Composing music is hard and the lack of inspiration can be daunting. A lot of elements are required to make it work: musical score, instruments, musicality, feeling, groove, originality. Music generation has been around for ages, even before the digital era, as a tool for musician to create new music and get inspired. What about machine learning? Can we use it as a tool for music generation? With Magenta, a music generation library based on Tensorflow, you can use the power of machine learning to help musical creation. We'll see why specific neural networks topologies, such as RNN, LSTM and VAE, have specific usage in music generation. We'll see how to train a model on your own style, used to then generate new rhythms, melodies and audio clips.
Hello, I’m a software engineer, conference speaker, open source maintainer and sound designer.
|Talks by tracks||Talks by session types||List of Speakers|