Computer Science – Learning
Scientific paper
2011-12-16
Computer Science
Learning
Preprint submitted to Neural Networks
Scientific paper
We investigate a recently proposed family of positive-definite kernels that mimic the computation in large neural networks. We examine the properties of these kernels using tools from differential geometry; specifically, we analyze the geometry of surfaces in Hilbert space that are induced by these kernels. When this geometry is described by a Riemannian manifold, we derive results for the metric, curvature, and volume element. Interestingly, though, we find that the simplest kernel in this family does not admit such an interpretation. We explore two variations of these kernels that mimic computation in neural networks with different activation functions. We experiment with these new kernels on several data sets and highlight their general trends in performance for classification.
Cho Youngmin
Saul Lawrence K.
No associations
LandOfFree
Analysis and Extension of Arc-Cosine Kernels for Large Margin Classification does not yet have a rating. At this time, there are no reviews or comments for this scientific paper.
If you have personal experience with Analysis and Extension of Arc-Cosine Kernels for Large Margin Classification, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Analysis and Extension of Arc-Cosine Kernels for Large Margin Classification will most certainly appreciate the feedback.
Profile ID: LFWR-SCP-O-136385