How to train simple AIs to balance a double pendulum
ฝัง
- เผยแพร่เมื่อ 4 มิ.ย. 2024
- Previous video: • How to train simple AIs
XPBD article: matthias-research.github.io/p...
Github github.com/johnBuffer/Pendulu...
Music used (in order of appearance):
- Winterbeams by Diffie Bosman
- Empyrean by Dear Gravity (4000 pendulums part)
- Poison Message by Man with Roses
- The Endurance by C.K. Martin - วิทยาศาสตร์และเทคโนโลยี
Imagine training for 46 years and the god says, "naaaah, you wiggle too much"
i know RIGHT
I feel like the evolution of quick oscillations was frankly quite cool and made sense with such a chaotic system
"but, but i did the task sir!"
*nope echoes in distance*
I would argue with him that in such a case "well, you made me so are at fault". Unless of course I had be given free will?
😂😂😂😂😂😂
I went from "Only a fool thinks he can balance a double pendulum" to "Praise the lord, he walked on water!"
irl it's impossible so you were right
@@extremechimpoutSo your claim is that all YT videos with robots doing that are fake? Even with triple ones.
In real life it does not wobble as much, everything has friction. It's actually easier irl.
@@MrHardzio4Fun I looked there are no such videos
@@extremechimpout I looked there are many of those video, even quadruple.
@@extremechimpoutLook harder.
My toxic trait is believing that I would be extremely good at balancing a double pendulum with zero practice.
That would be very impressive :D
@@PezzzasWork Your videos are always so clean and well put together. I've finished the video now and the final solution you came up with is great. The idea to incrementally ramp up the difficulty was a stroke of genius that felt so obvious after you said it. Great work showing that paper what's possible!
@@CriticalMonkey623 yeah maybe if you could somehow do that irl you'd actually be able to do a double pendulum balance
If single pendulum is like balancing a broomstick on your finger, then double is like balancing a broom stick on top of another broomstick, on top of your finger. Suddenly it does not seem so easy.
@@anonanon6596 nah i could do it
The idea of starting with lower gravity and higher friction is simply genius
Yeah he immediately solved a problem that even that professional research paper struggled with, and it intuitively makes perfect sense how the learning can transfer over from the simpler problem to the more complex problem. Goes to show that sometimes all you need is to just reframe the problem.
@@research417If you google it, you’ll find that it was a project report from an undergraduate machine learning course…
It’s still a cool approach, though.
For real, when I saw the research paper I thought Pezzz was going to show something else because the “professionals” said it was not possible. It’s amazing how it just took a small change to make it work. Great work pezzz
ingenious
the agent UI/visualization, the training graph hyper parameter view, the movitated loss function explanations.. all so well done wow
Thank you!
@@PezzzasWorkcan you make a video doing those?
@@gus2603absolutely support this 🎉, i want to avoid looking at numbers as much as possible.
Is it made using Python?
Everything is made using C++
The idea of gradually increasing gravity and reducing friction is just genius! Would be interesting to see if this method also would work for a triple pendulum 🤔
boyo do i have some news for you
"World's first video of 56 transition controls for a triple inverted pendulum : 3-body problem"
@@andrewferguson6901 That video is not using an AI controller, it's a controller calculated from methods found in control theory.
@@andrewferguson6901lmao i found the video and apparently i already started it at some point
Seems like engineers doing that for years th-cam.com/video/cyN-CRNrb3E/w-d-xo.html this one some years
Woah there buddy thats enough
it was both surprising and entertaining to see the AI doing the blender move whenever the pendulum drops like a kid throwing tantrums
It's less "doing a move" and more potential energy being converted into kinetic. If the pendulum falls from its highest position it will be moving fast without a correction.
Since it is more stable lower down, it will tend to spend more time rapidly spinning below the axis that slowly spinning above it, unless intervened.
And finally, the AI, like any control system has a maximum frequency/speed it can manage, above which it breaks down and essentially become random input.
So while the AI might be able to take the slow pendulum and balance it up high, if it then falls, it might not have the skill needed to recover. So every failure mode leads to a rapidly spinning "blender".
@@daniel.lupton i know how the doube pendulum works and i also know the situation is either what you said or the AI gaining reward through some loophole in the reward system since its technically over the parameters once every loop (thought this because blenders became less of a problem with each added condition)
But i just wanted to make a visual comparison between malfunctions of AI and toddler brains where they start giving seemingly nonsensical inputs to solve their impossible problems (which is your statement anyways)
@@xwxwvyz1 Yeah sorry I didn't mean to imply you didn't understand it. I think I was just looking for an excuse to explain why the blender thing happens.
I didnt even think that it would be possible, wow, also nice editing and everything is just put in a way that makes the video really enjoyable 🎉
Thank you very much!
All AI are essentially just function emulators.
@@PezzzasWorkDefinitely a cool video.
If you end up doing a follow-up, here are a couple of things I would like to see (as a control theorist):
- test with random disturbances that also take place during the swing-up control. You could also think of this as adding a little noise to the inputs of the NN. I think you would have to include this in the training as the swing-up didn’t look like it would tolerate any sort of disturbance.
