Project Magenta: Music and Art with Machine Learning (Google I/O '17)

Опубликовано 19 мая 2017, 23:21

Google Brain researcher Douglas Eck explores Magenta, an innovative project that leverages TensorFlow to create art and music through the power of deep neural networks and reinforcement learning. Discover how this initiative embraces the artistic and musical community. We dive into the technical details and challenges faced in building generative models. No machine learning expertise is necessary to engage with this Google I/O session.

Chapters:
0:00 - Intro
1:37 - Enable artists and musicians with ML
3:18 - Focus: Deep learning and reinforcement learning
4:24 - Project 1: Sketch-RNN
5:25 - Autoencoders
6:11 - Generative model for vector images
7:28 - Trained on Quickdraw
7:48 - AutoDraw
8:07 - Demo
12:19 - Generatlization, not memorization
14:10 - Project 2: NSynth
14:43 - WaveNet
18:15 - NSynth temporal autoencoder using WaveNet
23:47 - Nsynth dataset
25:18 - Resources

See all the talks from Google I/O '17 here: goo.gl/D0D4VE
Watch more Android talks at I/O '17 here: goo.gl/c0LWYl
Watch more Chrome talks at I/O '17 here: goo.gl/Q1bFGY
Watch more Firebase talks at I/O '17 here: goo.gl/pmO4Dr

Subscribe to Google for Developers → goo.gle/developers

#io17 #GoogleIO #GoogleIO2017 #Art #Music event: Google I/O 2017; re_ty: Publish;