Slashdot Mirror
AI Bots Pick The Hits of Tomorrow
Wolverine Inspector writes "The Music Industry uses a product called HSS (Hit Song Science) made by Spain's Polyphonic HMI. According to The Guardian "while no one's talking about it, it seems that the whole record industry is already using AI to choose hits. From unsigned acts dreaming in their garage, to multinationals such as Sony and Universal, everyone is clandestinely using a new and controversial technology to gain an edge on their competitors."
Even though it costs about $5,200 US/$6,500, many artists are starting to buy it to help them write succesfull songs."
That's just great.
Remember how video card manufacturers were tweaking their drivers to perform well in benchmarks? "Musicians", and I use that term loosely, will be tweaking their songs to score a "hit" on this service. Right, but it will be harder than ever to produce something out of the mainstream when a record exec will look only at the score on HSS and potential effect on the bottom line. Art for art's sake is virtually a thing of the past. Prepare for more of the same on the FM dial! (thank goodness for etunes.com)
Trolling is a art,
Remember the good old days when the listeners picked the hits?
Next up: bots that generate pop music.
Air Supply and Ashlee Simpson.
No mod points, no meta-moderating/Firehose/all the other free work Slashdot wants me to do.
This is not AI. The music companies are using clustering technology.
The basic idea is that you measure certain characteristics of a song,
such as voice quality, cadence, etc. I'm sure the actual
characteristics used are much more complicated, but the idea is the
same. Once you have your characteristics you can build a three
dimensional vector out of a song. After you have your three
dimensional vector, you can then use many different algorithms, one
such is the Bi-secting K-means algorithm to group the songs together.
After you have built your cluster, you take a new song, run it through
the process and check to see how close it falls to a "hit" cluster.
We use this same process for document classification at my work, and I
don't think it bears any relation on AI. As I stated above, it's a
rather simple grouping technique.
There is a downside to this technology though. By measuring how close a
song is to previous hits, you are guaranteeing that all new songs will
be similar to old hits. This type of system tends to minimize or
eliminate fresh new types of music.
(why the word wrapping? Emacs auto-fill-mode)
Doug Tolton
"The destruction of a value which is, will not bring value to that which isn't." -John Galt
many artists are starting to buy it to help them write succesfull songs."
Comercially successful != good
---
"I can't complain, but sometimes still do..." Joe Walsh
Is it just me or did the article quote music industry folks as saying the software must work becuase 95% of the hits of the last decade scored highly. The software is a mathmatical model based on the hits of the last century.. so of course it scores them highly.
The program works by applying the formula. It takes three variables.
Boobs
The artist must have boobs. The larger they are, the higher this value.
Blandness
The blander it is, the higher this value.
Beat
The stronger the beat, the higher this value.
These are multiplied together.
B * B * B = X
If X is greater than or equal to the Olivia Newton-John quotient, a recognized standard throughout the popular music business, the song will be a hit and we release an album.
If X is lower, we don't do one.
Q: Are there a lot of these kinds of artists?
You wouldn't believe.
Q: Which record label to do you work for?
A major one.
Unknown host pong.
This announcement from the producers of this record contains important information for radio program directors, and is not for broadcast.
The first cut on this record has been cross-format-focused for airplay success. As you well know, a record must break on radio in order to actually provide a living for the artists involved. Up until now, you've had to make these record-breaking decisions on your own, relying only on perplexing intangibilities like taste and intuition. But now, there's a better way.
The cut that follows is the product of newly-developed compositional techniques, based on state-of-the-art marketing analysis technology. This cut has been analytically designed to break on radio. And it will, sooner or later.
For the station that breaks it first, the benefits are obvious. You lead the pack. Yes, no matter what share of this crazy market you do business in, no other release is going to satisfy your corporation's current idea of good radio like this one. On this cut, we're working together, on the same wavelength, in scientific harmony.
But remember, this cut is constructed for multi-market-breaking NOW. Don't waste valuable research with needless delay. We've done the hard work of insuring your success; the final step is up to you.
SPECIAL DESIGNER SONG FOLLOWS IN 5.. 4.. 3.. 2.. 1.
In a related story, the IRS has recently ruled that the cost of Windows upgrades can NOT be deducted as a gambling loss.
This was discussed last November, which was a repeat of the same tech from February.
A quick search for "polyphonic" in the music category would've easily picked this up, they're the only 3 matches!
- The first step in the process for our technologies is to analyze a representative sample of music (up to date we have
- analized more than 1 Million tracks)
Analized? Analized? - what dedication these folks have. Brings tears to my eyes.The label's marketing department are promoting him to the Norah Jones audience. But Polyphonic's analysis has shown that the crooner's song patterns are more similar to Linkin Park, Aerosmith and JayZ.
future HSS developer: You know who I really hate? The record industry.
future HSS collaborator: Well, you should do something about that.
future HSS developer: You're right! Recording execs are really, really, stupid. I bet it'd be easy. I've got a plan.
future HSS collaborator: Sigh... fine, what's your plan?
future HSS developer: They pay us $6000, and we tell them if their song will be a hit or not, then give them some printouts with, you know, clusters of dots on them, random numbers, whatever. Then we say "Artificial intelligence! The magic boxes say this will be a hit because it resembles Tupac Shakur and Wagner!"
future HSS collaborator: You know, unlike your plan to hack people's PVRs to make them think they're gay.... this would actually work. Let's do it. Get me a dartboard.
