It's no surprise that SKIDATA works with face recognition. But it might surprise you to hear that we also recognize food! Felix Köppl, Vice President of Business Development, explains how we make use of artificial intelligence.
Felix, what does SKIDATA understand by artificial intelligence?
Unfortunately, artificial intelligence is an overused term in the media today. Many associate artificial intelligence with robots that behave and think humanely, and which are supposed to be superior to us in many ways. In fact, research today is still a long way from achieving anything like this. The overriding goal of artificial intelligence is for machines to solve problems independently. This is what we teach machines with the help of deep learning, a method of machine learning. We train an algorithm on a large amount of data in order to provide accurate results for a specific problem. In practice, these algorithms are already widely used today – for example, in predicting stock prices or product recommendations, as well as in the recognition of speech or traffic situations for autonomous driving.
Where is SKIDATA already using artificial intelligence today?
We have several projects. One project is Trayzi. This is a fully automated check-out service for restaurants and snack shops. With the help of cameras and deep algorithms, Trayzi recognizes what the customer has on their tray. With the help of a large quantity of food images, we teach the machine to recognize food and beverages. The algorithm can then price exactly what the customer has taken. Another project is face recognition for secure access to stadiums. Visitors are not only identified by their ticket, but also by unique facial features. This enables stadium operators to ensure that the person who bought the ticket is also the person entering the stadium.
Artificial intelligence is particularly critical in sensitive areas such as human security. But data protection is also an issue. How do you intend to create a climate of acceptance?
We deal with these issues very transparently and openly. In the case of Trayzi, we explain in detail to the restaurant operator how the technology works and what data we use. The operator can in turn communicate this openly to their customers and ensure that they always have a choice between our automatic checkout and a traditional checkout.
How does SKIDATA ensure an open and proactive use of new technologies within the company?
Before we test our prototypes in the field, we conduct internal trials. This helps us test technical processes, but we also hear feedback from our colleagues and get them excited about our new technologies. For example, we set up a pop-up canteen where employees could try out Trayzi for themselves over several days. Most colleagues were enthusiastic – and not just about the food!
The market for talent in these technology areas is highly competitive. Why should young developers and engineers come to SKIDATA?
Trayzi acts like a start-up embedded within a world market leader. This means we offer the financial security of an international group while being able to implement new product concepts quickly and dynamically. I think we are attractive to talent because we work on genuinely unique product innovations. I see this every day in the great passion and dedication of our employees to Trayzi. Our engineers work in a very open, dedicated and international team. They can use the latest frameworks and computing infrastructure, and we provide them with all the data they need to try out the latest models and deep-learning architectures. Anyone can contribute their ideas, no matter how crazy they are! We are currently growing rapidly at our location at the Ostbahnhof in Munich. I am looking forward to hearing from applicants with experience in deep learning!
Which products will make use of artificial intelligence in the future?
I believe that self-learning algorithms will be an integral part of every SKIDATA product in the future. They enable precise predictions, like the future utilization of parking spaces, automated processes, like our Trayzi Checkout, and they increase security, for example, in facial identification when entering stadiums. Accuracy will continue to improve with the availability and quality of data, and new model architectures, adding value to more and more products.