Artificial Intelligence

 

Artificial Intelligence (AI for short) is one of the key technologies of our time, with the potential to revolutionize entire industries. Various ways of progress brought fundamental successes in neural networks, rule-based AI systems and important advances in machine learning techniques. In particular, developments in deep learning (especially pattern recognition) have improved massively in the last decades and reached various potentially economically important application areas. Headlines in this field have been made by AlphaGo, for example, a computer program that was able to beat the Go world champion. In addition, autonomous driving and voice assistants are among the fields of application. At the RWTH, AI methods are used, for example in the development of robots in production logistics. But machine learning techniques are also used in applications such as hot rolling and plastic injection molding, as well as mechanistic hybrid modeling for disease progression and real-world evidence patient data analysis.

The research follows the goal to develop the necessary technologies for machines to learn from data. There are two main research areas: classical AI, which includes higher-level cognitive functions such as planning, reasoning, and abstract problem solving. Also, machine learning, which includes probabilistic and statistical learning methods for classification, regression, and prediction problems. AI is not limited to individual applications, but is changing entire domains, for example how mobility is affected by the availability of autonomous vehicles.

AI works together with Data Science, since statistical learning methods are also always based on data and their analysis. The semantics derived from AI/machine learning methods is important for Data Science applications and methods. However, the two fields, although related, have different goals. AI research focuses on developing methods to perform specific tasks in limited domains. Extensive supervised learning, in particular, has driven successes in Deep Learning and machine learning. As a result, representation learning, for example has become a weakly supervised or unsupervised method of machine learning. Due to the research in the field of automated speech recognition at RWTH, major ICT companies such as Amazon, eBay and Apple have started research and development labs on "Machine Learning for Speech & Language Technology" in the vicinity of Aachen.