Interactive Exhibition

The “Neural Network” exhibit was acquired for Aalen University of Applied Sciences in mid-2024 in cooperation with the Vienna Technical Museum. Due to the positive response, the loan agreement between the AI Workshop for SMEs and the Vienna Technical Museum was extended through summer 2028.

Create your own pattern and see how a neural network uses it to recognize a digit from 0 to 9. This interactive exhibit demonstrates, step by step, how artificial intelligence processes information and makes decisions.

Activate or deactivate individual pixels to create your own pattern. Each cell is stored as a 1 or a 0 and forms the basis for processing. The neural network then attempts to recognize a digit from 0 to 9 in the pattern.

Track how your input travels through the neural network. The information is weighted, linked together, and processed across multiple layers. In this way, the network gradually recognizes increasingly complex patterns and relationships.

Watch how the neural network makes its decision. The output layer evaluates all digits from 0 to 9 and selects the most likely one. If multiple results are equally likely, they are also displayed.

The neural network installation is currently located in the Explorhino exhibition area in the Digital Innovation Spaces (1st floor).

Test how artificial intelligence recognizes objects. The exhibit classifies four different object categories and displays the probability that the recognized classification is correct. This lets you experience how modern image recognition works.

Hold an object in front of the camera and watch as the neural network analyzes the image. Within a few seconds, the AI assigns the object to one of four classes. It can recognize horses, pens, cars, and dice.

Discover how the AI arrives at its decision. The exhibit not only displays the recognized object class and its probability but also highlights, using a heat map, the areas of the image that were decisive for the decision.

Build your own robot—no programming required. Connect the magnetic Cubelets to each other and discover how individual building blocks come together to create functioning robots. Each combination results in different behavior and encourages experimentation.

Discover the different functions of the Cubelets. Sensor blocks detect information such as distance or brightness, processing blocks modify the data, and action blocks convert it into movement, light, or rotation. Together, they form a simple robot.

Try out different combinations and observe how the robot’s behavior changes. Even small changes can cause a robot to react differently. This way, you’ll learn the basics of sensor technology, data processing, and robotics through play.

Can artificial intelligence be creative? Play on the piano mat and experience how the AI picks up on your melody and develops it in real time. This creates a musical dialogue between human and machine.

Every improvisation begins with your playing. Change the tempo, touch, or note density, and watch how the AI’s response changes as well. Every musical encounter is unique and unfolds in real time.

A generative AI model processes your inputs. It was trained using the Magenta MAESTRO dataset, which consists of approximately 200 hours of professional piano music. The data comes from artists who have consented to its use.

Can AI make production more efficient? The Geochecker analyzes metal components and determines whether they can be suspended in a way suitable for powder coating. In this way, the AI assists employees with the pre-sorting of components.

Every component is automatically inspected. The AI analyzes the shape, holes, U-profiles, and center of gravity of the workpiece. It then assesses whether and how the component can be safely suspended—directly during the ordering process in the online store.

AI takes over time-consuming routine tasks. Components are inspected and evaluated within seconds. This allows employees to focus on complex decisions, while the inspection process runs faster, more efficiently, and with consistent reliability.

Can AI hear when something is wrong? The Miniature Water Plant simulates a closed water circuit in which various defects can be deliberately triggered. This makes AI-powered anomaly detection immediately tangible.

Every defect sounds different. Microphones record the sounds of the water circuit and feed them to the AI for analysis. The AI is designed to detect deviations from normal conditions and thus alert users to potential anomalies in the system.

Detection is only the first step. In addition to the audio signals and results, the AI’s explanations are also visualized. They help users understand which characteristics were crucial for detecting an anomaly—and make the decision more transparent.

What’s actually in the yellow bag? Grade Up! shows how AI recognizes plastic packaging and classifies it into different types of plastic. Place an object in front of the camera and see how image recognition becomes a building block for better recycling.

A camera captures the shape and visible features of the packaging. An AI model analyzes the image, identifies the object, and classifies it as a specific type of plastic, such as PET, PE, PP, or PS. This allows different materials to be separated more effectively.

Pure-grade plastics are a key foundation for high-quality recycling. AI can help classify packaging more precisely. This enables higher purity, better recycled materials, and a more targeted return of plastics to the recycling loop.

What does a smart production line look like? The demo factory showcases artificial intelligence directly in industrial manufacturing. Materials are identified, processed, and guided through an automated production process in conjunction with robotics.

The Demo Factory is the result of a student project during mechanical engineering Studies. It integrates AI technologies into practical training and prepares students for the challenges of Industry 4.0. This links theoretical knowledge with real-world applications

How do humanoid robots interact with their environment? The Unitree G1, which stands about 1.3 meters tall, can walk, gesture, and react to its surroundings. This allows users to directly experience modern robotics and human-machine collaboration.

The G1 was acquired by the Department of General Mechanical Engineering at Aalen University of Applied Sciences. It is used in research, teaching, and public outreach, enabling students and researchers to gain hands-on experience with modern robotics, AI methods, and human-robot interaction.