ML is about developing algorithms for analyzing data. In a multitude of conceivable application areas they help to solve problems such as classification, recognition, inference or predictive modeling, amongst others. Apart from practical innovations regarding novel algorithms, architectures and learning strategies, we also need a thorough theoretical analysis to provide performance guarantees or stability analysis. Without such theoretical guarantees, the purely empirical successes are challenged by the doubt on reliability. ELLIS Munich welcomes contributions of talents to the great challenge of AI explicability and the development of novel, reliable and intelligent algorithms for future applications.
