray, microscope images, photos, videos etc.)
• Extraction of diagnostically relevant features from text or complex sensor streams (ECG, EEG etc.)
• Leading areas of disease types that the AI communities are currently tackling (cancer, nervous system disease, cardiovascular and urogenital, digestive, respiratory, skin endocrine diseases) and potential health impact and RoI 

Moderator: Thomas Wiegand, FG-AI4H Chair, Fraunhofer HHI, Germany [ Biography ]

• "AI in cancer diagnostics", Frederick Klauschen, Charité Berlin [ Presentation​ ]
• Julia Bohlius, Cancer Epidemiologist, University of Bern [ Biography I Presentation ]
• "FastMRI", Nafissa Yakubova, Visiting Researcher, Facebook AI Research [ Biography I Presentation ]
• "First Derm", Alexander Börve, CEO and Founder of First Derm [ Biography I Presentation ]
• "AI Technology Development and Applications in HealthCare by Tencent", Yen-Hsiang Wang, PhD, Tencent AI Lab ​​​[ Biography I Presentation ]
•  "Child Growth Monitor", Jochen Moninger, Director of Innovation, Welthungerhilfe [ Biography I Presentation ]

Session 3: Data Availability and Benchmarking
This session will address the issue of availability of data sources for training and testing of the algorithms in the context of evaluation and validation with open benchmarks. It will address the need of open data sources and opportunities for how to make algorithms fair and unbiased, how to promote algorithmic accountability and how to make predictions equitable. It will describe the importance of having standardized formats for data input and data output in order to come up with some assessment/benchmark for algorithms.

Key Topics:

• Medical data required for training the AI4H algorithms and the required data types to advance AI algorithms development
• Availability of health data repository/open data sets for each application domain and how to assess the quality of data in those datasets.
• How to access clinical and health data to train AI algorithms (e.g., demographics, medical notes, electronic recordings from medical devices, physical examinations and clinical laboratory and diagnostic images, case reports)
• Identify who can share, own, control and maintain a variety of datasets and where should this data be stored in order to support 24/7 operational benchmarking infrastructure.
• Explainability of AI
  

Moderator: Marcel Salathé, FG-AI4H Vice-Chair, EPFL, Switzerland [ Biography ]

• "Building Personalized Medicine in Switzerland", Christian Lovis, Chairman, Division of Medical Information Sciences, University Hospitals of Geneva (HUG) [ Biography I Presentation ]
• "Clinical terms and ICD 11", Robert Jakob, Clinical terms and ICD 11, WHO [ Presentation ]
• "Ada Health- Our Appraoch to Assess Ada's Diagnostic Performance" Henry Hoffmann, Chief Scientist, ADA
  [ Biography I Presentation ]
• "Skapa- Swedish Quality Registry for Caries and Periodontal Disease", Hans Östholm, Deputy Registry Manager, Skapa Initiator [ Biography I Presentation ]
• "Chinese Practice of Benchmarking and Algorithm Evaluation", Shan Xu, Smar​t Health Department, CAICT​
  [ Biography I Presentation ]
• "Transparent and Trustworthy AI for medical applications", Wojciech Samek, Head of Machine Learning Group, Fraunhofer HHI [ Biography I Presentation ]

This session will discuss demand and opportunities to fund AI for Health initiatives including the Focus Group on AI for Health, as well as regional priorities. It will also highlight how to assess the business case for AI for health solutions in order to raise awareness and attract interest from foundations, venture capitalists and other donors.

Key Topics:

• Critical areas of interest based on countries priorities and global public health objectives, and potential health problems to which AI4H interventions can be applied and assessed for impact and scalability
• Possible funders and funding opportunities for AI for Health initiatives and ways to engage with them.

Moderator: Ramesh Krishnamurthy FG-AI4H Vice-Chair [ Biography ]  and Sameer Pujari, WHO [ Biography ]