Masatoshi Hamanaka
Masatoshi Hamanaka researches music-information systems and computational music theory. Hamanaka shares authorship of work on music-analysis databases, musical tree structures and the deepGTTM series of learning methods. These methods examine local grouping, meter and other structural features of music. His research also includes CGBVS-DNN, a coauthored method for predicting compound–protein interactions with deep learning. He directs RIKEN’s Music Information Intelligence Team. His later work includes melodic morphing.
Timeline
-
2014–2015
Databases and musical trees
Hamanaka coauthored work on a musical structural-analysis database in 2014 and a learning-based time-span tree generator in 2015. The work developed computational representations of musical structure outside RIKEN.
-
2016–2017
The deepGTTM methods
Hamanaka, Keiji Hirata and Satoshi Tojo coauthored papers on deep-learning methods for local boundaries, metrical structure and joint learning of grouping and meter.
-
2016–2017
Compound–protein prediction
He coauthored CGBVS-DNN research for predicting compound–protein interactions with deep learning. The work dates to 2016, and its journal issue to 2017.
-
2023
Melodic morphing and team research
Hamanaka published a 2023 paper on melodic morphing based on generative tonal-music theory. He directs RIKEN’s Music Information Intelligence Team, whose research aims include computational media operations.
Sources and credits
Photo: www.riken.jp; creator not stated. License: Not stated by source. Source: riken.jp. Cropped to a square and resized. Ask us to remove or replace this photo.