Artificial Intelligence , Big Data , Broadcasting, TV, Radio & Film , Content & Media , Human Futures and AI , Rudy de Waele

Spotify using Deep Learning to Create the Ultimate Personalised Playlist


This post has already been read 416 times!

Spotify recently introduced Discover Weekly: the best-ever recommendations delivered to users as a weekly mixtape of fresh music. The streaming service will create weekly, personalized playlists based on people’s individual listening patterns.

This is a major improvement in the recommendations offered to you by Spotify – something I have been quite disappointed with from the music streaming service the last few years.

Spotify’s music-recommendation system has already become intelligent, thanks to machine learning that happens behind the scenes. But an academic researcher who’s been spending time at the company’s New York office has been looking at ways to use deep learning to surface songs based on their content.

What can you listen to on Spotify that you can’t on Rdio? Or, what’s a song on Google Play that isn’t on Apple Music? The problem with all you can eat streaming music is that everyone has the same buffets. Aside from a few high profile artists who are exclusive to one service or the other, these are basically commodity services with the same libraries. That means services have to compete on features. They have to be able to play, for example, the new Pitbull single for you, before you even think to request it. El Party! It’s why Apple has Beats 1, its human-curated radio station as a selling point. And it’s why Spotify now has a new automated playlist feature called Discover Weekly.

Read the full article.

Posted by Rudy de Waele aka @mtrends /

Rudy de Waele