I’m not really surprised, the main challenge of that game is motor control, something any machine can do with more precision than a human
My jaw hit the floor when I watched an AI master one of the world's toughest physical games in just six hours
Submitted 9 months ago by Lifecoach5000@lemmy.world to technology@lemmy.world
Comments
Taako_Tuesday@lemmy.ca 9 months ago
sushibowl@feddit.nl 9 months ago
I agree but also disagree. It’s true that machines are capable of fine motor control much more quickly and accurately than humans. But this by itself is often not enough.
This achievement should be somewhat surprising because of Moravec’s paradox: the observation that, opposite to what early AI researchers expected, intelligence and reasoning skills are comparatively easy for a computer to simulate, while sensorimotor skills are in fact incredibly hard. Notice how, for example, chess engines started beating human players in the 90s or so, but we still don’t have a robot that can do something as simple as pick raspberries (because surprise, for a machine picking a raspberry is actually hard as shit).
CluckN@lemmy.world 9 months ago
My eyes bursted out of my eye sockets when an AI was able to multiply 8 prime numbers faster than a human.
lvxferre@mander.xyz 9 months ago
They’re calling everything “AI” nowadays… this sort of learning algorithm is old as fuck, here’s a 8yo example. The main differences between both situations is 1) some sensor(s) being used to “tell” the algorithm about the board state, and 2) the barebones robotic arms.
Thorny_Insight@lemm.ee 9 months ago
I don’t get what the issue is calling it AI?
lvxferre@mander.xyz 9 months ago
Even if skipping completely the discussion about what is “intelligence”, the expression “artificial intelligence” has been used as a label for so many different technologies that it has become practically useless. It includes things like decision trees in games (even if a lot of them boil down to simple if/then statements), generative models, even theoretical artificial intelligence. And evolutionary models like the one in the OP and the one in my link.
So it’s basically the 20s version of what “smart” was in the 90s/00s. Like this:
Image
OK, I’m exaggerating with the image macro.ioslife@lemmy.ml 9 months ago
AI implies intelligence. This is just a simple algorithm
TheFriar@lemm.ee 9 months ago
Exactly. Not to mention, why the fuck is it a surprise that a computer twisting the knobs “at superhuman speed” would be better at this game than humans. Like, no shit. We can’t compute how the degrees at which we’re turning the knobs affects the speed of the ball, can’t store that information for next time, and find the best way not making the same mistakes twice. Because…we’re human. We don’t have that finely tuned ability…because we’re not machines.
So…this isn’t “AI” despite the robot hands they put in the thumbnail and no shit a dedicated computer could master this game. I’m surprised it took six hours.
lvxferre@mander.xyz 9 months ago
Additionally, this shit is really easy to compute. It’s all Newtonian physics, and there are only two relevant equations here, both simple: d = at²/2 + vt and a = g*sin(θ). It’s really easy for a computer to reach those formulas, cancelling the advantage that humans would have (insight and actual knowledge of the system).
maniel@lemmy.ml 9 months ago
You don’t need AI to do that, seriously, such a buzzword where a relatively simple algorithm would suffice, don’t tell me it’s harder than double pendulums or those ball bouncing contraptions tech students make since a decade or more
CrayonRosary@lemmy.world 9 months ago
Not needing AI isn’t the point. The point is that AI can do it, and AI doesn’t require a programmer to design and debug a bespoke algorithm to accomplish a task.
maniel@lemmy.ml 9 months ago
Test, you have a point here
squirrel@discuss.tchncs.de 9 months ago
I hope their jaw is alright
Lifecoach5000@lemmy.world 9 months ago
I cringed at the headline but just posted it as is and thought the article was kinda interesting.
tooLikeTheNope@lemmy.ml 9 months ago
It sure won’t when his wife is gonna leave him for her new AI driven dildo. It is just a matter of time
Annoyed_Crabby@monyet.cc 9 months ago
Here, the video the article is talking about. Save you from reading the author’s life story.
db0@lemmy.dbzer0.com 9 months ago
While the link is useful, the smug takedown is uncalled for. Humans relate way more through personal stories like this. Without the story, the video is not impressive at all, as most will have now idea how difficult this achievement is.
justJanne@startrek.website 9 months ago
It’s just like those shitty recipe sites that tell you their grandma’s life story for hours before giving the recipe. Get to the point, who cares about the anecdotes of some writer?
