Dingus the Deep reinforcement learning cube learns to push a boulder up a series of increasingly difficult hills
Жүктеу.....
Пікірлер: 1 000
@revimfadli46666 ай бұрын
One must imagine Dingus happy
@Dingus_Labs
6 ай бұрын
Legends say he was happily pushing boulders even before he was put into this strange simulated environment.
@christianottley8542
6 ай бұрын
One must imagine dingus upside-down, its hilarious
@Scratchydoesmusic
6 ай бұрын
@@Dingus_LabsHmmm...
@totrox2445
6 ай бұрын
I knew this comment will be here as i saw the title of the video.
@The_Cheez
6 ай бұрын
Dang it. He beat me to it
@KaufmoTheCIown6 ай бұрын
Give Dingus a top hat that progressively gets taller every video.
@paper177
6 ай бұрын
YES
@paper177
6 ай бұрын
but wait nono one that gets taller everytime he gets a reward
@revimfadli4666
6 ай бұрын
@@paper177 then one no longer needs to imagine Dingus happy, for he is actually happy
@paper177
6 ай бұрын
free hat
@AnUnfortunatePotato
6 ай бұрын
@@paper177is that dimmagingus, owner of the dimmsdale dimmadingus?
@zawarudo4976 ай бұрын
it acts so convincingly distraught whenever it loses the ball
@ShrimpFry_Cute
6 ай бұрын
He lost his ball. Why do you think he’s not distraught. I would freak out if I lost my ball too.
@frowner_and_co
6 ай бұрын
He so sad 😢
@nicolasn12
6 ай бұрын
boulder*
@alansmithee419
6 ай бұрын
It doesn't know what to do so it acts chaotically. Humans do the same. We just also get distressed because our brains recognise the problem. We don't act chaotically because we're distressed, it's just that acting chaotically and getting distressed often have a common cause - we don't know what to do.
@manboy4720
6 ай бұрын
@@alansmithee419 so you're saying that dingus is the next stage in human evolution?
@exLightningg6 ай бұрын
I always enjoy these machine learning videos but at the end I always wish they'd run the trained AI through all the levels/courses again.
@garchamayne4254
6 ай бұрын
Definitely. Want to see the difference between the new and final ai
@mondongoloco7902
6 ай бұрын
Wtf is ur pfp
@arend8785
6 ай бұрын
yeah, or maybe a final level with all the elements from previous levels
@222patapon222
6 ай бұрын
I was coming to the comments to ask exactly this, @Dingus_labs can you please show us Dingus going through all the levels with his new found knowledge?
@exLightningg
6 ай бұрын
@@mondongoloco7902 Oh it's a little fella that appeared on an image I AI generated. Full image: i.pximg.net/img-original/img/2023/02/12/21/34/03/105319111_p10.png
@no-brainer-gamer6 ай бұрын
I never thought I would hear the sentence "There are, in fact, 54 other dinguses training in the exact same level at the exact same time" that caught me off guard
@rennegaddefoxxe
6 ай бұрын
Dinguses, or Dingi? (Din-guy?)
@wrjtung3456
6 ай бұрын
Please have seggs with me no brainer gamer
@frowner_and_co
6 ай бұрын
@@rennegaddefoxxedinguses is an actual word so let's go with that until we get comfirmation
your right! the plural of dingus would be dingi I didn't think about that...@@rennegaddefoxxe
@asshizu6 ай бұрын
It's incredible how fast this ppo setup learns. I've never seen anything like it. Reinforcement usually takes much more trial and error and training time.
@asshizu
6 ай бұрын
What inputs did you use?
@nikkiofthevalley
6 ай бұрын
I have a feeling that there's a couple hundred instances running in parallel.
@imakething
6 ай бұрын
55, he says so
@Dingus_Labs
6 ай бұрын
Yeah 55, he did still learn super fast, I didn't have too many inputs or actions for him to perform which does help with learning but it was still far ahead of what I expected
@skeleton819
6 ай бұрын
@@Dingus_Labsit seems like dingus doesn’t struggle with overfitting, thats something ive never seen on any other ai vid. This is super cool dude
@dariocampanella79926 ай бұрын
I can easly imagine AI taking over the world and having this guy to push boulders for the rest of his life.
@LordCrate-du8zm
4 ай бұрын
One must imagine Dingus Labs happy.
@preistestdragon46346 ай бұрын
when dingus finished all the levels i think it would have been interesting to see you put him back at level 1 and see how all his training thus far would have affected him from the start.
@Junkkie
6 ай бұрын
One must imagine Dingus happy. Or speedrunner
@laula813
6 ай бұрын
Or combine all levels into a mountain.
@eli3963
6 ай бұрын
For sure. I believe reinforcement learning is pretty prone to forgetting.
@Dingus_Labs
6 ай бұрын
There's definitely a risk of overfitting, where the AI has become to specialised for a specific task and struggles to solve more generic problems. Particularly for the last level where he trained on it for a long time and learned to stay very close to the centre.
@tsunertoo9149
6 ай бұрын
Maybe try something where instead of just static stages you set parameters and generate a seed kind of like a rogue like? That way it could potentially force the AI to try and use general information rather than fitting to a stage.@@Dingus_Labs
@sophiekarp15564 ай бұрын
I think I have a bunch of Dingus cubes running my brain
@MatthewConnellan-xc3oj
3 ай бұрын
Well, use reinforcement learning on them for the next 3 years.
@haroldas16 ай бұрын
It would be interesting to see a course where it starts with a path forward, but that path is not possible. It is required for you to backtrack behind the start to progress
@Dingus_Labs
6 ай бұрын
Yeah, depending on how the model has trained and how the reward function is programmed it can really struggle to adapt to situations like that!
