New Perspectives in Science Education

The progress in development of Artificial Intelligence (AI) is receiving ever-increased attention in the public domain. This is accompanied by efforts to adapt teaching methods and content-creation in higher education. Supported by the German Federal Ministry of Research and Development (BMBF), we aim to integrate the fundamentals of this complex field into various faculties in an easily accessible manner. A particular focus is on novel approaches in teaching, especially with the use of practical examples both for entry- and advanced-level students as well as school education. As a research group we are focused on the sectors of medical technology and mechanical engineering. We consider the integration of AI with robotics by integrating image recognition systems and physical manipulation of objects with the long-term objective of manipulating medical instruments.

We implement three stages of teaching, which are marked by rising proficiency. In the first stage, we focus on school students with the aim to offer a first contact with an AI-empowered robotic system. We present a practical example, in which we teach students without prior programming knowledge the creation of a simple robot program. First, the algorithm is planned with abstract symbols and only after that “coding” of the algorithm in the graphical operating interface is carried out. As an advanced feature adding to the robot program, the aid of an Artificial Intelligence module can be integrated into the program, which offers insight into the implementation of such systems. The functionality of the AI module itself is merely explained to the students, neither coding of the AI nor details of the accompanying GUI are subject in this level of training exercise.

In the second level we focus on bachelor students. As an increased challenge, here the training of the AI module itself is added to the entry level task. This course level includes also text-based programming, development of algorithms and advanced configuration and training of the AI module.

In the third level we target master students. Here, also the data-handling between the robot, the camera and the AI module is to be planned and coded in a text-based programming language. Moreover, the algorithms and coding of neural networks are to be understood at the successful termination of the course.

This gradual increase in complexity allows learners on every level to build on their prior knowledge and gradually engage with more challenging aspects of AI in a real-world application.