- It would be nice if you gave the audience the ratio between maximum acceleration and gravity used.
- I definitely agree that it makes more sense to use acceleration as the control input.
@@PezzzasWorkalso, I think that “article” you mention on NN-based control of a double pendulum is just a project report from an undergraduate ML course.
Finally, if you’re unable to make your NN solution robust to disturbances or noise during the swing-up, you could use the ML approach for the “feed-forward” control, basically as a planned trajectory and add conventional control theory on top to stabilize the trajectory w.r.t. noise/disturbances.
The next step is teaching an AI to balance a million pendulums.
Keep up the great work, it's both educational and entertaining!
we just need a million dyson spheres so the model training doesnt take 100 years
nah dont need a million just the magic number 7
Do you mean balancing a rope? 😂
@@strecher777 no but I will share it when the time is best!
We need to balance the n-pendulum
Your videos are always a delightful blend of programming, puzzles, science and philosophy. Thank you for making YT a better place
Thank you very much!
9:03 "its a little better"
AI: starts swinging the pendulum around like crazy
keep up making this content!
I will try :)
Very nice!, I really enjoyed the first video, it helped me create my own NN. Keep up the good work
Thank you!
It's so interesting how friction and low gravity helped the AI learn. It's like humans practicing something complex in simpler variant, because trying to do something like balancing a double pendulum would completely overwhelm us too having no idea about how you would even start counteracting the imbalances
👌🏼 Magnificent.
It was really great idea to use air friction as the simplifying factor.
I world love to see the cost of work in the evaluation function.
Love this approach of setting up an easier environment, for the ai to learn from and slowly increasing the difficulty. I would also love to see how you are going to explore this aspect in the future :D
the progressive difficulty was an amazing training tool! very well put together video
starting easy and gradually increasing the difficulty is a very clever idea.
The editing and the feedback UI are brilliant. Cheers, Pez.
It's amazing how you go above and beyond. After that first working solution, I would've been really proud of myself and be done with it, but you just make it better and better.
give this man a Nobel prize! respect!
love the visuals, the clean colored borders... perfect!
This is amazing! I have been waiting for this second part and its finally here! Its amazing that you can produce a video with such a good quality and informative in such short time!
The UI and graphs for this video look amazing! Keep up the good work!
You made it absolutely wonderous and intrigued interest to the notion of chaos. Well done
Amazing work. It was _really_ engenius the way you break up the problem for the evolution algorithm! And the interface you put together to demonstrate the work just adds the cherry on top of it. Keep up the great work, I'm looking forward for the next one!
Your work never ceases to amaze me
Amazing video as always, I love the visuals on this one! Greetings from Paraguay 🇵🇾
Awesome! I've been looking forward to this video since I saw the first one. Did not disappoint
The UI is incredible and everything is very well presented!
I haven't worked with neural networks before, but you definitely inspired me to start looking into it and giving it a shot. Well done with your solution, and good work for not giving up!
I've been obsessed with double pendulums for a long time now, I absolutely loved this video. I honestly started the video thinking it would be impossible to balance.
I just LOVE your interface
I'm in love with the interface you designed!
I'm currently making my own neural network and your videos really helped me understand the concepts and mechanisms of this project
This is incredible, such amazing work
Wonderful, you have mastered the art of getting what you want from a neural network!
That pendulum rainbow animation is straight art
this is such an amazing visualization and intersection between so many beautiful forms of mathematics. thank you for this @Pezzza's Work
Love your ideas and visuals! 🥰
loving the visuals, great video!!
I love the interface, great video!!
Really amazing implementation! The first moment I saw the jerk movements I immediately blamed floats. I was very satisfied when you confirmed it, and decided to switch to doubles, as I was screaming at my monitor. :D
Great video. I love watching the refinements. Human and AI learning together.
This video is so beautiful, great work!
Beautiful visualizations!
Amazing content would be an extreme understatement.
This video is simply fascinating! Seeing how an AI can learn to balance a double pendulum is truly impressive. The explanations are clear, and the visuals are very well done, making the subject both understandable and captivating. I especially enjoyed the demonstration with different weights, and it made me want to try creating my own AI to tackle this challenge. Keep producing such inspiring and educational content, it's a real pleasure to watch!
Buddy graduated with a phd at 17. Subbed earned.
Incredible work!! Amazing video!
Really nice! I liked the velocity-commanded version better than the acceleration commanded one. In many practical cases that is also found to matter; 'in theory' equivalent neural networks should exist but controlling acceleration does bias towards smooth accelerations indeed. In practice a motor controller often has an internal PID control loop; so commanding a velocity (and having that quickly realized, up to physical constraints on acceleration; something like min(max_motor_accel, velocity_error/dt)) isnt necessarily an unnatural choice. It still allows for spiky torques to be learned where they are required in a natural manner.
beautiful vid! keep up the good work. subbed
Just subscribed. You're a genius
Amazing work
the poor guy trained for 46 years, stuck inside the matrix and you said there still room for improvement...