The good and new comes from no quarter where it is looked for, and is always something different from what is expected.
This sounds a bit familiar... I think the RIAA stole this idea from Southpark. Cartman dressed up like a robot (AWESOM-O) to get secrets from Butters... but he ended up in Hollywood creating blockbuster movie ideas. Here is how it went down:
...golden retriever, or something. ...boxer, or something. ...Yes, it's flawless!
Mitch: Punch-Drunk Billionaire!
Producer: Gentlemen, this little boy was kind enough to let us show you his robot. The AWESOM-O 4000. [approaches the robot, who's seated at one end of the table] I've already seen what he can do.
Staffer 1: Uh, excuse me sir, but uh, that's not a robot.
Producer: It's not?
Staffer 1: No, it clearly had bipedal movement, so the correct term is "computerized automatron."
Mitch: Oh, very nice, Mitch.
Staffer 2: You are the smart one.
Producer: Well, regardless, I believe maybe this automatron can help us come up with new movie ideas.
Staffer 2: How can a robot come up with better ideas for movies than us?
Producer: Watch this: AWESOM-O, given the current trends of the movie-going public can you come up with an idea for a movie that will break a hundred million box office?
Cartman: Um... okay. How about this: [the staffers take pen to paper and anticipate the ideas] Adam Sandler is like, in love with some girl, but then it turns out that the girl is actually a
Staffer 2: [thinking over this idea, then write it down] Oh, perfect!
Staffer 3: We'll call it "Puppy Love"!
Staffer 2: Give us another movie idea, AWESOM-O!
Mitch: Yeah yeah!
Staffer 3: Let's hear it!
Mitch: Yeah, we wanna hear it!
Staffer 3: Come on, come on!
Cartman: Okay, how about this: Adam Sandler... inherits like, a billion dollars, but first, he has to, like, become a
Staffer 3: [the producers start writing again]
90% of being smart is knowing what you're dumb at.
Does this remind anyone of the Monty Python skit where they use mathematicians to create the world's funniest joke, and use it to get Nazis to die laughing?
... are intricately related. Many AI techniques are forms of statistical inference or statistical classification techniques. Some neural nets implement grouping techniques not that different from k-means.
Any box which learns from a set of data in order to predict future data by implicitly extracting trends and patterns from that data is an implementation of some form of statistical inference algorithm and is subject to all of the general results statistics has to offer about such algorithms. Conversely, statistical inference algorithms are often implemented in ways associated with AI, for example as neural nets.
Given this situation, it's hard to define the boundaries that separate artificial intelligence, pattern recognition, statistical inference and classification and the rest. Of course, there is a legitimate question as to whether such techniques actually mimic genuine intelligence even in principle, and there are other approaches.
From the point of view of terminology, there is a huge range of techniques that can be called AI, and statistical inference is one of them. If you call a VLSI neural network implementing a statistical inference algorithm "AI", then why not call a normal computer implementing a statistical inference algorithm "AI"? Besides, AI sounds a hell of a lot sexier than statistics when you're trying to extract maximum dough from the ample coffers of the recording industry.
"The Milliard Gargantubrain? A mere abacus - mention it not."
Does this remind anyone of the Monty Python skit where they use mathematicians to create the world's funniest joke, and use it to get Nazis to die laughing?
Actually, it reminds me of some of Bradbury's more gloomy predictions of the sanitization of the culture that was happening then and is continuing to occur.
How long until books are written the same way, or at least evaluated by the same kind of tool? I suppose the news media will have it happen to them first: "Sorry, Dan, that story about political hanky-panky rated a 4.5 on our offensive-o-meter, way above the threshold of 3. Put some kittens in it and maybe we can get it to a 2.5."
--
$tar -xvf
Like LAUNCHCast before the RIAA leaned on them, and then Yahoo! acquired it and ripped out everything that made it innovative and cool?
You'd rate songs on a scale from 'never play this again' to 10, and the system would select new songs based on what you'd already rated and insert them into your personal "station" rotation.
But the real killer feature was that you could search for other users whose tastes were statistically similar to your own, subscribe to their stations, and learn about new and different music and artists as some of their favorites were added to your rotation. Want to buy a song? Click on it.
Absolutely cool collaborative software. Unfortunately, if you wanted to expend the effort, you could abuse the system to constuct a station that could (gasp!) play a specific song at a specific time for free, and the RIAA wouldn't allow that.
So the only thing that had gotten me to purchase any new music in years was eviscerated, stuffed full of ads, and then sold to Yahoo! as a 'service' with all the collaboration gone. You could pay money to lose the ads. Whee.
Bitter? Me?
A short story by Arthur C. Clarke describes one possible consequence of this sort of thing. The scientist involved builds a computer to study the underlying theory of music, harmonic relations, wave analysis, frequency distribution, etc. and how it interacts with the brain on a physiological level. His search is related to the notion that all existing tunes are crude approximations of the fundamental melody that has eluded composers for centuries (basically a rehash of Plato's theory of ideals applied to music.) The scientist is later found in a permanent catatonic state in his lab (by his tone-deaf assistant) with the Ultimate Melody repeating over and over in an endless loop. Because the overwhelming power of the Ultimate Melody (the ideal form on which all melodies in the universe are patterned after), his mind is completely dominated by it--much the same as when a catchy tune gets stuck in your head for days, only much more powerful. The melody formed a fugue in the pathways of his brain, going round and round forever, obliterating all other thoughts.