deepGTTM-II: Automatic Generation of Metrical Structure based on Deep Learning Technique

Publication Type:

Conference Paper


Proceedings of the Sound and Music Computing Conference 2016, SMC 2016, Hamburg, Germany (2016)





This paper describes an analyzer that automatically generates the metrical structure of a generative theory of tonal music (GTTM). Although a fully automatic time-span tree analyzer has been developed, musicologists have to correct the errors in the metrical structure. In light of this, we use a deep learning technique for generating the metrical structure of a GTTM. Because we only have 300 pieces of music with the metrical structure analyzed by musicologist, directly learning the relationship between the score and metrical structure is difficult due to the lack of training data. To solve this problem, we propose a multidimensional multitask learning analyzer called deepGTM-II that can learn the relationship between score and metrical structures in the following three steps. First, we conduct unsupervised pre-training of a network using 15,000 pieces in a non-labeled dataset. After pre-training, the network involves supervised fine-tuning by back propagation from output to input layers using a half-labeled dataset, which consists of 15,000 pieces labeled with an automatic analyzer that we previously constructed. Finally, the network involves supervised fine-tuning using a labeled dataset. The experimental results demonstrated that the deepGTTM-II outperformed the previous analyzers for a GTTM in F-measure for generating the metrical structure.

SMC2016_submission_32.pdf726.98 KB
SMC paper: