Decomposing high-dimensional data objects for chemistry

Decomposing high-dimensional data objects for chemistry 

Large datasets are common in all branches of science, especially in chemistry. The datasets are often so large that calculations on data become a bottleneck. The goal of this project is to develop efficient and accurate algorithms for so-called tensordecomposition. Here you see data as a kind of large box and try to represent the effect of the contents of the box using simpler objects. That is, regardless of whether the box contains a million complicated things, in different cases different aspects are typically important, such as size, weight, etc. However, it is generally difficult to determine the relevant aspects in a particular context, and algorithm development is important here. Such algorithms can be absolutely essential for analyzing experimental chemical data sets and for making theoretical quantum calculations on molecular systems. The grant will fund two postdocs as well as additional costs of the project.