Data Nerds Big Monthly Meetup

Data Nerds is a Data Science meetup that meets both weekly and monthly. Our weekly meetings involve working on data science related projects. For more information about these weekly meetups, please reach out to one of our meetup organizers. For our monthly meetup, we meet to discuss ideas or projects related to data. If you are unable to attend in person, we will be streaming the meetup live as well as posting the recorded meeting to YouTube, stay tuned for more information! Please try update your RSVP if you find yourself unable to attend the event in person.

This month we will be located in a room in the Belmont's Janet Ayres Academic Center. More details to come.

This month we have Scott Hawley, PhD, leading a conversation around his paper "SignalTrain." This is a project about a new ‘generic’ way of modeling musical audio signal processing effects: treat the audio effect as a black box, and by gathering lots of input and output audio, and recording what the “knobs” were set to each time, we train a neural network to do the same thing. Dynamic range compression has been a prohibitively hard problem up to now because it’s both nonlinear and time-dependent. I’ll present how we solved this -- designing the neural network, acquiring & structuring the data, training the model, etc, and show you the code & extra demos. The just-released paper & some demos are at http://www.signaltrain.ml

Scott Hawley (@drscotthawley) is a computational physicist with a Ph.D. in relativity, but has been teaching audio engineers at Belmont since 2006, and the Nashville pro-audio community has significantly altered the course of his research career, especially since he started developing machine learning apps for music production professionals.