Expectation-based parser for music

Seminar organized by Ken Satoh, Professor, National Institute of Informatics, 2022 Fondation France-Japon EHESS / Air Liquide Fellow, with Prof. Tojo, JAIST (Japan Advanced Institute of Science and Technology).

Abstract:

There have been proposed many grammar theories of music in comparison to language, and thus, the syntax has also been shown in tree. In their parsing processes, trees are usually constructed by the generative rules given the whole span of a music piece in top-down manner. In this paper, we propose an incremental parser of music, from the left to the right of the scores, building partial trees. In this process, we can predict the succeeding phrase from the beginning part. Therefore, the generation process of the tree itself represents our human expectation process of music. We apply our method to jazz music represented by chord sequences, implement the parser, and exemplify the incremental generating process in several standard pieces.

More information

Ken Satoh is a full professor of Principles of Informatics Research Division, NII (National Institute of Informatics), and Sokendai (The Graduate University of Advanced Studies), Japan.

Upcoming seminar sessions

• September 19, 2022 2pm - 4pm | Visitor: Prof. Takeda from NII, my institute
• September 28, 2022 2pm - 4pm | Visitor: Prof. Planas from NII, my institute