Machine learning is concerned with detecting patterns and extracting information from large sets of data, where the underlying probabilistic models may be unknown or unavailable. In contrast to statistical signal processing, where often a system model is assumed and a priori information is integrated into algorithms, machine learning methods attempt to extract the knowledge directly from the data. The algorithms which belong to the toolbox of machine learning, e.g., support vector machines, random forests, or deep neural networks, provide high degrees of freedoms and are capable of approximating highly complex functions. During the learning phase, these degrees of freedom can be leveraged to approximate the underlying statistical model of the data. This trained algorithm is subsequently used to classify previously unseen data. Due to the ever-increasing availability of processing power, e.g., GPUs, machine learning methods can be employed in various applications.
Machine learning is used in the following fields of research at our institute:
To learn more about machine learning and about fundamental tools that are relevant for machine learning, we kindly refer to the following lectures offered at our institute: