Research on Automatic Music Generation Based on LSTM

The traditional way of music creation has a long cycle and requires the creator to have professional music knowledge. With the rapid development of artificial intelligence, the use of artificial intelligence algorithms to automatically generate music melodies can be used to assist the creation of musicians, providing musicians with new creative ideas, the use of computer understanding of music has become one of the hot spots of research. MIDI music files are converted music sequence information, with small size, easy to edit and other characteristics, very suitable for training using machine learning methods. Based on this feature, this paper is oriented to Musical Instrument Digital Interface (MIDI) music, combined with word vector technology and Long Short-Term Memory (LSTM) algorithm, constructed a deep learning model, trained the model, and realized the music automatic generation system, and the experimental results show that the music automatically generated by the system can basically satisfy the initial creation needs of the creators, and it has certain reference value.