AI recommendations fail fans who like hard rock and hip hop – official science People who listen to rock or hip hop are harder for music recommendation algorithms to please, according to a study by machine-learning experts and data scientists. The problem appears to be this, we're told: People who prefer so-called hard music, which covers a wide set of genres from hard rock and punk to hip hop, aren't …
As a fan of Hip Hop and Metal I disagree. In fact I completely disagree with this statement:
"People who prefer so-called hard music, which covers a wide set of genres from hard rock and punk to hip hop, aren't much interested in music outside of their niche."
Total bull. My favourite band is Motorhead but I regularly listen to everything from metal to house, classical, jazz, hip hop and folk music. I think I've seen Steve Earle perform as many times as I have Lemmy.
I'd suggest the problem AI has with recommending music is music taste is exceedingly personal and not explainable in machine terms. I couldn't explain why I like the bands I do but I definitely like them and dislike others.
I take more offence to the next paragraph, by calling that bland stuff "music". It is not, and they know it. It might be entertaining to some, I do not doubt that. There are all kinds of people in the world, and wouldn't it be boring if we were all the same etc.
Much like an operette not being great art, but entertaining for sure (a guilty pleasure, I guess).
> "If none of the people on the stage are actually playing a musical instrument then it's not music."
That's a bit of a crude definition. It rules out choirs singing Allegri, a capella groups like Sweet Honey In The Rock and bands famous for their particular choice of instrument who sometimes forget to play them!
> Much like an operette not being great art, but entertaining for sure (a guilty pleasure, I guess).
You are the very model of a modern metal listener
Your favourite band is Motorhead but aren't afraid of Westenra.
You know the songs of ABBA and you quote the charts historical
From Voulez-Vous to Waterloo in order categorical.
Absolutely! And it reminds me of back in the 1980s listening to the Radio 1 John Peel show late one night when in between the "usual" stuff that daytime radio wouldn't go near, he played Sheena Easton's Nine-To-Five - explaining that he couldn't understand at all why he liked it, but he did, so he played it!
This reminds me of this investigation by Stereogum - why have some tracks that weren't even singles, B-sides in some cases, become the most listened to for some indie bands?
https://www.stereogum.com/2105993/pavement-harness-your-hopes-spotify/columns/sounding-board/
Quick answer is.. "auto-play" picking these because they sound more like mainstream songs..so yeah it totally chimes with this article.
Is this strictly a "hard music" issue, or is that judgement merely based on the limited musical horizons of the researchers? Suppose I like 15th century motets of the Flemish but not the Italian schools, some but by far not all by Schubert and Joni Mitchell but only early Joni Mitchell (these not being being "bland" but certainly not "hard" either), would the AI be able to cope?
Amazon always thinks I need more of the same that I already bought. Netflix cannot understand that I want to watch different genre movies and TV series, not more of the same I watched latest. Spotify recommendations are also bollocks as my taste spans a whole lot of music styles.
Agreed. Spotify things I like "Fallout! Boy" because I like "My Chemical Romance". They are sort of audibly similar, to a dumb computer, but not to squishy humans.
However, given enough time and enough data I suppose they might feasibly be able to find some great music for me. But not there yet.
I like Alt-J, Metallica, Tchaikovsky, Caro and hundreds (or thousands?) of other things.
Some of the things I like are "hard" Queens of the Stone Age (and I even like some Hip Hop from Public Enemy to NF).
Saying I don't like to listen outside my genre is utter rubbish.
Just recommend stuff that's "Good" because that's all I like to listen to.
But it has to be subjectively good (to me I mean) and not necessarily what the next person likes.
And some people (many people I think) don't like something because it doesn't fit into their self perception. I happen to also like some Britney Spears songs, in addition to the above... so good luck recommending things to me. But I know if I like it when I hear it.
The study should rather say that people who listen to "hard music" don't like the yoghurt garbage that other people are fed with by mainstream media. These people are often much more open to different styles of music, they just don't like the commercial $hit produced by the so-called "music" business.
I have been asked if I wanted to see a few cover bands, so far I have rejected a Thin Lizzie tribute and a Bon Jovi tribute.
Not a fan of tribute bands, but I am old enough to see quite a few originals back in the 80s.
I prefer my memories of live versions of Thin Lizzie songs to have come from Phil himself.
So it's easier to recommend songs to people who like the dull manufactured tripe that passes for pop music? Well DUH! You just recommend another dull manufactured piece of tripe and there is plenty of that available. It's much easier to recommend stuff to people who like all their music to sound the same.
A decent play list for me would be Peter Gabriel, Bolt Thrower, Celine Dion (in French), Bach, Motörhead, Rush, Hans Zimmer, Hawkwind, The Prodigy, Yes, and Nile's excellent Rameses - Bringer of War to finish, but all that disparity is just going to give the AI a nervous breakdown. <LOL>
If you're happy to be spoonfed "recommendations" generated by an algorithm for your listening choices, you'll have to be happy with whatever you're given.
If however you make a bit of an effort, you'll be able to find the good stuff, whatever your tastes.
