3 ms·

I wonder what would happen if you train an RNN like described with, say, the scores of all of Mozart's Chamber Music and then let it generate new music from the

by rsp1984 11y ago

I wonder what would happen if you train an RNN like described with, say, the scores of all of Mozart's Chamber Music and then let it generate new music from the learned pieces. How would it sound? Would it figure out beat? Chords? Harmonies? May it even sound a bit like Mozart?

Nadya 11y ago
There's a few such projects in existence. Perhaps not RNN-Mozart inspired, but I'm sure that exists too. Emily Howell https://www.youtube.com/watch?v=QEjdiE0AoCU https://www.youtube.com/watch?v=QEjdiE0AoCU Here's a Bach-inspired computer-generated song: https://www.youtube.com/watch?v=PczDLl92vlc https://www.youtube.com/watch?v=PczDLl92vlc
kastnerkyle 11y ago
The work of Nicolas Boulanger-Lewandowski was extensively focused on this topic, see his work [1]. He wrote a Theano deep learning tutorial on this topic [2], and several people (Kratarth Goel) [3][4] have advanced the work to use LSTM and deep belief networks. For a brief while RNN-NADE made an appearance as well, though I do not know of an open source implementation There are also a few of us who are working on more advanced versions of this model for speech synthesis, versus operating on the MIDI sequence. Stay tuned in the near future! I can say from experience that some of the samples from the LSTM-DBN are shockingly cool, and drove me to spend about a week using K-means coded speech. It made robo-voices at least but our research moved past that pretty fast. [1] http://www-etud.iro.umontreal.ca/~boulanni/ http://www-etud.iro.umontreal.ca/~boulanni/ [2] http://deeplearning.net/tutorial/rnnrbm.html http://deeplearning.net/tutorial/rnnrbm.html [3] http://arxiv.org/pdf/1412.6093.pdf http://arxiv.org/pdf/1412.6093.pdf [4] https://github.com/kratarth1203/NeuralNet/blob/master/rnndbn.py https://github.com/kratarth1203/NeuralNet/blob/master/rnndbn...
- JonnieCache 11y ago
  Is the robot-voice code published anywhere? You can make money out of that kind of thing btw! https://soniccharge.com/bitspeek https://soniccharge.com/bitspeek (Obviously not the same thing but the point is that silly robo-voice code is marketable :)

by rsp1984 11y ago

Nadya 11y ago

There's a few such projects in existence. Perhaps not RNN-Mozart inspired, but I'm sure that exists too. Emily Howell https://www.youtube.com/watch?v=QEjdiE0AoCU https://www.youtube.com/watch?v=QEjdiE0AoCU Here's a Bach-inspired computer-generated song: https://www.youtube.com/watch?v=PczDLl92vlc https://www.youtube.com/watch?v=PczDLl92vlc

kastnerkyle 11y ago

The work of Nicolas Boulanger-Lewandowski was extensively focused on this topic, see his work [1]. He wrote a Theano deep learning tutorial on this topic [2], and several people (Kratarth Goel) [3][4] have advanced the work to use LSTM and deep belief networks. For a brief while RNN-NADE made an appearance as well, though I do not know of an open source implementation There are also a few of us who are working on more advanced versions of this model for speech synthesis, versus operating on the MIDI sequence. Stay tuned in the near future! I can say from experience that some of the samples from the LSTM-DBN are shockingly cool, and drove me to spend about a week using K-means coded speech. It made robo-voices at least but our research moved past that pretty fast. [1] http://www-etud.iro.umontreal.ca/~boulanni/ http://www-etud.iro.umontreal.ca/~boulanni/ [2] http://deeplearning.net/tutorial/rnnrbm.html http://deeplearning.net/tutorial/rnnrbm.html [3] http://arxiv.org/pdf/1412.6093.pdf http://arxiv.org/pdf/1412.6093.pdf [4] https://github.com/kratarth1203/NeuralNet/blob/master/rnndbn.py https://github.com/kratarth1203/NeuralNet/blob/master/rnndbn...

JonnieCache 11y ago
Is the robot-voice code published anywhere? You can make money out of that kind of thing btw! https://soniccharge.com/bitspeek https://soniccharge.com/bitspeek (Obviously not the same thing but the point is that silly robo-voice code is marketable :)