The cosmos of our Deep Learning Lab comprises various aspects each providing distinct opportunities: It starts with the aim to provide students practice-oriented and innovative learning with real data and real tasks. This opens up the Deep Learning Lab think tank as a playground to develop fresh ideas in a fun way, choosing methods and solutions freely. One such playground is the annual Deep Learning Lab machine learning challenge. If new business ideas take form either during this challenge or in other contexts, we specifically foster emerging start-ups, and support the founding of businesses. Another aspect of the Deep Learning Lab cosmos is the first contact and exchange with companies. A result from such exchanges can be machine-learning projects between industry and university, but they also give the companies access to young machine learning experts as future employees. Next to projects with industry, also government-funded projects and fundamental research are in our scope. Last but not least, bachelor, master, and Ph.D. theses originate from around the Deep Learning Lab and, at the same time, are an essential part of it.
Each year in summer, 10 student teams participate in the computer lab called “Deep Learning Lab”. During the introductory phase of the computer lab, the students are introduced to the programming language Python, solve simple programming tasks, and learn how to apply support vector machines and neural networks to solve easy classification problems. Additionally, they become acquainted with the GPU cluster.
After the introductory phase, students apply their knowledge to develop the best possible solution for the given task in competition with the other participating teams. For the DLL Challenge task, students will be provided publicly available or real-world industry data, while the actual test data is non-disclosed to participants. Tasks and datasets change every year.
The Deep Learning Lab concludes with a closing event where all teams present their solutions to the Machine Learning Challenge and the best three teams are awarded. Furthermore, the closing event allows representatives from industry acting as a sponsor to present their company and to get in contact with young machine learning experts, as about 50 Master and Ph.D. students participate in the event each year. The award ceremony is usually accompanied by a sponsored buffet with food from the barbecue and cold beverages.
More information about the Deep Learning Lab as a university course can be found in the following articles from the past years:
2021: KI gegen Pixelfälscher verteidigen - TU Braunschweig | Blogs (tu-braunschweig.de)
2020: Praxislabor mit Perspektive - TU Braunschweig | Blogs (tu-braunschweig.de)
2019: Kleidungsstücke erkennen mit künstlicher Intelligenz - TU Braunschweig | Blogs (tu-braunschweig.de)
2018: Mit echten Daten neuronale Netze trainieren - TU Braunschweig | Blogs (tu-braunschweig.de)
One possibility is to cooperate with Ph.D. students from our research group in a research project on AI and machine learning. You can also contribute by providing internships to our students, and with that giving them the chance to gain experience in the industry. If you can provide an intriguing task for the machine learning challenge from your own real-world data, we would be happy if you reach out to us, and we could start a cooperation with you for the Machine Learning Challenge. The same holds, if you want to become a sponsor for the closing event as an industry partner to present your company, exchange ideas, and connect with young machine learning experts.