@Speed001
6 ай бұрын
So a maze
@vincenttrigg4521
6 ай бұрын
@Dingus_Labs I could imagine it working if it knows things like going from point A to point B (aka start to finish). Basically have it have a target location to go to (like the top of the hill here) and once it does that, the map changes its "finish" location. I'll use emoticons here as a diagram. 🟨 map ⏹️ end location 💗 Dingus 🟢 ball 🟦 obstacle 🟥 off map 🟪 reward location 1️⃣2️⃣3️⃣4️⃣ segmented finishes (Basically larger rewards or finish locations that change places once reached) Part 1 🟥⏹️⏹️⏹️⏹️⏹️🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟪🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟪🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 2 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟥🟥🟥🟥🟨🟪🟨🟥🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟨🟨🟪🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 3 - round 1 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨1️⃣🟨🟦🟨4️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟨🟨3️⃣🟨🟥🟥 🟥🟥🟨2️⃣🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 3 - round 2 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨1️⃣🟨🟦🟨3️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨2️⃣🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 3 - round 3 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨1️⃣🟨🟦🟨2️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟨🟨🟪🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 3 - round 4 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟦🟨2️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟨🟨1️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥
@Triikouu6 ай бұрын
the robotic urge to push a boulder up a hill for eternity
@dingus29246 ай бұрын
I dont remember pushing this boulder
@invader_guy11716 ай бұрын
I think you should've added a level where the boulders rolling in the opposite direction are randomized, cuz it looks like it just memorized the pattern in which they're rolling down instead of associating the boulders with moving out of their way
@Slappydafrog_
6 ай бұрын
the problem with that is that dingus has no vision. He's like a little blind man with a ball for a walking stick. if he loses it, he has no idea where he is
@Juuls04
6 ай бұрын
@@Slappydafrog_that makes no sense? Thats just not true and has nothing to do with the comment you replied to
@JerryHarding228
6 ай бұрын
I was going to leave this same comment. I would also want to see things like the ball/dingus spawn points on the start platform being randomized
@Juuls04
6 ай бұрын
@ it literally says in the first sentence of the video that it is deep reinforcement, which means he does have vision and that vision has nothing to do with the ball
@jessh4016
6 ай бұрын
@@Slappydafrog_ 0:06
@Roover81386 ай бұрын
In a way, we’re all just like Dingus
@nicooftheforest6 ай бұрын
this was the most emotional video i've ever watched watching dingus fall off at the very end made me cry and when he succeeded i was so proud of him 😢
@AlexandHuman6 ай бұрын
With these AI trainers that have levels, I really would enjoy if at the end of the entire training sequence, we could get a nice all--previous-levels-in-one level that Dingus could do. It would be so satisfying to see all the training pay off on what we've already seen to be really hard at first become easy at the end
@Gamesux4216 ай бұрын
So theoretically, if you let the AI run for long enough and teach it to get rewarded with faster times it would become the perfect speedrunner
@Delmworks
6 ай бұрын
In theory. I’m not 100% sure you could teach an AI to use glitches or exploits, nor could it find them on it’s own
@isleepy6801
6 ай бұрын
@@Delmworks I’ve seen reinforcements algorithms frequently find glitches/ exploits. I’m an ML Engineer.
@coldbluebrownie
6 ай бұрын
the fastest sisyphus
@arceyplayz6 ай бұрын
I now want to see a video of Dingus dealing with openings where Dingus can walk through but the ball cannot, and things like marble rails where Dingus can't walk on but the ball can, so basically, Dingus needs to learn to let go of the ball. That would be VERY interesting to see and another thing I want to see is alternate paths. Have a LOT of paths and see which one Dingus prioritizes? The easiest? The fastest?
@hehrhrthgthtgggrg4841
4 ай бұрын
car
@GlorifiedGremlin4 ай бұрын
6:35 It's literally like Naruto training with shadow clones and then absorbing all their experience 😂
@minecraftperson84126 ай бұрын
Dingus: what is my purpose? *You push rocks up hills* Dingus: Oh god
@mythalda00406 ай бұрын
woah, this is some high quality content, this man deserves more subscribers
@FurretAnimationsIsEpic
6 ай бұрын
agreed
@brentramsten2496 ай бұрын
having the ability to actually read its environment seems to have drastically increased the speed of the learning process. it at the very least seems a substantial improvement over 'blind' learning
@speenta4879
6 ай бұрын
I don’t know much about this subject what do you mean by “blind learning”?
@brentramsten249
6 ай бұрын
@@speenta4879 its when the ai has access to only the 'success/fail' information of its task. without some means of 'seeing' its environment the only thing the ai can do is spam random inputs until it gets a reward for accomplishing one of the preasigned tasks. when you give the ai the ability to determine what objects exist around it, it can actually use that information to make decisions. when you give a blind ai a new level its basically impossible for it to beat it first try. but if the ai can 'see' and is very well trained it can do a level its never seen before and first try it with relative ease
@speenta4879
6 ай бұрын
@@brentramsten249 oh that’s interesting so some ai doesn’t get information on it’s surroundings and only learns based on trial and error?
@brentramsten249
6 ай бұрын
@@speenta4879 "most" non-blind ai is so unusual that im actually surprised when i hear about one that does. theres a decent handful that get some very basic stats feed into it, but one that can actually see stuff around it even somewhat like a player playing a video game. then again i dont keep up with learning ai very much anymore... so maybe thats changed
@nikkiofthevalley
4 ай бұрын
@@brentramsten249I've seen more "non-blind" AI than "blind" AI. I have no idea what gave you that idea.
@icht63486 ай бұрын
Dingus master of the shadow clone jutsu using 57 clones to speed up is learning pace. One must be proud of this dingus and his vast chakra reserves.
@rennegaddefoxxe6 ай бұрын
"The Dingus you are watching is not the only Dingus training at this time." THERE ARE MULTIPLE DINGI!
