Nettsider med emneord «neural networks»

In this project, we  addresses fundamental question of “How much should we overparameterize a NN?” with a focus on genralizaiton and common practice in DL such as SGD, nonsmooth activations, and implicit/explicit regularizations.  For smooth activations and gradient descent, we established current best scaling on the number of parameters for fully-trained shallow NNs under standard initialization schemes [1].

In this thesis, you will investigate how modern structure-aware machine learning techniques can be applied to practical challenges usually approached using classic symbolic AI techniques. Specifically, you will develop novel algorithms based on graph neural networks (GNNs) for anomaly detection in knowledge graphs, and test them against existing approaches in synthetic and real-life settings.

Please read more in the Norwegian project page.