Program Summary
Detailed Program
8:30 - 8:40am Welcome Remarks 8:40 - 9:40am Keynote Presentation Project Magenta: Can Music Generation be Solved with Music Recommendation? Douglas Eck
Google Brain Magenta is a Google Brain project that investigates the question, “Can we use machine learning to create compelling art and music?” Our work is done in TensorFlow and is open sourced on GitHub at github.com/tensorflow/magenta. Probably the biggest challenge we face in this endeavor is that of evaluation. How do we judge the relative goodness of generated music and art? To address this, we're working to build a community of musicians, artists and consumers willing to experiment with generated media. With care, we should be able to use their feedback as a kind of reward for model improvement (regardless of whether the response is "I love it!" or "Go away, your stuff is boring!") I'll discuss the overall project and talk about some current research that I think warrants further investigation. I'll also discuss ways in which recommender systems might play a role in improving the quality of generated media. One important point: this is pretty speculative work. As such, I'll try to leave a lot of time for questions and discussion. 9:40 - 10:00am Invited Talks (20 min each)
Josh Moore
Cornell University (now at Facebook) 10:00 - 10:40am Coffee Break and Posters
10:40 - 12:00pm Invited Talks (20 min each)
Colin Raffel
Columbia University Justin Salamon
New York University
12:00 - 1:30pm Lunch
Lunch on your own. The Marriott offers three restaurant options internally and Times Square offers a variety of options nearby. Please be aware that finding inexpensive food will be very unlikely in this part of town, and make sure to take this into consideration during your search. 1:30 - 2:30pm Accepted Talks (15 min each)
Item2Vec: Neural Item Embedding for Collaborative Filtering
Oren Barkan1,2 and Noam Koenigstein1
Microsoft1 and Tel Aviv University1 Andreu Vall, Hamid Eghbal-zadeh, Matthias Dorfer, and Markus Schedl Department of Computational Perception, Johannes Kepler University Houssam Nassif1, Kemal Oral Cansizlar1, Mitchell Goodman1, and S. V. N. Vishwanathan1,2
Amazon1, University of California, Santa Cruz2 Brandon Morton and Youngmoo Kim
Drexel University 2:30 - 3:30pm Coffee Break and Posters 3:30 - 4:30pm Accepted Talks (15 min each)
Matthew Prockup1,2, Erik M. Schmidt2, and Youngmoo E. Kim1
Drexel University1, Pandora Media, Inc.2 Eamonn Bell1 and Jaan Altosaar2
Columbia University1, Princeton University2 Flavio Figueiredo1, Bruno Ribeiro2, Jussara M. Almeida3, and Christos Faloutsos4
IBM1, Purdue University2, Universidade Federal de Minas Gerais3, and Carnegie Mellon University4 Aloïs Gruson, Martin Laurent, Linnea Segerstedt, and Christophe Charbuillet
niland.io 4:30 - 4:45pm Closing Remarks 5:30 - 8:00pm Happy Hours and Discussion
The workshop will be followed by a lively happy hour and discussion session at Amity Hall craft beer bar. The venue is a short subway ride away in Greenwich Village. See the Happy Hour page for more information. We strongly encourage all participants to join :)
|