Professor Dr. Sönke Hartmann is in charge of the course and was also responsible for the bachelor theses. He is very pleased about the outstanding quality of the work and the good acceptance of the new elective by the students:
"Artificial intelligence (AI) is a topic that is of great interest to many students at the HSBA. This ultimately led to the idea of offering the topic as an elective subject in the Business Informatics course. Students can not only get to know the basics and functionalities of AI, but can also implement and program AI methods themselves.”
Sören Spiegel (Ipsos), Niklas Scholz (EY) and Denis Semko (EY) have dealt with artificial neural networks and their application in their bachelor theses. Neural networks reproduce the structures in the brain and thus enable the learning of patterns in given data - a little like a human learns and in a completely different way than normal algorithms.
Bachelor Theses on Artificial Intelligence:
Sören Spiegel: "Text Classification of Open Entries in Questionnaires with Artificial Neural Networks"
As part of his work, Sören Spiegel has developed various neural networks to automatically classify answer texts from questionnaires according to content. Dr. Jan Olzem, Head of Data Science/CSS, Ipsos Operations GmbH, explains the importance of this work for his partner company Ipsos - an international market research company:
"Our industry is currently in a phase of upheaval, as the classic methods of market research are increasingly challenged by technologically sophisticated and often cheaper/faster AI-based methods. In particular, the high costs for manual editing of texts from surveys are becoming increasingly difficult to communicate. Therefore, the topic of this bachelor thesis is very important for us. Practical applications of the method developed by Sören Spiegel have already crystallized before the completion of the thesis. We are currently working on a concept for implementation within the framework of several customer projects, both in Germany and in the USA. A first productive application is planned for the year 2020".
Niklas Scholz: "Use of Artificial Neural Networks to Support Fraud Detection in the Annual Audit"
Niklas Scholz has implemented a so-called auto-encoder network for his thesis, which can help with the detection of fraud in the context of auditing by detecting anomalies in the data. Dr. Christoph Wockel, Senior Manager in Forensic and Integrity Services at EY, confirms the high relevance the work has for Niklas Scholz's partner company:
"The Forensic Data Analysis department of EY Wirtschaftsprüfungsgesellschaft is currently working on incorporating Niklas Scholz's approach into one of our main products. It will be a while before we arrive at a truly practical solution here, but his work will definitely be the starting- and reference point."
Denis Semko: "Making Urban Planning Smarter Using Machine Learning"
Denis Semkos' work (partner company: EY) did not focus on direct practical application, but on the scientific aspect. As part of the EY Nextwave Data Science Challenge, he designed a neural network of the type LSTM (Long Short Term Memory), which can use position data (anonymous mobile phone movement data) to predict to which location a person will move next - for example to the centre of a city. This model is mainly used in the context of smart city solutions and urban planning.
For more information on BSc Business Informatics: https://www.hsba.de/en/studies/bachelor/business-informatics-bsc/