Helmholtz Doctoral Award for KIT Computer Scientist

The extraction of information from large amounts of data, so-called data mining, is of great importance in today's society: for example, finding suitable therapies for new diseases, crime prevention or increasing the efficiency of industrial production processes. It involves analyzing huge amounts of data and identifying patterns such as anomalies.

However, data mining is becoming increasingly complex because data is often multidimensional, meaning it has many attributes and continuously arrives in a streaming fashion. These factors make statistical analysis of data increasingly unstable and less reliable. Moreover, insights are required in real-time.

With his PhD thesis "Estimating Dependency, Monitoring and Knowledge Discovery in High-Dimensional Data Streams", Edouard Fouché provides novel approaches for improving the extraction of information from multidimensional data streams. He has now received an award from the Helmholtz Association for this work.

Edouard Fouché 08:50' (Video: Helmholtz Association)

In his work, the data scientist first establishes a new method (Monte Carlo Dependency Estimation) that helps estimating the dependencies in huge streams of data. In a next step, Fouché develops two new algorithms that can efficiently monitor and verify statistics (Scaling Multi-Armed Bandit) and reliably identify patterns and deviations (Streaming Greedy Maximum Random Deviation). These new tools for knowledge discovery in high-dimensional data streams help in practice in a wide range of applications in industry, for example, to detect faults or to perform predictive maintenance.

Edouard Fouché completed his PhD at KIT from 2016-2020 in the DFG Research Training Group (GRK) 2153 "Energy State Data - Informatics Methods for Acquisition, Analysis, and Use." The GRK offers PhD students from different disciplines the opportunity to conduct interdisciplinary research on concepts for the sustainable use of energy. In addition to assistance in research aspects, the program includes a broad qualification curriculum with further training, coaching, support and networking among themselves and with experts in Germany and abroad.

Since completing his doctorate (Dr.-Ing.), Edouard Fouché has been working as a post-doctoral researcher in the Data Science research area of Prof. Klemens Böhm's Information Management Systems group. Previously, he studied computer science in Paris and Karlsruhe.

The Helmholtz Association awards an annual prize for outstanding doctoral theses in six Helmholtz research fields. The aim is to provide targeted support for talented young scientists at an early stage and to attract them to research in the long term.