Hacker News Board Hacker News, browsed like a classic message board
Live data from Hacker News

Spotify’s Discover Weekly: How machine learning finds new music

hackernoon.com

Post reply on HN
Print view View original

151–160 of 275 posts

Re: Spotify’s Discover Weekly: How machine learning finds new music

#151

Post by mulletbum »

luckydata wrote: post #142

I spent quite a bit of time thinking about the recommendation algorithm and (after falling into a death metal hole I can't seem to escape on Spotify) I came to the conclusion that their analysis of the audio content of the song is way too shallow. You listen to songs because they have a common harmonic structure, or rythm, but not necessarily the same spectrum. That's why a metal fan might dig a cover of Metallica by…

I just always assumed it would be better to match up peoples likes/dislikes against one another. For example, I might really dig Brittany Spears, but I also like Pantera. Considering how those types of music have nothing to do with each other, it would be much better to match me up against other people who have similar taste in music as me.

So let's say for example there are 1,000 people who like Spears and Pantera (specific songs). Now take those 1,000 people and compare the other artists and songs I am listening to. Let's say, of the 1,000 people who like Spears and Pantera, 680 listen to a specific song by Creed. That's a high hit rate, which means that song should be offered to me as I will most likely enjoy it. It doesn't matter that Creed doesn't sound anything like Pantera or Spears.

Re: Spotify’s Discover Weekly: How machine learning finds new music

#152

Post by smeyer »

acdha wrote: post #148

I find the premise fascinating because of how bad the Discover Weekly recommendations were when I was looking for a replacement for Rdio. Rdio did a great job of suggesting new music but with Spotify I found myself constantly skipping tracks – combined with not having a way to play album tracks in order[1] I found no justification for using, much less paying for the service. 1. Yes, they claim to offer that for paid…

> combined with not having a way to play album tracks in order

As a paid subscriber for several years, I've had no difficulties playing album tracks in order. I'm not sure what issue you were having, though.

Re: Spotify’s Discover Weekly: How machine learning finds new music

#153

Post by e40 »

jasonrhaas wrote: post #141

I think that some people don't like the Spotify Discover Weekly because they don't want to admit that they like crappy music. Its kind of like when you hear your voice on a recording, you're like... "do I really sound like that?" Just face it, the algorithm has you pegged.

For the first few months, I really liked DW. The last few weeks, it's been giving me almost exclusively country rock. That is very different than it was before, though I would like an occasional country rock song (The Outlaws, for example).

I'm kinda bummed by this, since I really looked forward to listening to it on Monday going to work.

Re: Spotify’s Discover Weekly: How machine learning finds new music

#154

Post by mkane848 »

I'm going to have to go ahead and plug SAGE[0], because between Spotify recommending me the same bands/songs ad nauseum (or just being way off), friends already telling me they used it to find a new band to listen to on the first day it launched, and just being a fan of his work in general, it's a complete labor of love and I think it shows.

[0] https://medium.com/@hate5six/sage-an-artificially-intelligen...

Re: Spotify’s Discover Weekly: How machine learning finds new music

#155

Post by revx »

markussss wrote: post #140

The biggest problem with the Discover Weekly is its inability to understand _why_ you're listening to a specific subset of music. It might not be that your taste in music suddenly changed, or that you discovered a new genre that you're incredibly interested in, even though your most listened to genres or songs changed for a few weeks, or your listening patterns changed for a few weeks. A few examples: I'm Norwegian,…

My problem as well. I recently was in charge of putting together the playlist for a friend's get-together. They have significantly different taste in music from me, so now my recommendations are all messed up!

Re: Spotify’s Discover Weekly: How machine learning finds new music

#156

Post by mkane848 »

luckydata wrote: post #142

I spent quite a bit of time thinking about the recommendation algorithm and (after falling into a death metal hole I can't seem to escape on Spotify) I came to the conclusion that their analysis of the audio content of the song is way too shallow. You listen to songs because they have a common harmonic structure, or rythm, but not necessarily the same spectrum. That's why a metal fan might dig a cover of Metallica by…

If you're into metal, hardcore, or anything in between, you've seriously gotta check out https://medium.com/@hate5six/sage-an-artificially-intelligen... - it's from the guy who runs hate5six and it's fantastic

Re: Spotify’s Discover Weekly: How machine learning finds new music

#157

Post by madshiva »

criley2 wrote: post #138

Machine learning, lol. Ok. I can write a Discover weekly in about two lines of psuedocode. IF user.history.containsEvenOne("rap/r&b") THEN discover.weekly.setAll("rap/r&b") TA-DA~ you now have the complete "Discover Weekly" experience.

you have forget the badge!

Re: Spotify’s Discover Weekly: How machine learning finds new music

#159

Post by Dowwie »

Here's something I would never be recommended. I discovered the group by chance when listening to NPR one day -- The Frightnrs: https://www.youtube.com/watch?v=gwibcfkfuGQ&list=RDEMedGoIEt...

It's really unfortunate that their lead singer died of ALS as the group completed producing its album.

This recommendation was based on the preference of a third party (the DJ). I discovered by sharing his listening preferences.

Post reply on HN