Training mathematicians to tackle Big Data challenges
Recently, Horizon2020 granted the project proposal BIGMATH. Within this project, academia (Politecnico di Milano, TU Eindhoven, University of Novi Sad, and Universidade de Lisboa) and industry collaborate to train a group of seven young, creative mathematicians over a period of 3 to 4 years with strong theoretical and practical skills, needed to tackle the major challenges of the Big Data era. Two of the PhD’s will be appointed at TU/e’s department of Mathematics and Computer Science, where they will be supervised by Associate Professor Michiel Hochstenbach and Full Professor Wil Schilders.
Big Data is already big. However, in the foreseeable future, it will become even bigger; more and more data is gathered and used by the industry as well as government agencies. This comes with new challenges. For example, the new General Data Protection Regulation (GDPR) EU law on data protection and privacy requires that algorithms that make decisions about people become transparent. And for machine learning to become successful, more insight in the mathematical structures of the problem is needed.
Mathematical methods underexposed
We are not fully equipped to face many of these Big Data challenges just yet. One of the reasons for this is that mathematical methods currently are underexposed in the spectrum of data analysis. That is why the BIGMATH project aims to deliver highly skilled talents in mathematical fields of optimization, statistics, and large-scale linear algebra for Big Data, trained to develop new techniques that will bring data science to a new level, with a close link to practice. Therefore, they will also be trained in a wide set of 'soft skills' that enable them to transfer effectively their knowledge to the productive world, thus fostering the European market to create innovation.
BIGMATH focuses on 7 industrial Big Data problems spread across three domains: human facial data analysis, financial applications, and production systems. The human facial data analysis domain is about smart image processing. The financial applications domain focuses on gaining insight in the financial power of individuals, and making use of sensor data of industrial machines is the core challenge the production systems domain. These three domains have been defined by the industrial partners of the project, of which LIME, a spin-off company of the department of Mathematics and Computer Science, is one.
The collaboration between academia and industry is expected to make this project very effective. The industry is providing the link to the daily practice by exposing the PhD candidates to a set of Big Data-related real industrial problems, since Big Data challenges cannot be tackled only through theoretical studies. On the other hand, the academic partners will make sure the PhD candidates are provided with up-to-date training and knowledge on cutting-edge research on targeted mathematical disciplines, making it possible to reach a high level of competence.