Hackathon: Hack the Rock
We want to draw your attention to an upcoming Hackathon. On 26. of March we will meet in the Laboratory of IKG and explore two datasets. The first dataset consists of downhole logs from a sea bottom drill project. The second dataset contains raster data from electron microscopy of minerals. The goals are to develop novel approaches in the context of stratigraphic sequences and residence/eruption time estimation. If you want to participate, please register under firstname.lastname@example.org, this event is open to all scientists interested in machine learning.
1. Drill Holes
Drill holes on the earth surface are an effective, but expensive way to infer information about underground resources as well as past climate changes. In addition, obtaining the core is often necessary for analyses, which increases the drilling expenses even more. In order to reduce or even replace the necessity of the cores, geophysicists record several physical measurements along the drill hole. In this subproject of the hackathon, we will analyze these physical measurements and prototype approaches using machine learning.
2. Volcanic Crystals
A volcanic eruption is often associated with fear, but this natural phenomenon actually allows mineralogists to retrieve million years old information about the earth. The eruptions transport crystals, which have been preserved deep inside the earth’s body, to the surface. The chemical composition of these volcanic crystals contain valuable information about the processes in the earth’s interior. Therefore, crystals like Olivine, which are prepared as 2D raster images, are analyzed using optical microscopy, scanning electron microscopy (SEM), electron probe microanalysis (EPMA) and others. Unfortunately, the chemical analysis is expensive and time-consuming. This subproject of the hackathon will prototype approaches, which can replace or at least reduce the necessity of chemical analyses with machine learning methods.