I received my PhD (Doktor rer. nat.) in summer 2018 for my thesis with the topic “Cooperative and transparent machine learning for the context-senstive analysis of social interactions”.
My main research focus is on creating tools that help (non-)experts creating Machine Learning models.
The NOVA tool is an AI supported annotation tool, that allows to use Machine Learning already during the annotation of large continuous corpora, in order to speed up the task. We call this cooperative machine learning because both the machine and the user learn from each other during this process. NOVA also implements several eXplainable AI (XAI) Frameworks in order to get a better understanding of the inner workings of “black box” machine learning models.
NOVA is written in C# and runs on Windows. It further uses C++ and Python back-ends for feature extraction and the training process. NOVA can be found on Github:
For more information see my latest journal Article eXplainable Cooperative Machine Learning with NOVA in the KI – Künstliche Intelligenz journal
Once we trained these models we want to apply them in a real-time scenario, for example with a virtual agent or social robot. To this end I am part of the SSI development team. SSI (The Social Signal Interpretation Framework) is specialised on real-time processing of multiple sensor input. SSI is written in C++ and can also be found on Github:
With our tools we can help analyse affective and emotional reactions of people towards certain stimuli, e.g. advertisement videos, art installations, health-care applications and similar areas.
If you are interested in a cooperation or tutorial or want to know more details check out my github profile here:
or contact me at