Counterpoint by Convolution

Abstract

Machine learning models of music typically break up the task of composition into a chronological process, composing a piece of music in a single pass from beginning to end. On the contrary, human composers write music in a nonlinear fashion, scribbling motifs here and there, often revisiting choices previously made. In order to better approximate this process, we train a convolutional neural network to complete partial musical scores, and explore the use of blocked Gibbs sampling as an analogue to rewriting. Neither the model nor the generative procedure are tied to a particular causal direction of composition. Our model is an instance of orderless NADE (Uria et al., 2014), which allows more direct ancestral sampling. However, we find that Gibbs sampling greatly improves sample quality, which we demonstrate to be due to some conditional distributions being poorly modeled. Moreover, we show that even the cheap approximate blocked Gibbs procedure from Yao et al. (2014) yields better samples than ancestral sampling, based on both log-likelihood and human evaluation.

Proceedings of the 18th International Society for Music Information Retrieval Conference, ISMIR 2017

References in corpus (4)

• Modeling Temporal Dependencies in High-Dimensional Sequences: Application to Polyphonic Music Generation and Transcription
• Professor Forcing: A New Algorithm for Training Recurrent Networks
• A Deep and Tractable Density Estimator
• Polyphonic Music Generation by Modeling Temporal Dependencies Using a RNN-DBN

Cited by in corpus (33)

• Jukebox: A Generative Model for Music
• A Comprehensive Survey on Deep Music Generation: Multi-level Representations, Algorithms, Evaluations, and Future Directions
• The Bach Doodle: Approachable music composition with machine learning at scale
• Creativity and Machine Learning: A Survey
• MMM : Exploring Conditional Multi-Track Music Generation with the Transformer
• Learning a Latent Space of Multitrack Measures
• Convolutional Generative Adversarial Networks with Binary Neurons for Polyphonic Music Generation
• The Effect of Explicit Structure Encoding of Deep Neural Networks for Symbolic Music Generation
• LakhNES: Improving multi-instrumental music generation with cross-domain pre-training
• Music Style Transfer: A Position Paper
• DAWSON: A Domain Adaptive Few Shot Generation Framework
• PIANOTREE VAE: Structured Representation Learning for Polyphonic Music
• Fair Generative Modeling via Weak Supervision
• From Artificial Neural Networks to Deep Learning for Music Generation -- History, Concepts and Trends
• Symbolic Music Generation with Diffusion Models
• Melody Generation for Pop Music via Word Representation of Musical Properties
• BandNet: A Neural Network-based, Multi-Instrument Beatles-Style MIDI Music Composition Machine
• Automatic Melody Harmonization with Triad Chords: A Comparative Study
• The NES Music Database: A multi-instrumental dataset with expressive performance attributes
• Bach or Mock? A Grading Function for Chorales in the Style of J.S. Bach
• Coupled Recurrent Models for Polyphonic Music Composition
• Generating Music with a Self-Correcting Non-Chronological Autoregressive Model
• Modeling Baroque Two-Part Counterpoint with Neural Machine Translation
• Melody Structure Transfer Network: Generating Music with Separable Self-Attention
• BacHMMachine: An Interpretable and Scalable Model for Algorithmic Harmonization for Four-part Baroque Chorales
• Melody Harmonization Using Orderless NADE, Chord Balancing, and Blocked Gibbs Sampling
• Harmonic Recomposition using Conditional Autoregressive Modeling
• Modeling Musical Structure with Artificial Neural Networks
• Actions Speak Louder than Listening: Evaluating Music Style Transfer based on Editing Experience
• RL-Duet: Online Music Accompaniment Generation Using Deep Reinforcement Learning
• Shallow Art: Art Extension Through Simple Machine Learning
• BumbleBee: A Transformer for Music
• Musical Speech: A Transformer-based Composition Tool