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

221–230 of 275 posts

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

#221

Post by a1k0n »

cjCamel wrote: post #13

For anyone interested in the Raw Audio Models section of that articles, there are some fun endpoints[1][2] from the Echonest API that provide those models. They've moved over to the Spotify API since I last had a play, but it's great that they still provide them. You can get the audio breakdown of a track, as well as a summary of the track features including fun stuff like "danceability" and musical positiveness ("va…

I also wanted to nitpick this part of the article -- the neural network isn't determining the audio features; these are all hand-engineered features developed by The Echo Nest.

The neural network is trained to mimic collaborative filtering vectors from raw audio. It's a separate model from time signature, key, mode, tempo, loudness, etc.

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

#222

Post by MisterKent »

Asdfbla wrote: post #29

Sometimes I wish Spotify added a bit more 'noise' to their recommendations, so to speak. If I don't listen to much music except Discover Weekly for a few weeks, I (subjectively) find that what's recommended to me more or less sounds the same after a while. Either they are afraid to insert new things that stray too far from an optimal recommendation or they forget too much of my listening history.

I get constant metal versions of pop songs. Winterplay's Billie Jean was recommended to me for weeks on end. This week I have Turn down for what meets metal.

Seems like a small tweak to let us tell them we dislike this style of music. According to the article a -1 should be able to be put in that matrix (if that's actually how it works).

Also, I dislike that I lose music after a week. Which means I've lost a few good songs from their churn that I forgot to save.

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

#223

Post by GasketDan »

ikornaselur wrote: post #90
arkadiytehgraet wrote: post #80

If there is at least one song that I like in Discover Weekly, then I consider it to be a good mix. However, this happens once a month at best. After reading the article and all three methods for recommendations, I cannot understand how on Earth it comes up with its suggestions. Do I feed it bad data? Does it not take into account the relative "weight" of a particular track, i.e. my preference to listen to it multiple…

Yeah, I was surprised when I read the start of the article. On Mondays, I will start the day by playing my discover weekly playlist and I'll usually have it running throughout the day, saving 1-4 songs from it to my library. Sometimes none. I just checked and this week I didn't even save a single song for example. I guess it doesn't help that I do listen to really different types of music depending on what I'm doing.…

You can follow and listen to other people's Discover Weekly mixes. I listen to Tobias Van Schneider's Discover Weekly more than my own.

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

#224

Post by confounded »

chrissnell wrote: post #68
cr1895 wrote: post #54

Earlier quoted context omitted.

>“Didn’t I” by Darondo seems to be a good example of this. A fantastic, forgotten soul song (and from the Bay Area!) that was assumed to be a recognition of my friend’s unique taste in his car... until the other four people inside revealed they’d all had it in the last two weeks, too. It could be popularized by DJs who are playing it, e.g. Four Tet, whose personal playlist on Spotify now has close to 39 thousand foll…

It was also used in “Breaking Bad”, in the scene where Walter sets “Ken Wins” BMW on fire. Thomas Golubić was the music supervisor for BB and Better Call Saul and these shows’ playlists make superb additions for Spotify listening, in my opinion. They’re chock-full of great discoveries like Darondo. It doesn’t surprise me that “Didn't I” comes up in many playlists, but it does surprise me that the other great songs on…

Good context, thanks both!

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

#225

Post by zkomp »

I wish there was a way to turn this feature completely off. I used to like Spotify but lately the experience has been constant bugs and to be frank, insultingly bad recomendations... stop this nonsense!

Every time I open Spotify this crap shown on the top, before my own lists, it makes me angry every time.

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

#226

Post by ben174 »

nobleach wrote: post #172

I find that my music tastes can vary too much for Spotify. While working, I mostly listen to Electronic music. But I can also take a hard right turn and listen to Hall and Oates... or Fleetwood Mac, or Steely Dan. Then some form of 80's throwback. But as far as recommendations go, I really only want to expand in that first genre of Electronic Music. I don't want Jethro Tull to be recommended to me. My feed used to be…

The importance of a good Discover Weekly has caused me to be super 'careful' with what I listen to. I have a separate Spotify account on my family plan where I listen to stuff I don't like. For instance, it's connected to my Sonos so guests at my house can listen to whatever they want.

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

#227

Post by zkomp »

KozmoNau7 wrote: post #124
mixedbit wrote: post #121

Earlier quoted context omitted.

My friend likes and uses the Discover Weekly and to get good recommendations he uses different Spotify account when he plays something for his kids or plays some generic background music at home.

He can use the "private session" option for that instead.

I have this problem also, I know of private session and afaik it is not supposed to be used to hide the kids music from algorithms, rather the other way around: hide explicit music you listen to from your kids and coworkers etc...

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

#228

Post by systemride »

steve-benjamins wrote: post #89

As an artist, Discover Weekly has been the best thing to ever happen to me. Every Monday I get a big infusion of listeners (around 5,000)— many of who stick around and check out my other music :) Prior to that, the best press I could get was the tedious process of cold-emailing bloggers (a practice which is now dying off).

Never thought I would find a new artist on hacker news that I would like. Purification ritual is great!!

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

#229

Post by youareterrible »

emodendroket wrote: post #147
slhck wrote: post #119

Earlier quoted context omitted.

How exactly? — Ah, figured it out: - Go to the DW playlist - Click the "..." options button - Then to "Go to Playlist Radio" - "Follow" the DW playlist radio - You can now up/downvote songs

Wow, what a great user experience.

If you had it all on screen you would talk of button fatigue. Users are the worst.

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

#230

Post by jonknee »

ben174 wrote: post #226
nobleach wrote: post #172

I find that my music tastes can vary too much for Spotify. While working, I mostly listen to Electronic music. But I can also take a hard right turn and listen to Hall and Oates... or Fleetwood Mac, or Steely Dan. Then some form of 80's throwback. But as far as recommendations go, I really only want to expand in that first genre of Electronic Music. I don't want Jethro Tull to be recommended to me. My feed used to be…

The importance of a good Discover Weekly has caused me to be super 'careful' with what I listen to. I have a separate Spotify account on my family plan where I listen to stuff I don't like. For instance, it's connected to my Sonos so guests at my house can listen to whatever they want.

Something like an "Incognito Mode" would be a good idea for Spotify and similar services.
Post reply on HN