Frankly the simple Youtube recommendations are better, because they're generated by seeing what other people who listen to a piece of music on there also choose and seek out, so the elements of personal taste and music knowledge get incorporated in the data.
This post has been deleted by its author
There's your problem.
Now if you'd mentioned shows like John Peel who played what he wanted you to hear, that's more like it. I think it was Tommy Vance on Capital Radio (maybe in the days when it was on 539) who had similar free rein on his Saturday night show. If he decided he was in the mood for a bit of Chicago [Transit Authority] then this was good reason to stay in with the tape recorder running. This morphed into TV on Reggae, which was arguably the first mainstream UK station to devote a programme to the genre and its variants. Nicky Horn had a similar half hour slot at the end of his programme which introduced me to the likes of Pink Floyd and Traffic. The free rein curation by someone who is seen as an authority, without pandering to what management dictate, is what people respond positively to.
Audioscrobbler did a great job simply because it correlates the choices of people who happen to listen to the same artists at the same expanding their taste. I don't know how much that can be improved by AI but then again I don't do AI.
As a dnb fan which has as many spectrums as all other genres combined, I myself am into the stuff that's as dark as a traffic warden's soul and rough as your missus' chest hairs, quite specific, yet whenever I was recommended something like Fear Factory or Machine Head was always spot on. And surprise surprise the latter borrow from heavy metal drum patterns and so on. Seeing all my fave artists (metal and electronica) at gigs and talking with other fans also confirms that the better half of my peers are indeed into the same stuff including off-genre inspirations. (With hardcore metalheads being perhaps a bit more close minded to other genres as expected).
Basically fix your AI, surely there's a pattern that can be applied in training based on experiences like mine over the last 15 years.
Live music was the catalyst for a lot of people in the last century.
Those that sat through some horrendous support bands were sometimes rewarded with an absolute gem. I'm sure I've mentioned this before, but I went to see John Martyn at Sadlers Wells. He had not one, but two support acts, so you'd think maybe a trip to The Shakespeare's Head across the road was quite in order (which is probably where you'd find him, pre-gig). Tanita Takiram was more or less boo'ed off stage for her curled up pork sandwiches. But then on comes another solo guitarist, no fanfare, no introduction, no waiting for hush to descend, she just started to play and after a few bars we were entranced. People in the audience were whispering "who is this?" and very few people knew it was Tracy Chapman. History in the making, as it were. How any AI app can connect Martyn and Chapman would be anyone's guess, but I suspect a lot of people who saw Martyn on that tour still appreciate Tracy Chapman too. In those days it was probably agents ringing round "I need someone, anyone to support a Glaswegian folk singer at a gig at Sadlers Wells."
A lot of the so-called rock radio shows would get their inspiration from the many pub rock venues (Red Cow, Golden Lion, Hope & Anchor, etc.). I believe that's how e.g., Dire Straits, The Police and Tom Robinson found fame. Who remembers bands such as (and I quote) "They Shoot Horses Don't They?" I was into a band called Bethnal who did a ground-breaking (literally) version of Baba O'Reilly featuring a violin solo (when played live) which was streets ahead of the original (IMHO). Doesn't sound much on YouTube... in fact, please don't! But in the pub, or Hammersmith Odeon (which they did get to play) you can truly immerse yourself in the atmosphere.
One of the problems of today is "Listener's Digest". Who listens to an album in sequence from first to last track these days? I think this is one of Neil Young's beefs for temporarily, at least, pulling his back catalogue off music sites. Digressing slightly: who has watched The Italian Job from start to finish? Most people just want to see the chase, and once you've picked that out of the highlights, why bother to watch the whole thing from the beginning? This has been going on for years with classical music, but rock is arguably now the same.
A lot of the true rock classics don't fit into a standard single slot "Stairway to Heaven" and "Layla" immediately come to mind. Maybe why Mountain's "Nantucket Sleighride" rarely got air play.
Sorry, now where are my meds?
I got in to Pandora decades ago and was really surprised how well the algorithm worked, recently I used Google Play Music which again was really good at picking similar artists and on a few occasions I was genuinely surprised at who it put together but it did work, I'd even found new artists I now love from this.
Since being forced to switch to YouTube music though feels a massive step backwards as it clearly uses a new algorithm that is just appalling shite in comparison and feels like it's directed more by sales targets than actual preference matching. I appreciate a band like Skunk Anansie are hard to classify over the course of their repertoire, but to put the likes of Ed fucking Sheeran, Bruno Mars and a One Directioner gone solo is unquestionably wrong and also doesn't factor the other listening tastes I may have expressed on previous selections and playlists.
It's so bad that after years of being a happy Google Play Music customer I've switched to Spotify (that YouTube Music's UI appears to be "inspired" by). Whilst not great (it takes too much of an inspiration from my previous playlists so might through a heavy metal song in to a electontica playlist) it's certain not tried to play Ed fucking Sheeran at me..
The Register 
Biting the hand that feeds IT
Copyright. All rights reserved © 1998–2024
