Thoughts about Intelligent Machines

... and Things that Aspire to Become it!


"Computer science is no more about computers than astronomy is about telescopes." Edsger Dijkstra

Explainable Machine Learning Systems
Machine Learning System are often black box solutions after a model is learned. This is not desirable in many contexts. This project is linked to my ongoing Bachelor thesis at TU Darmstadt and will be deveoped during the next months.
Links: Upcoming Links will be published here
Android Machine Learning Application
Development of an AI-based Android App that learns from the users device interactions and can predict possible future actions in order to optimize device interaction. This project is in a conceptual stage. Google released a Neural Networks API for Android developers.
Links: Googles Neural Networks API
Privacy-preserving Classification
Most classification methods do not offer any privacy-preserving qualities, neither for the query-giver nor the model owner. This seems to be inappropriate for both parties as there is much value in the private information of a customer as well as in the prediction model. This project aims to work out an overview of privacy-preserving techniqes and to test implementations.