Loved the video, the visualizations, the breakdown and the thought of using not python for everything is whats impressive, immediate sub, hope ill learn how to do this stuff
fantastic! this has so many uses! hope you publish code soon! I have a few ideas what to use it for! I worked for a brief time on this exact problem!
Really great video and visualisation. Please add some tests next for pendulum setups that are already in motion to see if it can also stabilize these and not only from resting position. This will be fun!
Honestly I thought it would be impossible, you're a real beast !
Amazing work :-) I will send this to my students in a couple of years when I'll teach intelligent automation.
This is the kind of top tier content youtube was made for!
Such a beautiful project!
Yess! I was waiting for this video :D
You can probably improve things by giving the two pendulums a random offset at the beginning of each run instead of having them both straight down. You can even start off with a very tiny offset +/- 1degree, then steadily increase the offset range over time. Then maybe the system can learn to start from any starting position? Also you can do the same thing with initial angular velocities
as AI most generaly does not impress me, your work and dedication does. also i love the attention to details like graphics, that s truly amazing!
Awesome video! Really shows the complexity in tasks like this, and brings into light the massive challenge of doing this in real life with double and tripple pendulums. I wonder how they managed to get such good control!
I also find it very interesting how it hits a wall every once in a while than makes 1 small change and just doubles it performance. Thats really interesting!
This video is truly beautiful. And interesting too!
The idea to increase difficulty was so good. Loved watching it go through periods of great success and fast difficulty increasing.
phenomenal video
Amazing video, gives great insight how to make AI tackle seemingly impossible task.
Increasing the difficulty is a cery good approach, if I ever make a neural network I will probably try it.
Watched the first video, was amazing. Watching this. Hopefully it's even better
Triple pendulum when?
Look up "World's first video of 56 transition controls for a triple inverted pendulum". It's the world's first video of 56 transition controls for a triple inverted pendulum.
@@themoran2 wow
@@themoran2 It can't balance it as long and consistently as this video's double pendulum. Seeing the AI balance it for much longer than 15 seconds would be nice.
It's still really impressive though and I was on the edge of my seat seeing a triple pendulum being balanced.
@@Thomas-ko1nf Seeing as how it's a triple pendulum balanced in real life, it's hardly surprising an AI in a simulated environment is more stable.
Then again, it's the world's first video of 56 transition controls for a triple inverted pendulum, not the world's first video of a triple pendulum perpetual balancing act.
@@themoran2 That video is not using an AI. It's from a lab researching control theory. They developped their own method based on control theory.
Starting from easy conditions and gradually making them complex is simply a stroke of genius! I bet this will be a thing in the near future. You should write a paper
Beautiful as always! :-)
This video is beyond impressive 🤯
One of the best videos I've ever watched about artificial intelligence, please make more videos like this, this was incredible!
J'attendais cette vidéo avec impatience !!
That's very great ! Even if the numerical scheme can hide some dissipation, it is the same for a near-perfect physical system.
its interesting that in the final result you can so clearly see oscillation at two different frequencies in the output velocity thats cool
you should've add to the score function how are the three points are aligned to avoid exploids such ase the last solution you showed us, also i think that we would like to see the result of the training method of controling the acceleration using the last score function when you was still giving the ai control over speed of the cart not the acceleration, and in the end it was a good video, and it was pretty informative and entertaining in the same time!
Ok, that's hella impressive. I fondly remember visiting "Hannover Messe" with my parents. A convention with a lot of technical stuff. There was a booth where they had a pendulum with a flywheel on the end that automatically uprighted itself and balanced after that. It's stuff like that and what you do here that inspires young people to make stuff like that. Thank you! (And now build a machine IRL that replicates the double pendulum simulation and is controlled by the neural net you trained in the simulation :D)
Man I love this content, I am young and want to know more about this, thank you!
WOW! New favorite video of yours
Just finished my first implementation of a standard nn with back propagation. I'm 16, so it took a bit, but I'm proud of it. Right now i have it fitting a given function, but it is pretty versatile.
Very impressive. Super useful for my work.
The exploit solution at the end is amazing
I feel like you now should write the scientific paper on this. Going from "This problem is yet to be solved" to "I solved it with about 2 hours of training for my model" is huge progress!
amazing work!
Oooh, you're real good
I would've thought it was good enough as soon as it was able to balance at all
that is amazing! thank you
Great work Pez ;)
Very nice. I'm most interested in learning how you visualize things.
i CAN spend hours watching 4000 pendulum trajectories that form moving geometry. It's just amazing!
mad respect, whole scientist team can get your interview to update the study 😂
To give it a greater intuition for how to recover the pendulums from any chaotic state, you could start each simulation with some applied random motion, instead of hanging dead center.
Trying to learn how to stabilise chaos! That’s a pretty steep goal Pez !
I love the video, especially all the graphs. Can you make a video about how you did the animations and what software you used if any. :)
oh wow, the training methodology is amazingly smart
Or perhaps, the next step would be to balance the double pendulum in 3D space? Each joint can rotate freely in any direction. That, or balancing a triple pendulum in 2D, I couldn't miss!! Anyway this was extremely enjoyable and a wonderful achievement! My favorite video of yours!