I primarily write technical documents for educational or work purposes, but from time to time I also enjoy writing a blog post here. That’s why I invested some time to find an easy way of publishing blog posts on this site. Because I live most of my editing time in the Emacs world, my priority searching for a workflow was to not leave that world and stick with the workflow I already use for technical documents.

The vast scale of the universe is a fascinating thing. Unfortunately most of it is quite far away and therefore hard to reach or even see with our blank eye. Fortunately for science and us, we invented tools that allows us to perceive things that are outside of our range of perception. For example high precision scales measure/perceive weight differences of objects by several orders of magnitudes better than a human ever could.

The current state of CI pipeline definition files Currently most generic CI systems (for example GitHub Actions or GitLab CI) let their users design pipelines in a single or multiple yaml files. This is fine for project tailored CI pipelines, because most steps are probably really project specific. These specific pipeline definition files can get really long without necessarily describing complex pipelines. Many of these pipelines execute the same step with different parameters, this can lead to a lot of duplicate code.

Introduction This article summarizes and gives an in depth explanation on the paper by Thomas Roddick and Roberto Cipolla, ‘Predicting Semantic Map Representations from images using Pyramid Occupancy Networks’ roddick2020predicting. The authors present a novel approach for creating a orthographic top down map of the environment based on a single or series of monocular images. This approach uses a wide variety of techniques and incorporates them into a single end to end trainable machine learning model.

Unsere Welt wird durch technologische Fortschritte immer digitalisierter. Auch im Bereich der Justiz, wird bereits Forschung betrieben um durch die Nutzung von Big Data und Algorithmen, Ermittler zu unterstützen. Ob nun finanzielle Transaktionen algorithmisch ausgewertet werden, oder automatisierte Gesichtserkennung auf öffentlichen Überwachungskameras eingesetzt wird, es lässt sich nicht abstreiten, dass diese Technologien bereits an vielen Orten eingesetzt werden. Da sich der technologische Fortschritt deutlich schneller entwickelt als gesetzliche Regulierungen, ist es zwingend notwendig sich mit ethischen Fragestellungen auseinanderzusetzen.

SVM Introduction Support vector machines (in short SVMs) are a binary classification machine learning algorithm. SVMs try to find an $n$-dimensional hyperplane, separating two classes from each other. The hyperplane should maximize the margin between both classes, because of this the generalization potential of SVMs is superior to most other similar classification algorithms. This makes support vector machines a very popular choice for a wide range of use cases. SVMs use a linear function $f(x) = wx+b$ that should separate two classes ($-1$ and $1$).