Several research groups are working in parallel on various approaches to contact tracing apps that could allow us to carefully move away from the extreme social distancing/lockdowns and reopen the economy while avoiding another wave of COVID-19 pandemics and saving thousands of lives (i.e. 'smart' social distancing). There are many challenges not least of which include privacy, risk estimation accuracy, scalability and efficiency, stability and robustness of the proposed approaches to various sources of noise or malicious attacks, and the sheer complexity of arriving at policies that allow for optimal trade-offs across multiple individual vs. societal objectives - a truly multiscale and multicriteria optimization problem.
The joint workshop will focus on various technical aspects of app design and contact tracing approaches more broadly, including but not limited to distributed approaches to probabilistic inference and machine learning, federated learning, differential privacy, and potential links to control in dynamical systems and reinforcement learning. The speakers include Alan Turing Award winner Yoshua Bengio, Lenka Zdeborova, Alfredo Braunstein, Ralf Herbrich, Manuel Gomez Rodriguez, and Kevin Murphy.
The workshop also serves the goal to further stimulate exchange between the research fellows of ELLIS and CIFAR. Important elements of ELLIS were originally inspired by the Canadian CIFAR program, and in December last year, CIFAR and ELLIS signed a memorandum of understanding.