SCCS Colloquium - Aug 7, 2019

From Sccswiki
Revision as of 06:59, 30 July 2019 by Makis (talk | contribs)
Jump to navigation Jump to search
Date: August 7, 2019
Room: 02.07.023
Time: 15:00 - 16:00


Samuel Weber: Exploiting the Data Hierarchy with Geometry Aware Sparse Grids for Image Classification

This is a Bachelor's thesis submission talk. Samuel is advised by Kilian Röhner.

Through establishing a data hierarchy and suitable stencils, we were able to improve the accuracy for image classifications using geometry aware sparse grids. In the case of the MNIST dataset, the improvement is around 23%. The new class of stencils, the hierarchical parent stencils, was defined to exploit the data hierarchy, achieving smaller grids and better accuracy for image classification.

Keywords: Geometry-aware sparse grid, image classification, data hierarchy

Language: English

We are looking for a second talk! Please use the Colloquium registration form.

Details will be announced later.