< Back to previous page

Project

Policy support for managing the COVID pandemic through artificial intelligence (FWOAL983)

Epidemiological modeling recently made important progress. The
large variety of models allow for simulations, which can be combined
with advanced optimization approaches using artificial intelligence, in
order to identify the most suitable prevention and containment
measures. We will extend state-of-the-art Reinforcement Learning
(RL) techniques, which have been shown to outperform the currently
used techniques by epidemiologists, and can deal well with
uncertainties. Next to epidemic factors, health (e.g. hospital load and
death counts) and economic factors will be included. It is clear from
cognitive sciences, that the way people and groups react to the
epidemic itself, and to the prevention and containment measures,
has a big impact on how the epidemic evolves. By taking these
cognitive variables into account the impact assessment and choice of
optimal measures will be improved. We allow for multi-criteria
optimization, such that policy makers can trade-off different aspects
by simulating and assessing the potential impact of each measure.
We will also pay attention to the communication of the outcome of
the learning process to the user, by building upon research on
explainable RL. This way the user better understands why certain
simulations have the impact that they have. The research will form
the basis for a valuable interactive tool for decision makers for the
current COVID-19 pandemic even when information on epidemics
only gradually becomes available
Date:1 Nov 2020 →  31 Oct 2021
Keywords:Reinforcement Learning, epidemiological modeling, cognitive biases, social psychology, behavioral economics
Disciplines:Adaptive agents and intelligent robotics, Bioinformatics of disease, Cognitive processes, Machine learning and decision making, Motivation and emotion