[ONLINE] Douglas Eck: "Music recommendation and discovery at scale"

The rebroadcast of a Distinguished Lecture by a guest from Google (USA) followed by an online live discussion with Prof. Eck himself.

November 19, 2020
17:00 - 19:00
Douglas Eck

 

About the event

This session will feature the rebroadcast of the lecture presented by Douglas Eck on September 30, 2014 followed by a one-hour live discussion with Prof. Eck himself. The main goal is to revisit the topic, and then, in the discussion that will follow, evaluate what has changed since the research was first presented. Participants are encouraged to submit their questions and comments in the chat of the platform used.

To access the event: www.cirmmt.org/join/DL3

CIRMMT FUNDING ELIGIBILITY REMINDER: CIRMMT student members should note that attendance at Distinguished Lectures is necessary to fulfill the eligibility requirements for funding opportunities. Attendance will be tracked via sign-up sheets for in-person events, and the chat of the online meeting platform used. 

Abstract

I will discuss recent work done at Google to tackle the problem of music recommendation in a streaming music service.  I'll look at some strategies for using audio features as a means to improve quality and to go deep into the long tail. I'll also look at embeddings-based methods for collaborative filtering.  Finally I'll discuss the use of Knowledge Graph as a means for providing structured data about the world of music. An overarching theme in the talk is, "What do listeners actually want from a music streaming service?" I don't have a complete answer for this, but think it's worth talking about, especially since it motivates collaborations across domains relevant for CIRMMT. Though I'll address some of the technical and algorithmic details involved in building a music recommendation system, the talk is geared for a general audience.

Biography

Douglas Eck is a Principal Scientist at Google Research and a research director on the Brain Team.  His work lies at the intersection of machine learning and human-computer interaction (HCI). Doug created and helps lead Magenta, an ongoing research project exploring the role of machine learning in the process of creating art and music. He is also a leader of PAIR, a multidisciplinary team that explores the human side of AI through fundamental research, building tools, creating design frameworks, and working with diverse communities.  Doug is active in many areas of basic machine learning research, including natural language processing (NLP) and reinforcement learning (RL). In the past, Doug worked on music perception, aspects of music performance, machine learning for large audio datasets and music recommendation. He completed his PhD in Computer Science and Cognitive Science at Indiana University in 2000 and went on to a postdoctoral fellowship with Juergen Schmidhuber at IDSIA in Lugano Switzerland. Before joining Google in 2010, Doug was faculty in Computer Science at the University of Montreal (MILA machine learning lab) where he became Associate Professor. 

Video Archive

APA video citation:

Eck, D. (2014, October 30). Music recommendation and discovery at scale - 
CIRMMT Distinguished Lectures in the Science and Technology of Music. [Video file]. 
Retrieved from https://www.youtube.com/watch?v=GAzVZOZe2f8