Development of Mutual Adaptation Model of Communication based on Mutual Prediction and Its Verification Using Music

2018 Fiscal Year Final Research Report

• Project/Area Number: 16K16180
• Research Category: Grant-in-Aid for Young Scientists (B)
• Allocation Type: Multi-year Fund
• Research Field: Entertainment and game informatics 1
• Research Institution: Nihon University
• Principal Investigator: KITAHARA Tetsuro 日本大学, 文理学部, 准教授 (00454710)
• Project Period (FY): 2016-04-01 – 2019-03-31
• Keywords: 音楽情報処理 / 自動作曲 / 演奏支援 / 音楽予測

Outline of Final Research Achievements

The main achievements are as follows:
1. We developed the JamSketch system, which automatically generate melodies according to melodic outlines which are drawn by the user. It has been confirmed that melodies generated by the system have quality close to those composed by non-expert human players.
2. We developed a method for generating jazz bass lines from a jazz database using a hidden Markov model. By learning emission probability distributions separately for each beat positions, we achieved to generate high quality bass lines.

Free Research Field

音楽情報処理、知能情報処理

Academic Significance and Societal Importance of the Research Achievements

音楽演奏という観点からコミュニケーションのモデル化を行った。それに加え、近年、機械学習によって芸術コンテンツの生成をすることで創造性のアルゴリズム的側面を探求する試みが増えてきている。本研究は、機械学習により自動で演奏を生成する方法に関して新たな手法を提案しており、当該分野に一定の貢献をもたらしたと考える。