Learning without examples
In her winning thesis, Sindy Löwe describes an original and innovative concept within Artificial Intelligence, which she conceived and developed in collaboration with her supervisors Bas Veeling and Peter O'Connor. Unlike the common algorithms in AI, its "Greedy InfoMax" does not need examples to learn from. It provides a better explanation for and insight into the way in which cerebral neuronal networks learn.
Her thesis was chosen to be presented on the main stage of the NeurIPS conference, the most prestigious peer-reviewed conference in machine learning. There she also received one of the six 'outstanding new directions paper awards'.
UvA thesis prize 2020
In October, Sindy also won the UvA thesis prize 2020 for her graduation work.