RWTH team wins Siemens AI-DA Challenge
Three RWTH students took home the first prize with an AI model for safety applications in mobility
Artificial Intelligence is a potent instrument to increase efficiency and sustainability in the mobility sector. By far the biggest hurdles on the journey towards this goal are safety concerns. Siemens Mobility wants to tackle these problems – with the help of students. In the Artificial Intelligence Dependability Assessment (AI-DA) Challenge, students were given the chance to submit their solutions for a safety application based on machine learning techniques. Among the more than 430 registrations, there were three students from RWTH Aachen University.
Applying knowledge, learning by doing, and contributing to the solution of relevant issues
Eric Tillmann Bill, Robert Giesler and Felix Knispel are enrolled as bachelor students in Computer Science at RWTH Aachen University. And it was in one of their seminars that they learned about the AI-DA Challenge by Siemens Mobility. On the reasons for participation, Robert states: “For me, the motivation to take part was my wish to apply the theoretical knowledge from the proseminar in practice. The Challenge offered a great opportunity to delve into a new programming language and to think intensively about current problems in the field of machine learning.” He adds: “Besides, I hadn’t planned anything for the semester break.” With the first phase of the competition, the submission of an initial solution proposal, starting in February of this year, the team could make sure the challenge would not interfere with their exams. And as the duration of three months reveals, the complexity of the task in fact required the participants to fully focus on their proposals.
But why would they choose to spend their semester break working on a complicated extracurricular activity? For Eric, the AI-DA Challenge was a welcome twist to his original plans: “During the last semester break, I had set myself the goal of learning more about neural networks and started reading books about it.” Given the coincidence with his academic plans, he didn’t have to think twice: “When I heard about the challenge, I immediately joined because it was not only a project that would help me learn and understand the subject matter, but also a way to contribute to the solution of an important problem of ANN classifiers.” Felix, a third-year Computer Science student, completed the RWTH team. When Robert and Eric asked him to join, he was excited to help: “Since I attended a lecture on artificial intelligence and a seminar on neuroinspired computing, the challenge was the perfect opportunity to put what I had learned into practice. I also have a lot to do with classification problems in my student assistant job at the Institute of Medical Informatics, which turned out to be very helpful.”
AI model with “Most Balanced Safety Orientation”
Just how big of an impact the three students made became clear in their final pitch in front of Siemens experts. After passing the first expert evaluation in May, the team successfully validated their AI model. As one of three pre-selected teams, Eric, Robert and Felix finally delivered an elaborate presentation of their solution, including assumptions and challenges of their approach as well as an outlook on further steps, and were challenged to answer critical questions by the participating AI experts.
Due to the strength and diversity of the submitted proposals, the expert jury decided to split the total prize money of €15,000. In addition to five runner-up prizes (€1,000 each), the experts announced all three finalists as winners in distinct categories (€3,000 each). The RWTH team convinced the jury with the great significance of their proposal “Safety assertions in neural network classification”, which was rewarded with the prize in the category “Most Balanced Safety Orientation” and an additional fund of €1,000 for the quality of their report, analysis and presentation.
Mutually rewarding cooperation with long and fruitful history
Since 2003, Siemens has had close connections to RWTH Aachen University, which is one of eight universities worldwide to host a Center of Knowledge Interchange (CKI) as a productive platform for cooperation. As a key account partner, the company organises the CKI Conference, an annual get-together of researchers and industry experts. Last year’s topic, “Technology Architecture for Artificial Intelligence – Smart Cities & Digital Enterprises”, put the discussion on AI applications in urban areas high on the industry and research agenda. As the AI-DA Challenge reveals, implementing AI in the mobility sector has now become a top priority for cooperation.
The next CKI Conference will take place on 11 November 2021 and focus on "Sustainability in Production beyond Zero". If you are interested in joining, contact Christina using info@cki.rwth-aachen.de.