[1]苗北辰,郭为安,汪镭.隐式特征和循环神经网络的多声部音乐生成系统[J].智能系统学报,2019,14(01):158-164.[doi:10.11992/tis.201804009]
 MIAO Beichen,GUO Weian,WANG Lei.A polyphony music generation system based on latent features and a recurrent neural network[J].CAAI Transactions on Intelligent Systems,2019,14(01):158-164.[doi:10.11992/tis.201804009]
点击复制

隐式特征和循环神经网络的多声部音乐生成系统(/HTML)
分享到:

《智能系统学报》[ISSN:1673-4785/CN:23-1538/TP]

卷:
第14卷
期数:
2019年01期
页码:
158-164
栏目:
出版日期:
2019-01-05

文章信息/Info

Title:
A polyphony music generation system based on latent features and a recurrent neural network
作者:
苗北辰1 郭为安2 汪镭1
1. 同济大学 电子与信息工程学院, 上海 201804;
2. 同济大学 中德学院, 上海 201804
Author(s):
MIAO Beichen1 GUO Weian2 WANG Lei1
1. College of Electronics and Information Engineering, Tongji University, Shanghai 201804, China;
2. College of China and German, Tongji University, Shanghai 201804, China
关键词:
音乐生成隐式特征提取循环神经网络栈式自编码器多声部音乐序列预测长短期记忆循环神经网络生成模型
Keywords:
music generationlatent feature extractionrecurrent neural networkstacked autoencoderpolyphony musicsequence predictionlong short-term memorygeneration model
分类号:
TP393.04
DOI:
10.11992/tis.201804009
摘要:
音乐生成是一种使用算法来生成音乐序列的研究。本文针对音乐样本特征提取以及自动作曲问题提出了一种基于音乐隐式特征和循环神经网络(recurrent neural network, RNN)的多声部音乐生成算法。该方法通过使用栈式自编码器对多声部音乐序列每个时间步的音符隐式特征进行提取,结合长短期记忆循环神经网络(long short-term memory, LSTM),以序列预测的方式搭建了基于隐式特征的音乐生成模型。仿真结果表明,该音乐生成算法在使用相同风格的音乐数据训练后,得到的模型可以生成旋律与和弦匹配较好的多声部音乐数据。
Abstract:
Music generation is a research area that uses algorithms to generate sequences with characteristics of music. Focusing on the problem of feature extraction from music samples and automatic music compositions, this paper proposes a polyphony music generation algorithm based on musical latent features and a recurrent neural network (RNN). The proposed algorithm uses a stacked autoencoder to extract latent features from of music sequence notes at each time step; the algorithm then uses long-short term memory RNNs to build a music generation system in the form of sequence prediction. The simulation results show that this algorithm can generate polyphony music with better melody and chord matching.

参考文献/References:

[1] ECK D, SCHMIDHUBER J. A first look at music composition using LSTM recurrent neural networks. Technical Report No. IDSIA-07-02[R]. Manno, Switzerland:Istituto Dalle Molle Di Studi Sull Intelligenza Artificiale, 2002:1-11.
[2] STURM B L, SANTOS J F, BENTAL O, et al. Music transcription modelling and composition using deep learning[EB/OL]. (2016-04-29)[2018-03-23]. https://arxiv.org/abs/1604.08723.
[3] CHOI K, FAZEKAS G, SANDLER M. Text-based LSTM networks for Automatic Music Composition[EB/OL].(2016-04-18)[2018-03-23].https://arxiv.org/abs/1604.05358.
[4] 王程, 周婉, 何军. 面向自动音乐生成的深度递归神经网络方法[J]. 小型微型计算机系统, 2017, 38(10):2412-2416 WANG Cheng, ZHOU Wan, HE Jun. Recurrent neural network method for automatic generation of music[J]. Journal of Chinese computer systems, 2017, 38(10):2412-2416
[5] LACKNER K. Composing a melody with long-short term memory (LSTM) recurrent neural networks[D]. Munich, Germany:Technische Universität München, 2016.
[6] CHU H, URTASUN R, FILDER S. Song from PI:a musically plausible network for Pop music generation[EB/OL].(2016-11-10)[2018-03-23].https://arxiv.org/abs/1611.03477.
[7] MAKRIS D, KALIAKATSOS-PAPAKOSTAS M, KARYDIS I, et al. Combining LSTM and feed forward neural networks for conditional rhythm composition[C]//18th International Conference on Engineering Applications of Neural Networks. Athens, Greece, 2017:570-582.
[8] OORD A V D, DIELEMAN S, ZEN H, et al. WaveNet:a generative model for raw audio[EB/OL].(2016-09-19)[2018-03-23].https://arxiv.org/abs/1609.03499.
[9] MOGREN O. C-RNN-GAN:continuous recurrent neural networks with adversarial training[C]//Constructive Machine Learning Workshop (NIPS 2016). Barcelona, Spain, 2016:1-6.
[10] SHEN Yan, XIE Meiping. Ship motion extreme short time prediction of ship pitch based on diagonal recurrent neural network[J]. Journal of marine science and application, 2005, 4(2):56-60.
[11] HOCHREITER S, SCHMIDHUBER J. Long short-term memory[J]. Neural computation, 1997, 9(8):1735-1780.

备注/Memo

备注/Memo:
收稿日期:2018-04-08。
基金项目:国家自然科学基金项目(71771176,61503287).
作者简介:苗北辰,男,1994年生,硕士研究生,主要研究方向为音乐生成的自动化;郭为安,男,1985生,副教授,博士,IEEE会员,主要研究方向为人工智能理论和应用。作为独立PI主持的项目包括国家自然科学基金青年基金、面上基金、上海市科学技术委员会等国家级和省部级项目。发表学术论文20余篇,被SCI检索10篇;汪镭,男,1970年生,教授,博士生导师,主要研究方向为群体智能、并行实现技术。发表学术论文90余篇,出版专著4部。
通讯作者:苗北辰.E-mail:m1104193501@163.com
更新日期/Last Update: 1900-01-01