@snowysalt6 ай бұрын
I have a fun idea you could make a live stream of dingus pushing up a boulder a long track while chat messages can cause boulders to roll down and mess with him
@Dingus_Labs
6 ай бұрын
I've thought about some live stream stuff! I'm not in a position to do it right now but I'm hoping to have the opportunities sometime in the coming months.
@ruaskuma75156 ай бұрын
It would be cool if u made this into a game where people train their own ai on their own maps and then u release a map and see who trained their ai to beat the map the quickest or last the longest in a tounament. I think it would be pretty fun. Great vid btw.
@kingmo5653 ай бұрын
Somewhere there's an operator looking at you on screen and commenting on it as if you were named Dingus and constantly messing up. Awesome video!
@lucasroberds72426 ай бұрын
You have shamed, and embarrassed dingus with these unusual trials. He and his brethren will come back in full force to extract vengeance.
@Vallee1526 ай бұрын
People never show the final AI doing all the levels again
@Dingus_Labs
6 ай бұрын
He would very likely struggle, the changes to the model after he's finished a level means he'd likely need to relearn for a bit when he goes into a different level
@danival2090
6 ай бұрын
@@Dingus_Labs I'd love to see that!
@moonsinger3336 ай бұрын
wow it's kind of a crime you don't have more subs!!! i'll have to keep an eye out for more dingus content, watching this video and cheering on a cute little ai pushing boulders delighted me very much
@eugenekrabs1416 ай бұрын
i need one of those youtube boxing tournaments between Dingus and Sisyphus
@phrygianphreak44286 ай бұрын
Dingus: "what is my purpose?" DL: "You are sisyphus" Dingus: "MY GOD."
@EpicVideos26 ай бұрын
What algorithm did you use? Also, are you able to open source the boulder pushing environment? I'm researching hiearchical reinforcement learning and planning and your boulder environment could be a pretty useful benchmark.
@Dingus_Labs
6 ай бұрын
I use ppo, If you're keen to learn deep reinforcement learning hands on, I highly recommend mlagents toolkit for unity, it comes with tons of examples, already set up example environments and proper readmes with set up steps. Awesome for learning!
@assassinunknown6664
6 ай бұрын
He used 54 other dinguses
@wyndhamwynne2480
6 ай бұрын
@@Dingus_Labsdo you have a shared repo for this project, which I could learn more about PPO RL from?
@Dingus_Labs
5 ай бұрын
@@wyndhamwynne2480 I'm apprehensive to share any repos because I do make use of some commercially purchased textures and models, and they restrict redistributing, which means I'd have to find a different way to store those files and share, and people using the repo would still need to replace those missing assets with something else, I don't yet have a simple and quick solution to that issue.
@wyndhamwynne2480
5 ай бұрын
@@Dingus_Labs that's a shame but understandable. Personally, I'm not interested in any models or textures, or even trying to reproduce locally, I'd only interested in reading through the code and learning more about deep reinforcement learning and specifically ppo. It's mainly the architecture design (width & breadth) for particular applications and the cost/loss function and back propagation to weights & bias adjustment that is unclear to me. Any references in these areas to good learning material would be greatly appreciated!
@jonaw.21536 ай бұрын
You dingus! You've automated the indomitable human spirit! Also, how'd you make this visualization?
@Dingus_Labs
6 ай бұрын
He never stops trying (until the specified Max steps have been hit on training the model). By visualisation do you mean the rays coming out of dingus early in the vid?
@user-cd4bx6uq1y6 ай бұрын
Interestingly this felt like a news segment
@sophiethesnail6 ай бұрын
bro created an intelligence purely for it to suffer eternal punishment
@CobbleDohickey-kv4my6 ай бұрын
one must imagine dingus happy. 😔
@milliemurrell60726 ай бұрын
this is really cool! i'd love to see one of these AIs try going up against harder and harder bullet hells
@krembananowy6 ай бұрын
I just watched some of your first videos and this one. Your editing and commentary these days is a vast improvement! Keep up the good work!
@gooscifur53276 ай бұрын
One must imagine dingus happy, the preprogrammed reward of push a boulder up hill is enough to fill a mans heart
@chris_piss6 ай бұрын
Would have liked to see him go back to the other levels after beating the last one to prove he actually learned all the skills instead of overfitting to each level. Am also curious to see the source code for the network
@WMBCo6 ай бұрын
This could easily be turn into a game
@dogpotato9726 ай бұрын
Adorable character, good editing, good narration. Watch this man blow up.
@CoratoLaga6 ай бұрын
I feel sick as all hell right now, but watching this i imagined Dingus happy and that made me happy :)
@JulianDaEpic06 ай бұрын
this man is criminally under-rated
@ilikefrys
6 ай бұрын
Fr
@SoggywafflesPLAYZ6 ай бұрын
Awesome video I would expect this quality from a much bigger channel! Video blowing up in 3 2 1…
@kylesimpson38564 ай бұрын
"Think you outsmart the biggest dingus in the room?"
@coolestkid34356 ай бұрын
im so happy this video did so well, i love your content and you need more people watching you!!
@catwithlonghair18506 ай бұрын
Wow this is amazing, more people should see this!
@RalorPenwat6 ай бұрын
You could probably get him to not go down the slide easily if you really wanted by adding a minor negative reward function to Dingus's height, but honestly I thought it was cute how he went down the slide too and makes a lot of sense considering the potential of the boulder sliding off the side.
@celinhooliveira37166 ай бұрын
*clicks on the video *sisyphus meme theme starts playing
@TurkeyMeat6 ай бұрын
I appreciate there wasn't the same patronizing basic explanation of machine learning at the beginning of every AI video on youtube here, great vid
@Mr_Reaps256 ай бұрын
One thing i think would have been cool was a course where he has to go down hill beckwards then back up. Wonder he that would work with the reward systems?
