Spotify News reposted this
How did Spotify build an algorithm that actually seems to know what music you want to listen to? By consistently pushing to improve the experience for listeners. “We built similarity vectors between songs and playlists,” says Spotify Co-President, CPO & CTO Gustav Söderström. While these similarity vectors offered unique discovery experiences for listeners, the product still had limitations. "We had no genre understanding," says Soderstrom. To solve this challenge, Spotify acquired The Echo Nest, an intelligence company skilled in classifying music. “Then we could start building great sessions where we could actually promise you that this is going to be Swedish pop, or this is going to be this vibe, not just these songs are similar in a vector space," explains Söderström. Which products or companies have you seen leveraging AI in forward-thinking ways? Tell us in the comments. #BuildingOne, a podcast hosted by LinkedIn's Chief Product Officer Tomer Cohen, features insights from top product leaders. Click here to listen and follow: 🎧 Apple: https://lnkd.in/eevtFBQu 🎧 Spotify: https://lnkd.in/eZn_ymh9 🎬 Mary Wilson Aulistar Mark
Alex Babinin
2w
I generally play my music according to my mood - but I also know to NOT use the Internet as a loudspeaker to broadcast what mood I'm in. As a result, Spotify does NOT have any data on what my music preferences or moods are. I compiled my music collection into .mp3 files long before LinkedIn or Spotify went big. I don't stream my music, I just play what I have from a pretty big stash of offline files. It directly follows that Spotify's recommendations are way off-base in my case. And I don't let Alexa/Siri/Google/Cortana/CoPilot listen in, either. Onii-chan, I prefer my privacy, please. 😒
Spotify is just a Playlist Generation tool. Spotify uses machine learning algorithms and AI models to create personalized recommendations for each user based on Genre (not Human Emotion). Well, This is not a Emotion based recommendation system. Genre is a choice of human not mood and mood is the Emotion not choice. Require research and development in Music Streaming industry. LinkedIn News
While there's been improvisation in recommending songs, working on the 'lyric' feature of songs will also facilitate a great listening experience for the users. Especially in a country like India where there's a wide variety of languages, this feature will help increase the engagement with the application. I mean, who would want to browse websites for lyrics during a JAM with a group of people?
Unfortunately, Spotify recommendation system is [censored] and there is still a lot of space for improvements and a long journey to go. Hopefully, one day it’s gonna work just right. At least I want to believe:) I like Gustav. I think he is great because he has a good skill, to explain complicated things in simple words.
FLORA BABY
2w
Spotify's innovative use ✨ of AI is truly impressive! I've also seen many companies leveraging AI for groundbreaking innovations, such as Tesla's application in autonomous driving technology and IBM Watson's contributions to healthcare. 🎉 Looking forward to seeing more companies driving change with AI.
For me it's not rocket science... Hahahaha Just a spreadsheet with metadata, so songs with as many points in common are more likely to be liked by a listener who liked a particular song... Innit?
Dandai Alves
2w
Parabéns! www.dandai.net
Jili Yu
1w
GREAT
Freelance Motion Designer | Ex Schibsted
2wAm I the only one who feels that Spotify, for the most part, regurgitates things I've listened to a ton in the past? If that really is the work of a "powerful" algorithm, well, then I'd rather have a more tailored experience or playlist put together by another human.