"An event like this is the perfect way to apply our academic knowledge in real life. The great response from the participants also shows that we should definitely repeat this," said Prof. Dr. Christian Bauckhage. In his keynote speech "Machine Learning in Industry", the Professor of Computer Science (Intelligent Learning Systems) at the University of Bonn and Lead Scientist for Machine Learning at Fraunhofer IAIS brought together his expertise from research and industry. The Co-Director of the LAMARR Institute for Machine Learning and Artificial Intelligence was a Senior Research Scientist at Deutsche Telekom Laboratories in Berlin before his appointment at the University of Bonn in 2008 and is therefore also very familiar with the practical side of machine learning. He then gave the participating students and young scientists the following advice on hacking: "Don't overthink it. Keep it simple."
Another relevant point was made in the keynote speech on the topic of Sustainable AI. Sophia Falk, a doctoral student at the Institute for Science and Ethics and the Sustainable AI Lab in Bonn, stood in for Prof. Dr. Aimee van Wynsberghe (Alexander von Humboldt Professor for Applied Ethics of Artificial Intelligence) and emphasized the importance of CO2 emissions in connection with AI. Given the enormous energy consumption caused by training AI, for example, this should always be seen in relation to the usefulness of the respective application and whether this justifies the emissions. Falk also presented measures for better "code efficiency" as a solution for programmers. "Our primary aim here is to raise awareness of the issue of sustainability so that these aspects are already taken into account when programming," explained Falk.
One day „Happy Hacking“
Christin Zörner, SVP Customer Experience & Data Analytics at Deutsche Telekom, described the focus of the Hackathon Challenge in her presentation. She emphasized how important the user-centric focus is for the company and explained what specific needs Deutsche Telekom has in terms of data analysis in this area. The aim is to turn customers into fans.
And so the task at #Hack4BestCX was to predict the Net Promoter Score (NPS) based on technical data about the home network and broadband connection. The NPS is a metric used to measure customer satisfaction with a company, product or service. It is typically measured by a single question: "How likely are you to recommend our company/product/service to a friend or colleague?" on a scale of 0 to 10. A high NPS indicates satisfied and loyal customers, while a low NPS indicates dissatisfaction and potential churn. For companies like Deutsche Telekom, it therefore makes sense to predict NPS using machine learning (ML) techniques. ML models can analyze large amounts of customer data, including demographics, usage behavior, interaction history and sentiment from various channels such as calls, messages and social media. By utilizing this data, ML algorithms can identify patterns and factors that influence NPS. "The challenge is really hard and you will be working on a real data set. You will need to bring all your data science skills to the table," said Benno Willoweit, Chapter Lead Data Scientists in the Customer Experience & Data department at Deutsche Telekom.
The teams then worked on the challenge for over six hours in a highly concentrated manner and without interruption. "I was impressed how concentrated everybody worked till the last minute. And the results showed that, too," Sophia Falk opened the award ceremony on stage together with Benno Willoweit. And the best result was delivered by Julia Dany and Felix Lehmann from the team "The Overfitting Overlords", who took first place. The two Master's students in Computer Science at the University of Bonn were delighted to accept their prize of 250 euros. The team "ML-DJ" came second: the four doctoral students Jonas Arruda, Manuel Huth, Lea Seep and Domagoj Doresic from the Computational Life Science, Bioinformatics and Mathematics research departments at the University of Bonn received a prize of 150 euros. As an individual, Zain Ul-Haq impressively made it to third place and thus won the third prize of 100 euros.
More Hackathons to come
However, not only the winners, but also the Telekom team was completely enthusiastic, especially about the high quality of the results. Dr. Heide Gesa Löhlein, Customer Experience & Data, gave a brief outlook on how to continue working with the data. Following the award ceremony, the discussion in the plenary session showed the high level of interest and commitment of the participants, who contributed many assessments to the discussion with the two Telekom Data Scientists Benno Willoweit and Dr. Heide Gesa Löhlein. Above all, the challenge but also the fun of working with real data was emphasized. In addition, a whole series of ideas and wishes for the next hackathon(s) have already been mentioned. „Great that we have been able to realize our joint hackathon idea together at Telekom Campus Bonn – one year after meeting at the German Data Science Society e.V. in Munich“, says Dr. Daniela Treutlein, Innovation Scout at the Transfer Center enaCom.
"Today has shown why hackathons are ideal as a format for knowledge transfer. Here, useful solutions were developed for real questions from practice. We enable students and young scientists to be directly involved in the development of innovations and bring them into contact with relevant companies," says Sandra Speer, Head of the Transfer Center enaCom. The Transfer Center enaCom regularly organizes hackathons on various topics, and more of these events are planned for 2024.