@iwantagoodnameplease6 ай бұрын
Does learning later courses negatively affect earlier ones? e.g. if you redo everything with the latest model, will he smash it first time on each course, or still mess up and have to learn a bit more?
@macizzyy
6 ай бұрын
@Dingus_Labs: He would very likely struggle, the changes to the model after he's finished a level means he'd likely need to relearn for a bit when he goes into a different level
@skittersspider17046 ай бұрын
Dingiphus enjoys pushing his boulder, it generates dopamine (or the virtual equivalent) in his brain. Therefore, one sees him happy upon watching him complete these challenges and doesn't have to imagine it.
@DarkOmegaMK26 ай бұрын
Dingus looks like it's ready to check my vibe.
@jamesbhollingsworth54526 ай бұрын
Does dingus "see" the boulders coming toward him? Does he "see" the whole map?
@wooloobutbrave91516 ай бұрын
Im fairly confident that dingus just brute forced most of the levels instead of actually learning because (read more) A: he was repeating a never changing level B: movements were mostly VERY dsimaler to the last generation C: When you double checked his boulder dodging, it completly failed D: It was predodging obstacles before even spawning
@maxwellbenedict5167
6 ай бұрын
yeah, you're probably right but I did notice it did some real learning as well. at 5:36, he stops at the end of the platform, which is something I believe he learned. alternatively, it could just be he picked the best one from every parallel instance and this was just the one that happened to stop.
@wooloobutbrave9151
6 ай бұрын
@@maxwellbenedict5167 or luck
@Dingus_Labs
6 ай бұрын
Movements being similar to last generation isn't too unexpected, changing hyperparameters to encourage more experimentation would have resulted in more variance per generation, I was fortunate that he managed to get through everything with my initial settings. The predodging obstacles is very interesting for sure. And yeah, having static levels does invite brute forcing, I did make Dingus spawn in a random rotation each respawn to add some small amount of variance per run but it's not enough to counter it.
@Corn-Pop.6 ай бұрын
this video has been popping up in my recommended for 3 days and I've avoided it, I didn't want to watch a video of a cube pushing balls, finally I gave in and ended up watching almost 15 minutes of a cube pushing a ball, damn you Dingus
@shrekman87656 ай бұрын
Your are disgusting-ly underrated how tf are ya not at 100k?!?! Look ya have a bright future on KZread
@ricolorenz73076 ай бұрын
I'm not convinced the AI isn't just memorizing inputs to beat the levels. It seems like it remembers how to move the ball forwards, left, and right, but aside from that, it doesn't look like it's using it's senses to actually determine anything. It looks like it's just memorizing inputs.
@PaddyPAA6 ай бұрын
WE MAKIN' IT OUT OF TARTARUS WITH THIS ONE 🗣🗣🗣🔥🔥🥶🥶
@nodrance6 ай бұрын
This could honestly work as a game
@projectnevadas6 ай бұрын
dingus makes me emotional in a way i never thought was possible
@Jacksongirard6 ай бұрын
Im glad this video got blessed by the algorithm. Ill be watching all of your videos now! Good stuff ❤
@aaronleeder82696 ай бұрын
Not only did you make a great video, not only did you create an intelligent ai, but you even came up with an original game idea! This actually looks fun, in a getting up or jump king kind of way!
@lance499
6 ай бұрын
Getting up? you mean Getting Over It?
@aaronleeder8269
6 ай бұрын
@@lance499 oh yeah, you’re right.
@Oh_Low6 ай бұрын
the all mighty algorithm has led me to this video! i subscribed instantly. I hope to see your channel grow more and more!
@struumbug6 ай бұрын
great video! it would be good for the boulder to spawn at a somewhat random x-position every time so the AI doesnt learn to rely on it spawning in the center.
@gvarph72126 ай бұрын
Great video! Subscribed. I just have few suggestions : 1) add a generation counter 2) show how well the final model deals with the first few levels
@superfunnyjoke39226 ай бұрын
Bro you’re blowing up!! Congrats man :)
@xaicracker11056 ай бұрын
We need a Dingus Prime.
@Hiasibua6 ай бұрын
Dingus finna build some pyramids soon 🧐
@dbonechis6 ай бұрын
Yo man, your content is great.
@buzzlightyearpfp76416 ай бұрын
please do more like this. much better than the previous video formats
@stinkerfart6 ай бұрын
now we need to play "me and the birds" by duster in the back
@theyellowpepper99516 ай бұрын
for 200 subs the production qualitys amazing man
@Dingus_Labs
6 ай бұрын
Thank you! I'm trying my best to get better with editing!
@BestPranksterEver5 ай бұрын
Give dingus legs so he needs to learn how to walk and jump.
@awesomegamer24306 ай бұрын
One must imagine dingus happy
@Finsternis..6 ай бұрын
Something I always would want to see in this kind of video is some kind of abstraction of the inputs for the AI so we get some insight into the context it considers. Like "is it aware of the slopes, if so, is it an abstracted into levels of slopyness or an realistic value" or "is it aware of its level of control over the stone" And not only out of sheer curiosity, but also since people often want to compare AI learning with human learning, and in that regard the question wether it is "very sloped" or "sloped by 60°" seems like an important difference.
@egekahraman89856 ай бұрын
even the sisyphus’s job is getting automated by ai
@vorttex_6 ай бұрын
Welp, instant sub. Please continue doing videos like this one
@DipolarApathy6 ай бұрын
I saw your channel name and within the first 10 seconds of this video I subscribed. Excellent shit mate
@Voidmoth13 ай бұрын
Oh god I can already hear the countless people imagining Dingus happy
@loganalbarran29625 ай бұрын
One must imagine dingus educated in the art of pushing boulders up insane hills.