Annoyed_Crabby@monyet.cc 9 months ago
I get that, some people prefer to have some personal story mixed in the article, but personally i’d like to have my time respected, more than 2 paragraphs of that and i’m out. With that bloated life story and a baitest of the clickbait headline, it deserved to be call out.
TimeSquirrel@kbin.social 9 months ago
I'm ASD and I'm also human, Gimme the cold hard facts. Everything else is useless to me.
Kbin_space_program@kbin.social 9 months ago
The point of journalism is to get the facts across and inform viewers. I don't care about the journalist other than them being impartial and reporting on the facts.
surewhynotlem@lemmy.world 9 months ago
Oh yeah? Can it tilt the board all the way to one corner, then pop the other corner and send the ball flying right to the end?
No, it’s amateur at best.
Blooper@lemmy.world 9 months ago
That’s addressed in the article actually. They had to program it so as not to cheat when they found it actually trying to cheat.
JohnEdwa@sopuli.xyz 9 months ago
The true ability of AI/machine learning is to find and abuse all the loopholes and errors that exist in the training set.
“The only winning move is not to play” was simply WORP maximising its reward function.
menemen@lemmy.world 9 months ago
This is pretty much what I’d expect AI to be best at.
INeedMana@lemmy.world 9 months ago
It’s cool but my question is (I did not see this addressed in the article nor video but might have missed it) did it learn to win the game in general terms or only this one example? I mean, if the layout of the board was changed, would it still solve it?
just_another_person@lemmy.world 9 months ago
They don’t discuss it here, but it’s most likely an reinforcement model that operates different generations of learned behavior to decide if it’s improving or not.
It would know that the ball going in the hole is “bad”, and then try to avoid that happening. Each move that is "good’ is then kept in a list of moves it should perform in the next generation of its plan to avoid the “bad” things. Loop -> fail -> logic build -> retry. After 6 hours, it has mapped a complete list of “good” moves to affect it’s final outcome.
The thing about these models is less that they will work (it is assumed they eventually will through trial and error), but how efficiently they will work. The number of generational cycles and retries is usually the benchmark when dealing with reinforcement, but they don’t discuss that data here either.
INeedMana@lemmy.world 9 months ago
Yes, but that’s kind of my point
We see it learn something with insane precision but most often it is almost an effect of over-training. It probably would require less time to learn another layout but it’s not learning the general rules (can’t go through walls, holes are bad, we want to get to X), it learns the specific layout. Each time a layout changes, it would have to re-learn it
It is impressive and enables automation in a lot of areas, but in the end it is still only machine learning, adapting weights to specific scenario
indomara@lemmy.world 9 months ago
It did learn to use shortcuts to skip parts of the maze, and had to be told not to. Super interesting!
INeedMana@lemmy.world 9 months ago
Yes, but that’s only because a generation found some random, specific motion that scored better. Not because it analyzed that doing a skip should be possible
jordanlund@lemmy.world 9 months ago
When the AI can solve one of these I’ll be impressed:
gian@lemmy.grys.it 9 months ago
A blast from the past… Damn, now I have the urge to recover mine… from somewhere in the storage room… if it still exist…
datendefekt@lemmy.ml 9 months ago
Hey, I also had a toy like that! Cool!
MonkderZweite@feddit.ch 9 months ago
Oh! That thing! Takes me back.
gian@lemmy.grys.it 9 months ago
The only thing that is hard about this game is to control the board, which is the concept of it.
dangblingus@lemmy.dbzer0.com 9 months ago
Not sure if it’s more interesting that an AI taught itself the PID instructions in order to deftly move the ball around, or if it’s more interesting if a human programs the PID instructions to move the ball around. Sounds like a lot of electricity was used doing it the first way.
frazorth@feddit.uk 9 months ago
But this sounds exactly the sort of thing that machines are better at that people, so it just feels completely unsurprising that it was good at the task.
Turning multiple dials to manage speed and direction is not normally how humans interact with the world, so we can we pretty shit at it.
A basic motor is completely designed to turn like this.
This feels no different to the machine learning tools used to train on Mario a decade ago.
tsonfeir@lemm.ee 9 months ago
Right. Computers doing the shit that we don’t want to do for a living while giving us time to do things like paint cows that don’t have two heads.
1984@lemmy.today 9 months ago
Technology has removed a lot of time consumin or boring jobs, but it also made us spend our time in front of the computers. The idea from the start was that we could live our lives when computers do our tasks. But we ended up on social media or in front of computer games.