Discover Weekly


Spotify's Discover Weekly: How machine learning finds your new music

#artificialintelligence

It's called Discover Weekly, and it's pretty much magic. I'm a huge fan of Spotify, and particularly Discover Weekly. As it turns out, I'm not alone in my obsession with Discover Weekly--the user base went crazy for it, which has driven Spotify to completely rethink its focus, investing more resources into algorithm-based playlists. Ever since Discover Weekly debuted in 2015, I've been dying to know how it worked (plus I'm a fangirl of the company, so sometimes I like to pretend I work there and research their products.)


Spotify's Discover Weekly: How machine learning finds your new music

#artificialintelligence

It's called Discover Weekly, and it's pretty much magic. I'm a huge fan of Spotify, and particularly Discover Weekly. As it turns out, I'm not alone in my obsession with Discover Weekly--the user base went crazy for it, which has driven Spotify to completely rethink its focus, investing more resources into algorithm-based playlists. Ever since Discover Weekly debuted in 2015, I've been dying to know how it worked (plus I'm a fangirl of the company, so sometimes I like to pretend I work there and research their products.)


Google's Gadget Vision: Same Stuff, Different Screens

WIRED

Your Google Home Max (because you wanted the best-sounding one, price be damned) reads out the weather and traffic conditions, lists off the day's events, and starts playing your Discover Weekly playlist as you turn on the shower. On one hand, it's a laptop: keyboard, trackpad, 12.3-inch screen. It can automatically connect to LTE through your phone, looks just like a Pixel, and runs Google Assistant and Snapchat and Instagram. Eventually, Google imagines a world where you switch gadgets because sometimes you want a pocket-sized, handheld device, and other times you want something with a keyboard and trackpad.


A succinct guide to machine learning for product managers

#artificialintelligence

It's now becoming common for me to hear that product owners, technical managers and designers are turning to popular online courses to learn about machine learning (ML). Instead, there is data that shows how Foursquare users were visiting venues that week -- and patterns in that data create trending venues. Songs that are recommended for a playlist looks like an unsupervised learning problem: the ML algorithm is looking for co-occurrence patterns in millions of playlists, to find songs that are commonly added to playlists that contain the songs that you have already added to yours. Many products can start to collect useful customer feedback using simple baselines; in the document, Martin quotes an example of sorting apps in an app store by download count (or popularity).


Machine Learning for Product Managers

#artificialintelligence

It's now becoming common for me to hear that product owners/managers, technical managers and designers are turning to popular online courses to learn about machine learning (ML). I always encourage it -- in fact, I did one of those courses myself (and blogged about it). However, it's not always clear how much benefit someone whose goal is to design, support, manage, or plan for products that use machine learning will get from doing an online course in ML. These courses throw you into the deep end, asking you to start programming classifiers, when many non-technical team mates are only looking for sufficient knowledge to be able to work in teams that are creating an ML-driven product. It's a bit like wanting to drive a car, and'therefore' signing up to a course on combustion engines -- probably a little bit too detailed for practical day-to-day driving!


How Did Spotify Get So Good At Machine Learning?

Forbes

I was at Spotify 2008–2015 and built up the machine learning team. Around 2013–2015 I built up a team around music recommendations that focused a lot on related artists, radio, and eventually Discover Weekly (which was released after I left). In my mind, the best way to get good recommendations is 90% through collaborative filtering then use deep learning models to get the extra 10%. It's been over two years since I left, but AFAIK, Discover Weekly is entirely powered by collaborative filtering, in particular, a few extensions to word2vec that the machine learning team in NYC built.


Commoditizing Music Machine Learning : Services

#artificialintelligence

Now there were multiple teams producing data; multiple teams developing personalization features. This was being used in services powering playlist recommendations on the Start Page, new user personalization on the Start page and Discover page. Our models were typically trained on a snapshot of training data to musical entity vectors. This will involve both systems work and modifying the machine learning models themselves.


How Spotify Is Leveraging Deep Learning To Shake Up The Music Streaming Industry

#artificialintelligence

Building on the already insanely popular Discover Weekly, the weekly roundup of songs Spotify thinks you'll love based on your user data and listening habits. The streaming giant is capturing the nostalgia of the enjoyment that comes with discovering new music by using big data and deep learning algorithms in their new Release Radar feature, with a little help from The Echo Nest, a music intelligence agency acquired in 2014. The latter is built on streaming data and playlist data, to tailor a playlist of songs related to your user streaming history that you'll more than likely love to listen to next – usually including some long forgotten gems you've not heard in a while! Spotify is using deep learning techniques to research the audio itself to pair with users streaming data to create new music recommendations.


The Little Hack That Could: The Story of Spotify's "Discover Weekly" Recommendation Engine

#artificialintelligence

He was responsible for one of those amazing things: a way to help Spotify users discover new music called Discover Weekly. Newett joined Spotify in 2013, initially working on a team developing a web page with personalized information, news about artists, album releases, and local concerts, along with a recommender system that offered suggestions of albums a user might find appealing. Discover Weekly would give users a personalized playlist of music they'd never listened to, designed to fit their musical tastes. Their system looks at what the user is already listening to, and then find connections between those songs and artists and other songs and artists, crawling through user activity logs, playlists of other users, general news from around the web, and spectragrams of audio.