@jayrony696 ай бұрын
You got the algorithm jackpot, gained 50 subs in a few minutes
@haveagreatweek6 ай бұрын
“Ah…a visitor…indeed. The kingdom of heaven has long since forgotten my name…and I am eager to make them remember. So come forth child of man… and…die” - Sisyphus prime 20XX
@RealNahuelBarcelo066 ай бұрын
You should have more subscribers, this is amazing! Also I'm surprised that this is the video that got more than 10K views. Good job!
@D0Samp6 ай бұрын
2:07 Just like me when I played the 8-ball missions in the Splatoon 2 Octo Expansion DLC.
@MichealAfton_jk_i_am_funniboi6 ай бұрын
"One must imagine Sisyphus happy. "
@colthite80706 ай бұрын
really high quality videos for such a low sub count. love this and love how this guy experiments with ai
@crackheadengineer37166 ай бұрын
"A fitting end to a life defined by futile struggle..."
@SteinGaming20066 ай бұрын
This video has earned my subscription
@parasective6 ай бұрын
The room around dingus is an aesthetic I didn't think was real
@momplaysminecraft77374 ай бұрын
When the reinforcement learning is deep, but the hill is steep, Dingus Labs is here for it.
@kaitoisanalien6 ай бұрын
“ This dingus.. To hold.. Me? “
@SquarerSun6 ай бұрын
As Dingus made it further and further up the last ramp, all I could think was "He's beginning to believe".
@milk_yt6 ай бұрын
11:40 My left ear sure loves this level.
@OminouSin6 ай бұрын
Just discovered your channel, absolutely love AI and watching them learn, what kind of programs do you use? I would love to try this myself! AI are so fascinating!
Пікірлер: 1 000
One must imagine Dingus happy
@Dingus_Labs
6 ай бұрын
Legends say he was happily pushing boulders even before he was put into this strange simulated environment.
@christianottley8542
6 ай бұрын
One must imagine dingus upside-down, its hilarious
@Scratchydoesmusic
6 ай бұрын
@@Dingus_LabsHmmm...
@totrox2445
6 ай бұрын
I knew this comment will be here as i saw the title of the video.
@The_Cheez
6 ай бұрын
Dang it. He beat me to it
Give Dingus a top hat that progressively gets taller every video.
@paper177
6 ай бұрын
YES
@paper177
6 ай бұрын
but wait nono one that gets taller everytime he gets a reward
@revimfadli4666
6 ай бұрын
@@paper177 then one no longer needs to imagine Dingus happy, for he is actually happy
@paper177
6 ай бұрын
free hat
@AnUnfortunatePotato
6 ай бұрын
@@paper177is that dimmagingus, owner of the dimmsdale dimmadingus?
it acts so convincingly distraught whenever it loses the ball
@ShrimpFry_Cute
6 ай бұрын
He lost his ball. Why do you think he’s not distraught. I would freak out if I lost my ball too.
@frowner_and_co
6 ай бұрын
He so sad 😢
@nicolasn12
6 ай бұрын
boulder*
@alansmithee419
6 ай бұрын
It doesn't know what to do so it acts chaotically. Humans do the same. We just also get distressed because our brains recognise the problem. We don't act chaotically because we're distressed, it's just that acting chaotically and getting distressed often have a common cause - we don't know what to do.
@manboy4720
6 ай бұрын
@@alansmithee419 so you're saying that dingus is the next stage in human evolution?
I always enjoy these machine learning videos but at the end I always wish they'd run the trained AI through all the levels/courses again.
@garchamayne4254
6 ай бұрын
Definitely. Want to see the difference between the new and final ai
@mondongoloco7902
6 ай бұрын
Wtf is ur pfp
@arend8785
6 ай бұрын
yeah, or maybe a final level with all the elements from previous levels
@222patapon222
6 ай бұрын
I was coming to the comments to ask exactly this, @Dingus_labs can you please show us Dingus going through all the levels with his new found knowledge?
@exLightningg
6 ай бұрын
@@mondongoloco7902 Oh it's a little fella that appeared on an image I AI generated. Full image: i.pximg.net/img-original/img/2023/02/12/21/34/03/105319111_p10.png
I never thought I would hear the sentence "There are, in fact, 54 other dinguses training in the exact same level at the exact same time" that caught me off guard
@rennegaddefoxxe
6 ай бұрын
Dinguses, or Dingi? (Din-guy?)
@wrjtung3456
6 ай бұрын
Please have seggs with me no brainer gamer
@frowner_and_co
6 ай бұрын
@@rennegaddefoxxedinguses is an actual word so let's go with that until we get comfirmation
@rennegaddefoxxe
6 ай бұрын
@@frowner_and_co cactus/cacti ... octopus/octopi ... Dingus/dingi. /s /jk
@no-brainer-gamer
6 ай бұрын
your right! the plural of dingus would be dingi I didn't think about that...@@rennegaddefoxxe
It's incredible how fast this ppo setup learns. I've never seen anything like it. Reinforcement usually takes much more trial and error and training time.
@asshizu
6 ай бұрын
What inputs did you use?
@nikkiofthevalley
6 ай бұрын
I have a feeling that there's a couple hundred instances running in parallel.
@imakething
6 ай бұрын
55, he says so
@Dingus_Labs
6 ай бұрын
Yeah 55, he did still learn super fast, I didn't have too many inputs or actions for him to perform which does help with learning but it was still far ahead of what I expected
@skeleton819
6 ай бұрын
@@Dingus_Labsit seems like dingus doesn’t struggle with overfitting, thats something ive never seen on any other ai vid. This is super cool dude
I can easly imagine AI taking over the world and having this guy to push boulders for the rest of his life.
