3月15日13時に第15期公募情報を公開いたしました。
4月1日13時より応募者登録サイトへの登録が可能です。
Information on the 15th call for applications was opened at 13:00 on 15 March.
Applicants can register on the registration website from 1 April at 13:00.
Intelligent Music Information Processing and Statistical Learning
While our lives are surrounded by music, it is not easy for everyone to create or transcribe music. High-level intelligence is at work in such processes as composition, performance, classification, and transcription of music. By revealing the mechanism of these processes, we can understand the properties of high-level human intelligence and develop technologies including automatic music generation and transcription that can improve the quality of our lives. In this seminar, I will explain attempts to analyze, emulate, and predict intelligent musical behaviours from the viewpoint of time-series information processing and discuss the importance of statistical learning in understanding the relevant knowledge acquisition processes in a unified way. For example, knowledge of features of a specific time period, region, genre, or individual creator is necessary for musical composition. It has been shown that completely describing the knowledge of these expression styles using grammatical rules is difficult and that knowledge about statistical features such as the frequency of combinations of musical notes is important. We can also find the roles of refined learning mechanisms including unsupervised learning and Bayesian learning in understanding musical syntactic structure and in classification of expression styles. I will present studies towards understanding the implicit knowledge behind musical culture in the perspective of statistical learning and introduce recent technologies of intelligent music information processing.
Opening talk
"An observatory in Argentina"