Automated Music Generation Using Recurrent Neural Networks

The paper aims to devise a set of machine learning models based on recurrent neural networks with emphasis on utilizing LSTM layers. These models are meant to be able to generate musical features such as melody notes or chords in sequence, or in other words generate music. Authors has decided to implement methods for music notation generation. Moreover, the paper contains a thorough description of the preprocessing of the obtained dataset along with the used ML technology and the latest research in related fields. In the paper, the authors elaborate on the process of training the devised models and example results of prediction done by the neural networks.