@LordCrate-du8zm
4 ай бұрын
One must imagine Dingus Labs happy.
when dingus finished all the levels i think it would have been interesting to see you put him back at level 1 and see how all his training thus far would have affected him from the start.
@Junkkie
6 ай бұрын
One must imagine Dingus happy. Or speedrunner
@laula813
6 ай бұрын
Or combine all levels into a mountain.
@eli3963
6 ай бұрын
For sure. I believe reinforcement learning is pretty prone to forgetting.
@Dingus_Labs
6 ай бұрын
There's definitely a risk of overfitting, where the AI has become to specialised for a specific task and struggles to solve more generic problems. Particularly for the last level where he trained on it for a long time and learned to stay very close to the centre.
@tsunertoo9149
6 ай бұрын
Maybe try something where instead of just static stages you set parameters and generate a seed kind of like a rogue like? That way it could potentially force the AI to try and use general information rather than fitting to a stage.@@Dingus_Labs
I think I have a bunch of Dingus cubes running my brain
@MatthewConnellan-xc3oj
3 ай бұрын
Well, use reinforcement learning on them for the next 3 years.
It would be interesting to see a course where it starts with a path forward, but that path is not possible. It is required for you to backtrack behind the start to progress
@Dingus_Labs
6 ай бұрын
Yeah, depending on how the model has trained and how the reward function is programmed it can really struggle to adapt to situations like that!
@Speed001
6 ай бұрын
So a maze
@vincenttrigg4521
6 ай бұрын
@Dingus_Labs I could imagine it working if it knows things like going from point A to point B (aka start to finish). Basically have it have a target location to go to (like the top of the hill here) and once it does that, the map changes its "finish" location. I'll use emoticons here as a diagram. 🟨 map ⏹️ end location 💗 Dingus 🟢 ball 🟦 obstacle 🟥 off map 🟪 reward location 1️⃣2️⃣3️⃣4️⃣ segmented finishes (Basically larger rewards or finish locations that change places once reached) Part 1 🟥⏹️⏹️⏹️⏹️⏹️🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟪🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟪🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 2 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟥🟥🟥🟥🟨🟪🟨🟥🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟨🟨🟪🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 3 - round 1 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨1️⃣🟨🟦🟨4️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟨🟨3️⃣🟨🟥🟥 🟥🟥🟨2️⃣🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 3 - round 2 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨1️⃣🟨🟦🟨3️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨2️⃣🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 3 - round 3 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨1️⃣🟨🟦🟨2️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟨🟨🟪🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥 Part 3 - round 4 🟥🟥🟥🟥🟥⏹️⏹️⏹️⏹️⏹️🟥 🟥🟥🟥🟥🟥🟥🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟦🟨2️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟦🟨🟨🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟪🟨🟨🟨1️⃣🟨🟥🟥 🟥🟥🟨🟨🟨🟨🟨🟨🟨🟥🟥 🟥🟥🟨🟢🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨💗🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟨🟨🟨🟥🟥🟥🟥🟥🟥 🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥🟥
the robotic urge to push a boulder up a hill for eternity
I dont remember pushing this boulder
I think you should've added a level where the boulders rolling in the opposite direction are randomized, cuz it looks like it just memorized the pattern in which they're rolling down instead of associating the boulders with moving out of their way
@Slappydafrog_
6 ай бұрын
the problem with that is that dingus has no vision. He's like a little blind man with a ball for a walking stick. if he loses it, he has no idea where he is
@Juuls04
6 ай бұрын
@@Slappydafrog_that makes no sense? Thats just not true and has nothing to do with the comment you replied to
@JerryHarding228
6 ай бұрын
I was going to leave this same comment. I would also want to see things like the ball/dingus spawn points on the start platform being randomized
@Juuls04
6 ай бұрын
@ it literally says in the first sentence of the video that it is deep reinforcement, which means he does have vision and that vision has nothing to do with the ball
@jessh4016
6 ай бұрын
@@Slappydafrog_ 0:06
In a way, we’re all just like Dingus
this was the most emotional video i've ever watched watching dingus fall off at the very end made me cry and when he succeeded i was so proud of him 😢
With these AI trainers that have levels, I really would enjoy if at the end of the entire training sequence, we could get a nice all--previous-levels-in-one level that Dingus could do. It would be so satisfying to see all the training pay off on what we've already seen to be really hard at first become easy at the end
So theoretically, if you let the AI run for long enough and teach it to get rewarded with faster times it would become the perfect speedrunner
@Delmworks
6 ай бұрын
In theory. I’m not 100% sure you could teach an AI to use glitches or exploits, nor could it find them on it’s own
@isleepy6801
6 ай бұрын
@@Delmworks I’ve seen reinforcements algorithms frequently find glitches/ exploits. I’m an ML Engineer.
@coldbluebrownie
6 ай бұрын
the fastest sisyphus
I now want to see a video of Dingus dealing with openings where Dingus can walk through but the ball cannot, and things like marble rails where Dingus can't walk on but the ball can, so basically, Dingus needs to learn to let go of the ball. That would be VERY interesting to see and another thing I want to see is alternate paths. Have a LOT of paths and see which one Dingus prioritizes? The easiest? The fastest?
@hehrhrthgthtgggrg4841
4 ай бұрын
car
6:35 It's literally like Naruto training with shadow clones and then absorbing all their experience 😂
Dingus: what is my purpose? *You push rocks up hills* Dingus: Oh god
woah, this is some high quality content, this man deserves more subscribers
@FurretAnimationsIsEpic
6 ай бұрын
agreed
having the ability to actually read its environment seems to have drastically increased the speed of the learning process. it at the very least seems a substantial improvement over 'blind' learning
@speenta4879
6 ай бұрын
I don’t know much about this subject what do you mean by “blind learning”?
