In Search of the Horowitz Factor
(invited lecture for ALT 2002 and DS 2002)

Author: Gerhard Widmer

Affiliation: Department of Medical Cybernetics and Artificial Intelligence, University of Vienna, and Austrian Research Institute for Artificial Intelligence, Vienna

Abstract. The potential of machine learning and automatic scientific discovery for various branches of science has been convincingly demonstrated in recent years, mainly in the natural sciences ((bio)chemistry, genetics, physics, etc.). But is computer-based scientific discovery also possible in less easily quantifiable domains like art?

This talk will present the latest results of a long-term interdisciplinary research project that uses AI technology to investigate one of the most fascinating - and at the same time highly elusive - phenomena in music: expressive music performance. We study how skilled musicians (concert pianists, in particular) make music 'come alive', how they express and communicate their understanding of the musical and emotional content of the pieces by shaping various parameters like tempo, timing, dynamics, articulation, etc. The audience will be taken on a grand tour of a complex discovery enterprise, from the intricacies of data gathering (which already require new AI methods) through novel approaches to data visualisation all the way to automated data analysis and inductive learning. It will be shown that even a seemingly intangible phenomenon like musical expression can be transformed into something that can be studied formally, and that the computer can indeed discover some fundamental (and sometimes surprising) principles underlying the art of music performance.

The title of the talk refers to the late Vladimir Horowitz, one of the most famous pianists of the 20th century, who symbolizes like few others the fascination that great performers hold for the general audience. Explaining the secret behind the art and magic of such a great master is an incredibly thrilling goal. Whether the 'Horowitz Factor' will indeed be revealed in this presentation is, however, rather doubtful ...

©Copyright 2002 Autor and Springer