-
at different locations and at different times. The end-goal is ML-driven detection of complex activities, by designing neural networks for evolving spatio-temporal graphs to integrate multi-source on-the-fly
-
. You will implement reinforcement learning methods and/or graph based neural-networks incorporating process, thermal and optical data to predict the local and global material properties on-the-fly
-
combine chemical process engineering knowledge with data-driven approaches in an effective way. The methods that you will develop and apply can potentially include hybrid modeling, graph neural networks
-
process spreads over an underlying network topology or contact graph, that is generally changing over time. Standard models used so far are mostly based on Markovian processes and mean-field approximations