@brentramsten249
6 ай бұрын
@@speenta4879 its when the ai has access to only the 'success/fail' information of its task. without some means of 'seeing' its environment the only thing the ai can do is spam random inputs until it gets a reward for accomplishing one of the preasigned tasks. when you give the ai the ability to determine what objects exist around it, it can actually use that information to make decisions. when you give a blind ai a new level its basically impossible for it to beat it first try. but if the ai can 'see' and is very well trained it can do a level its never seen before and first try it with relative ease
@speenta4879
6 ай бұрын
@@brentramsten249 oh that’s interesting so some ai doesn’t get information on it’s surroundings and only learns based on trial and error?
@brentramsten249
6 ай бұрын
@@speenta4879 "most" non-blind ai is so unusual that im actually surprised when i hear about one that does. theres a decent handful that get some very basic stats feed into it, but one that can actually see stuff around it even somewhat like a player playing a video game. then again i dont keep up with learning ai very much anymore... so maybe thats changed
@nikkiofthevalley
4 ай бұрын
@@brentramsten249I've seen more "non-blind" AI than "blind" AI. I have no idea what gave you that idea.
Dingus master of the shadow clone jutsu using 57 clones to speed up is learning pace. One must be proud of this dingus and his vast chakra reserves.
"The Dingus you are watching is not the only Dingus training at this time." THERE ARE MULTIPLE DINGI!
I have a fun idea you could make a live stream of dingus pushing up a boulder a long track while chat messages can cause boulders to roll down and mess with him
@Dingus_Labs
6 ай бұрын
I've thought about some live stream stuff! I'm not in a position to do it right now but I'm hoping to have the opportunities sometime in the coming months.
It would be cool if u made this into a game where people train their own ai on their own maps and then u release a map and see who trained their ai to beat the map the quickest or last the longest in a tounament. I think it would be pretty fun. Great vid btw.
Somewhere there's an operator looking at you on screen and commenting on it as if you were named Dingus and constantly messing up. Awesome video!
You have shamed, and embarrassed dingus with these unusual trials. He and his brethren will come back in full force to extract vengeance.
People never show the final AI doing all the levels again
@Dingus_Labs
6 ай бұрын
He would very likely struggle, the changes to the model after he's finished a level means he'd likely need to relearn for a bit when he goes into a different level
@danival2090
6 ай бұрын
@@Dingus_Labs I'd love to see that!
wow it's kind of a crime you don't have more subs!!! i'll have to keep an eye out for more dingus content, watching this video and cheering on a cute little ai pushing boulders delighted me very much
i need one of those youtube boxing tournaments between Dingus and Sisyphus
Dingus: "what is my purpose?" DL: "You are sisyphus" Dingus: "MY GOD."
What algorithm did you use? Also, are you able to open source the boulder pushing environment? I'm researching hiearchical reinforcement learning and planning and your boulder environment could be a pretty useful benchmark.
@Dingus_Labs
6 ай бұрын
I use ppo, If you're keen to learn deep reinforcement learning hands on, I highly recommend mlagents toolkit for unity, it comes with tons of examples, already set up example environments and proper readmes with set up steps. Awesome for learning!
@assassinunknown6664
6 ай бұрын
He used 54 other dinguses
@wyndhamwynne2480
6 ай бұрын
@@Dingus_Labsdo you have a shared repo for this project, which I could learn more about PPO RL from?
@Dingus_Labs
5 ай бұрын
@@wyndhamwynne2480 I'm apprehensive to share any repos because I do make use of some commercially purchased textures and models, and they restrict redistributing, which means I'd have to find a different way to store those files and share, and people using the repo would still need to replace those missing assets with something else, I don't yet have a simple and quick solution to that issue.
@wyndhamwynne2480
5 ай бұрын
@@Dingus_Labs that's a shame but understandable. Personally, I'm not interested in any models or textures, or even trying to reproduce locally, I'd only interested in reading through the code and learning more about deep reinforcement learning and specifically ppo. It's mainly the architecture design (width & breadth) for particular applications and the cost/loss function and back propagation to weights & bias adjustment that is unclear to me. Any references in these areas to good learning material would be greatly appreciated!
You dingus! You've automated the indomitable human spirit! Also, how'd you make this visualization?
@Dingus_Labs
6 ай бұрын
He never stops trying (until the specified Max steps have been hit on training the model). By visualisation do you mean the rays coming out of dingus early in the vid?
Interestingly this felt like a news segment
bro created an intelligence purely for it to suffer eternal punishment
one must imagine dingus happy. 😔
this is really cool! i'd love to see one of these AIs try going up against harder and harder bullet hells
I just watched some of your first videos and this one. Your editing and commentary these days is a vast improvement! Keep up the good work!
One must imagine dingus happy, the preprogrammed reward of push a boulder up hill is enough to fill a mans heart
Would have liked to see him go back to the other levels after beating the last one to prove he actually learned all the skills instead of overfitting to each level. Am also curious to see the source code for the network
This could easily be turn into a game
Adorable character, good editing, good narration. Watch this man blow up.
I feel sick as all hell right now, but watching this i imagined Dingus happy and that made me happy :)
this man is criminally under-rated
@ilikefrys
6 ай бұрын
Fr
Awesome video I would expect this quality from a much bigger channel! Video blowing up in 3 2 1…
"Think you outsmart the biggest dingus in the room?"
im so happy this video did so well, i love your content and you need more people watching you!!
Wow this is amazing, more people should see this!
You could probably get him to not go down the slide easily if you really wanted by adding a minor negative reward function to Dingus's height, but honestly I thought it was cute how he went down the slide too and makes a lot of sense considering the potential of the boulder sliding off the side.
