On April 29, 2014, Prof. Dr. Eyke Hüllermeier will give a lecture on "How to make computers behave intelligently: About algorithms, knowledge and data".
Abstract:
The first part of this talk elaborates on the evolution of intelligent systems design during the past decades and the development of machine learning into one of the core methodologies of modern artificial intelligence. This also includes a discussion of the influence of machine learning on computer science in general, its use in various domains of application, as well as its meaning for the "information society" in the era of "big data". For the sake of illustration, and to convey an idea of machine learning as a scientific discipline, the second part of the talk will be devoted to the formalization and analysis of a concrete learning problem. More specifically, an adaptive sampling algorithm for the recent problem of preference-based online learning in the setting of multi-armed bandits will be presented, along with an analysis of its sample complexity.