I am a PhD student in Computing Science, part of the Intelligent Systems group at the Bernouilli Institute for Mathematics, Computer Science and Artificial Intelligence at the University of Groningen. My project combines machine learning methods with aspects from system identification and differential geometry to achieve explainable results.
ABSTRACT: By applying concepts from the statistical physics of learning, we study layered neural networksof rectified linear units (ReLU). The comparison with conventional, sigmoidal activation functions is in the center of interest. We compute typical learning curves for large shallow networks with K hidden units in matching student teacher scenarios. The systems undergo phase transitions, i.e. sudden changes of the generalization performance via the process of hidden unit specialization at critical sizes of the training set. Surprisingly, our results show that the training behavior of ReLU networks is qualitatively different from that of networks with sigmoidal activations.
April 10-11, 2023
September 12-16, 2022
August 22-24, 2022