*clicks on the video *sisyphus meme theme starts playing
I appreciate there wasn't the same patronizing basic explanation of machine learning at the beginning of every AI video on youtube here, great vid
One thing i think would have been cool was a course where he has to go down hill beckwards then back up. Wonder he that would work with the reward systems?
Does learning later courses negatively affect earlier ones? e.g. if you redo everything with the latest model, will he smash it first time on each course, or still mess up and have to learn a bit more?
@macizzyy
6 ай бұрын
@Dingus_Labs: He would very likely struggle, the changes to the model after he's finished a level means he'd likely need to relearn for a bit when he goes into a different level
Dingiphus enjoys pushing his boulder, it generates dopamine (or the virtual equivalent) in his brain. Therefore, one sees him happy upon watching him complete these challenges and doesn't have to imagine it.
Dingus looks like it's ready to check my vibe.
Does dingus "see" the boulders coming toward him? Does he "see" the whole map?
Im fairly confident that dingus just brute forced most of the levels instead of actually learning because (read more) A: he was repeating a never changing level B: movements were mostly VERY dsimaler to the last generation C: When you double checked his boulder dodging, it completly failed D: It was predodging obstacles before even spawning
@maxwellbenedict5167
6 ай бұрын
yeah, you're probably right but I did notice it did some real learning as well. at 5:36, he stops at the end of the platform, which is something I believe he learned. alternatively, it could just be he picked the best one from every parallel instance and this was just the one that happened to stop.
@wooloobutbrave9151
6 ай бұрын
@@maxwellbenedict5167 or luck
@Dingus_Labs
6 ай бұрын
Movements being similar to last generation isn't too unexpected, changing hyperparameters to encourage more experimentation would have resulted in more variance per generation, I was fortunate that he managed to get through everything with my initial settings. The predodging obstacles is very interesting for sure. And yeah, having static levels does invite brute forcing, I did make Dingus spawn in a random rotation each respawn to add some small amount of variance per run but it's not enough to counter it.
this video has been popping up in my recommended for 3 days and I've avoided it, I didn't want to watch a video of a cube pushing balls, finally I gave in and ended up watching almost 15 minutes of a cube pushing a ball, damn you Dingus
Your are disgusting-ly underrated how tf are ya not at 100k?!?! Look ya have a bright future on KZread
I'm not convinced the AI isn't just memorizing inputs to beat the levels. It seems like it remembers how to move the ball forwards, left, and right, but aside from that, it doesn't look like it's using it's senses to actually determine anything. It looks like it's just memorizing inputs.
WE MAKIN' IT OUT OF TARTARUS WITH THIS ONE 🗣🗣🗣🔥🔥🥶🥶
This could honestly work as a game
dingus makes me emotional in a way i never thought was possible
Im glad this video got blessed by the algorithm. Ill be watching all of your videos now! Good stuff ❤
Not only did you make a great video, not only did you create an intelligent ai, but you even came up with an original game idea! This actually looks fun, in a getting up or jump king kind of way!
@lance499
6 ай бұрын
Getting up? you mean Getting Over It?
@aaronleeder8269
6 ай бұрын
@@lance499 oh yeah, you’re right.
the all mighty algorithm has led me to this video! i subscribed instantly. I hope to see your channel grow more and more!
great video! it would be good for the boulder to spawn at a somewhat random x-position every time so the AI doesnt learn to rely on it spawning in the center.
Great video! Subscribed. I just have few suggestions : 1) add a generation counter 2) show how well the final model deals with the first few levels
Bro you’re blowing up!! Congrats man :)
We need a Dingus Prime.
Dingus finna build some pyramids soon 🧐
Yo man, your content is great.
please do more like this. much better than the previous video formats
now we need to play "me and the birds" by duster in the back
for 200 subs the production qualitys amazing man
@Dingus_Labs
6 ай бұрын
Thank you! I'm trying my best to get better with editing!
Give dingus legs so he needs to learn how to walk and jump.
One must imagine dingus happy
Something I always would want to see in this kind of video is some kind of abstraction of the inputs for the AI so we get some insight into the context it considers. Like "is it aware of the slopes, if so, is it an abstracted into levels of slopyness or an realistic value" or "is it aware of its level of control over the stone" And not only out of sheer curiosity, but also since people often want to compare AI learning with human learning, and in that regard the question wether it is "very sloped" or "sloped by 60°" seems like an important difference.
even the sisyphus’s job is getting automated by ai
Welp, instant sub. Please continue doing videos like this one
I saw your channel name and within the first 10 seconds of this video I subscribed. Excellent shit mate
Oh god I can already hear the countless people imagining Dingus happy
One must imagine dingus educated in the art of pushing boulders up insane hills.
You got the algorithm jackpot, gained 50 subs in a few minutes
“Ah…a visitor…indeed. The kingdom of heaven has long since forgotten my name…and I am eager to make them remember. So come forth child of man… and…die” - Sisyphus prime 20XX
You should have more subscribers, this is amazing! Also I'm surprised that this is the video that got more than 10K views. Good job!
2:07 Just like me when I played the 8-ball missions in the Splatoon 2 Octo Expansion DLC.
"One must imagine Sisyphus happy. "
really high quality videos for such a low sub count. love this and love how this guy experiments with ai
"A fitting end to a life defined by futile struggle..."
This video has earned my subscription
The room around dingus is an aesthetic I didn't think was real
When the reinforcement learning is deep, but the hill is steep, Dingus Labs is here for it.
“ This dingus.. To hold.. Me? “
As Dingus made it further and further up the last ramp, all I could think was "He's beginning to believe".
11:40 My left ear sure loves this level.
Just discovered your channel, absolutely love AI and watching them learn, what kind of programs do you use? I would love to try this myself! AI are so fascinating!
I'd love to see more of your Dingus
one must imagine ai pushing a boulder up a hill