Elisa Oostwal

Elisa Oostwal
University of Groningen | RUG · Johann Bernoulli Institute for Mathematics and Computer Science (JBI)

Master of Science

About

3
Publications
267
Reads
How we measure 'reads'
A 'read' is counted each time someone views a publication summary (such as the title, abstract, and list of authors), clicks on a figure, or views or downloads the full-text. Learn more
63
Citations

Publications

Publications (3)
View
Fig. 1. (a) Illustration of the network architecture with an N-dim....
Fig. 2. Sigmoidal activation. Learning curves for K = 2 (a,b) and K = 5...
Fig. 3. ReLU activation: Learning curves of perfectly matching student...
Fig. 4. ReLU activation. Learning curves of the perfectly matching...
+2 Fig. 5. Student cross overlap C . (a) Sigmoidal activations, here K =...
Hidden unit specialization in layered neural networks: ReLU vs. sigmoidal activation
Article
Full-text available
  • Feb 2021
By applying concepts from the statistical physics of learning, we study layered neural networks of 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...
View
Figure 2: Sigmoidal activation. Learning curves for K = 2 (a,b) and K =...
Figure 3: ReLU activation: Learning curves of perfectly matching...
Figure 4: ReLU activation. Learning curves of the perfectly matching...
Figure 5: Student cross overlap C. (a) Sigmoidal activations, here K =...
+1 Figure 6: Monte Carlo simulations of the ReLU system. (a) the...
Hidden Unit Specialization in Layered Neural Networks: ReLU vs. Sigmoidal Activation
Preprint
Full-text available
  • Oct 2019
We study layered neural networks of rectified linear units (ReLU) in a modelling framework for stochastic training processes. The comparison with sigmoidal activation functions is in the center of interest. We compute typical learning curves for shallow networks with K hidden units in matching student teacher scenarios. The systems exhibit sudden c...
View

Network

Cited
View All
Cited By
View All