Have a UROP opening you would like to submit?
Please fill out the form.
MAS: Media Arts and Sciences
Jan 1, 2020
Use generative neural networks to change the voice of newscasters into those of well recognised voices like Mick Jagger or Leonard Cohen. We have a set of news snippets from DirectTV, that can be tweaked We also have scripts that can run on GPU-CUDA machines. The following tasks need to be accomplished * Get set up on the Necsys GPU machines, at the media lab * Scrape recordings of the singer/celebrity we are interested in * Use https://github.com/andabi/deep-voice-conversion or other libraries to train and generate news recordings in the new voice. * Repeat the process for several celebrities.
Experience with machine learning/neural networks/GANs Experience with scripting/programming in Python Familiarity with Docker preferred