Addressing Data Challenges for next-gen Digital Therapeutics Development

The largest bottleneck medical research organisations at the forefront of technology face is access to useful data. These difficulties stem from strict policies that limit access to ML teams instead of allowing technology to enable and ensure responsible data use. Potential ML solutions are greatly stifled due to insufficient ability to access, standardise, and aggregate valuable data that continues to be siloed away. Yet, over the last years we have seen two general paradigms emerge to address important technical challenges. The first approaches the problem from a privacy-preserving ML perspective, with techniques such as federated learning, differential privacy, or use of secure execution environments, among others. The second tries to utilize limited data in the most efficient manner by employing techniques such as data augmentation or even complete replacement via synthetic data generation.

The goal of this workshop is to create an important dialogue to shape a data collection/processing paradigm for new digital therapeutic (DTx) approaches. The workshop will prepare a foundation for conducting massive longitudinal studies, employing technology that scales, empowering individuals through data dignity advocacy, and evaluating the potential of synthetic data within the coming wave of generative AI applications. Furthermore, the advantages and limitations of the current formats and protocols used for storing and transferring medical data will be discussed. The workshop aims to provide a comprehensive overview, fostering collaborative efforts to advance a new digital therapeutics initiative via addressing the challenges of our most important asset as ML practitioners in applied medical domains. 

The Warehouse at the Champalimaud Center for the Unknown is a multidisciplinary initiative to create the first Institute for Digital Neurotherapeutics in Europe. The goal is to bring together the best of industry and academia in an innovative ecosystem that creates new technological approaches for tracking and treating neurological and psychiatric conditions. The Institute is directed by John Krakauer and Joseph (Joe) Paton.

John Krakauer is Co-director of the Warehouse at the Champalimaud Center for the Unknown. 
He is currently John C. Malone Professor, Professor of Neurology, Neuroscience, and Physical Medicine and Rehabilitation The Johns Hopkins University School of Medicine. He is also an External Professor at the Santa Fe Institute and a Visiting Scholar at The Champalimaud Centre for the Unknown. He is Chief Medical Advisor to MindMaze.

Joseph (Joe) Paton received his undergraduate degree in biology from Tufts University in 2000. In 2008, he received his Ph.D. with distinction from Columbia University in Neurobiology and Behavior, and shortly thereafter joined the Champalimaud Neuroscience Programme, where he was a fellow from 2008 to 2012. He is currently a Principal Investigator and Director of the Champalimaud Neuroscience Programme (CNP) at the Champalimaud Foundation (CF) in Lisbon, Portugal. He is an alumnus of the Simons Foundation on the Global Brain, a Howard Hughes Medical Institute International Research Scholar, and a current holder of a consolidator grant from the European Research Council. His laboratory focuses on the neural mechanisms and computational principles of learning, decision-making, action selection and timing. In the course of trying to understand how the brain combines these functions to produce intelligent behaviour, he has increasingly moved into areas that overlap with modern AI research. During his mandate as CNP Director, Paton has also spearheaded the development of a new Digital Therapeutics centre at CF, as well as representing CF as a member of the recently announced Centre for Responsible AI.

Organising Committee

John Krakauer
Joe Paton
Eric Lacosse 
Alexander Loktyushin
João Santinha 
Tiago Marques
Johannes Stelzer
Niklas Fricke 
Fatemeh Molaei

Note: This is a closed event workshop; registrations are not open to the public.