In hardly any other German city to employ so many experts in artificial intelligence such as in Munich. Four Examples Of This.By Linus Frey mark and Pia Ratzesberger Pia Ratzesberger
the study of Economics, politics and Intercultural communication in Erlangen, Bologna and Munich. Scholarship holder at the Institute to promote journalistic talent (Ifp). Traineeship at the SZ with stops in Berlin and Brussels.on Twitter, send an E-Mail
Autonomous consequences of driving without accidents
The "lot of possibilities" for the work of Matthias Althoff is essential. It is calculated in seconds in advance, can hazards be eliminated and accidents autonomously moving vehicles, with a probability bordering on certainty be avoided. Althoff has studied in Munich in mechanical engineering and in electrical engineering, his doctorate. Since 2013, he is Junior Professor for computer science at the Technical University of Munich (TU) and is working on a method to make Autonomous Driving safe and 100 percent, if neglected a tiny residual risk.
Is a self-propelled vehicle is in motion, calculated Althoff's System, the amount of all the possibilities, the situations, the fraction of a second can occur in parts later. A braking vehicle in front, a suddenly crossing cyclist, a back-anrasendes motorcycle. The variants in which road hazards can arise are infinite. Nevertheless, Althoff says that his special method takes into account all the possible situations in which self-driving vehicles in an accident could cause - at least, if its assumptions don't fit, for example, a person speeds up when you Go to 40 kilometres an hour. "The process is proof," he says self-consciously. Before the first cars can be equipped, it takes but a couple of years.
sounds for virtual reality
If it is in the morning at the Track, to the screeching of the trains not hear, and if you are sitting in the evening in the Bar, the conversation from the table next door - assuming you don't want to hear it. Dagmar Schuller, 43, developed a Software, the unpleasant noise to filter out. To create around you a better, new world. The Software should at some point recognize even yourself, what you feel as a nuisance, without specifying it before. Dagmar Schuller says: "The systems are supposed to work for the people." She is the Head of Audeering; with computer scientists from the Technical University of Munich, she has founded the company two years ago, they had seven employees, but since this year you are more than 50 people. They work for clients such as BMW, or Huawei, or even Red Bull, they take care of everything that has to do with acoustic artificial intelligence. For example, the noise in the worlds of Virtual Reality, is to give an as realistic as possible, you can actually just motorcycle.
It is also important to recognize, on the basis of the language, as a product arrives at a customer, emotion reading. Dagmar Schuller had already started in the school, to program, then on a PC with black Background and green font. If one of the students had programmed something incorrectly, crashed all the computers in the room. Then you had to go to the teacher, and there was a black dot on the list - there were three, came out of the reference.
Franz Pfister is a medical doctor and a data specialist. He is working on a Software that is designed to help physicians in the diagnosis of brain diseases.(photo: Robert Haas)
machine views of medical data
Actually, Franz Pfister, a doctor, but he is working on a Software, with the help of his former colleagues in the hospital once easier. The program could avoid mistakes, the people are still running. Pfister says: "Because no one talks about it, but ten per cent of the deaths in the hospital due to a diagnostic error." He studied medicine because he wanted to understand the man and his disease better. However, at the age of 28, he decided to study once again to understand the man and his diseases, better still, with the help of Data Science. Franz Pfister had seen little time in the clinics for patients remained, of how complex the investigation was, and how overloaded the Physicians. In the evening, for example, the probability is higher that a doctor says a wrong diagnosis as in the Morning, Pfister, and if a doctor in a CT or MRI image have discovered a disease, sinking the probability that a second one, are most remarkable.
people are affected by their environment - a machine not. Franz Pfister, 31, and six colleagues have now developed an artificial intelligence, abnormalities in medical data. You be alerted, for example, after a CT the doctor right away, if there is an emergency, you will highlight which brain regions should he look at in more detail. Just Pfisters Team built a prototype of your project Deepc. In March, she will test the for the first Time, at the Klinikum rechts der Isar.
A Partner for us people
Dongheui Lee would have prepared your Pizza like a robot. Or in other words: you just tried, the robots with which it works, to teach how to bake Pizza and as a person it in front of the eight. "The robot should learn to understand human behavior and to do," Lee explains the goal of their research. The 41-Year-old has studied in South Korea, mechanical engineering, after a stop in Tokyo since 2009 she has been Junior Professor of electrical engineering at the TU and want to train yourself to point robots, human behavior. Lee still plans a second, bigger research: you just don't want the robot to analyze human behavior. You should also be able to react to it. An interaction between man and machine.
For their research, has employed Lee for a long time with human behavior. "It is always difficult to predict how a person reacts," she says. Almost as difficult is it for you to predict how your research developed. Maybe you can present it in four or five years a human - like robot, maybe, but in ten or 20 years ago. Lee is successful, you could have a chat sometime with a robot about the weather and the recent holiday. And by the way, a Pizza from him. If you show him how to do that.