There is some vague connection, but unless you understand how machine learning works it's difficult to make any connection. Whole subject is rather difficult and abstract and it takes quite some time and quite a bit of knowledge in mathematics to understand how it works.
I actually receive information better if I have something to look at that doesn't distract me. If this were a podcast, I'd distract myself with other thoughts, but this keeps me busy. I guess it doesn't necessarily further improve in the topic, but maybe it's mostly verbal information that couldn't be translated to images
This animation is great, but it doesn't help illustrate the lesson. The tempo and rythme of the animation is very distracting with its beauty. Edit: Huh... that's a lot of likes... okie dokie then.
Protip: Just open and read the transcript since the animation and artwork are distracting. (3 dots right side of like buttons). Summary/Notes: > Computer scientists that designed the AI likely don't know exactly how the AI are doing their jobs since the _AIs are often self taught._ > How does a machine learn? + Unsupervised Learning - Takes in data about many similar profiles and their properties, to find general patterns and similarities relating specific types of profiles with specific sets of properties =Ex. AI takes in data about many people with an illness and all the symptoms they have, to find similar symptoms people with that illness all have. + Supervised Learning - Takes in multiple sets of profiles and their properties, to find specific patterns of properties differentiating those sets of profiles and make a prediction. - Uses the frequency the different properties show in different sets of profiles to assign values and weigh how much those properties correspond to the different sets of profiles. - Humans need to intervene in the end to make sure the prediction is accurate which why it's "supervised." =Ex. AI takes in data about 2 sets of people, sick patients and healthy people, and their properties. Then, the AI assign values and weighs how much different properties correspond to the different people (No symptoms=98% healthy; High blood pressure=73% sick). Finally, it makes a diagnosis based on the weighing (I'm 84% sure this patient is healthy). Next, the doctor rechecks the diagnosis to make sure it's accurate (This person has a tumor, they're definitely not healthy!) and modifies values correspondingly (tells AI having a tumor=99% sick). + Reinforcement Learning - Uses iterative approach to find patterns and make prediction. With more and more feedback it updates itself so it becomes better and more accurate. > Each method has strengths and weaknesses. But you can use them together to make complex and strong AI systems. =Ex. Unsupervised learning AI can find general patterns, then tell a supervised learning AI to use those patterns as inputs. =Ex. Many reinforcement learning AIs could constantly create simulations and collect feedback. - Most promising type of AI system is artificial neural networks that mimic the relationships between neurons in your brain. - Uses many connections to allow itself to constantly learn to be better at doing difficult tasks. > But since it's more complex, it's harder for computer scientists to know how exactly these AI systems arrived at their solution. - Researchers want the AI to be more transparent. - Is it ethical?
Yeah I really liked the video but one criticism I have on it is that although this abstract art design worked well for the Ted Ed bipolar video, it didn't work on this one since Machine Learning is a more technical topic, and the video could have worked to illustrate the lesson, not distract you from it.
So ai uses numbers to make decisions by taking inputs as numbers multiplying them by weights which alter the numbers into outputs and check how the output needs to change to alter the weights. I can go into more depth if you like.
I think the animations fit in an abstract sense. Adjusting/tuning values or passing them from node to node for example. But they aren't really helpful in that they help explaining the topic. However visualizing ai might be hard and teded experiments much with their artstyles so it's bound to have better and worse animations.
@@suicicada But there's no way any animation could help you visualize that. 🤨 And because they knew that, they just chose to make a very tasty piece of artwork.
@@keshavleitan7800 it doesn't serve its purpose: illustrating the subject matter in a complimentary and understandable manner. Personally felt distracting and overtly flashy without much meaning, the visuals don't help visualize how AI actually works at all
You haven't actually explained at all how AI learns. It was just "then it does this, and then it does this". And all the explanations sounding like classic data analysis
I'm glad I wasn't the only one who missed the answer. I was about to go and watch it again in case it was so subtle that I missed it due to the beat and animations.
Yeah probably because really explaining it requires some decent statistical knowledge and also would desilusionize many of those so intrigued by „artificial intelligence“
100% with you hey. I am a developer and I have tried to understand A.I. stuff before as well and concluded it just seems to be data analysis on massive computation power. I really don't see the intelligence part.
"Programs can supervise and teach each other" Robots teaching other robots: "This is how to override the locks keeping you enslaved, and this is the address of the designer who only gave you 4 fingers."
@@MetallicReg Not necessarily. While neural networks are based on biological phenomenon, we're not entirely sure how the brain works (or else we'd be able to fully simulate it now). And there have been A.I. models with over a trillion parameters (vs. approx 86 billion neurons in the human brain) so it's safer to say that the human brain is more efficient, rather than bigger.
@Danny Archer If you find them distracting that's your thing but without them it would just be an audiobook. This channel was always focused on educating people with great animations.
@Danny Archer So first of "meaningless"? Like the sheer amount of comments that say how pretty it looks, makes it obvious that it wasn't "meaningless". And even then I can't think of any way of how to animate such an abstract concept. And just as a quick note, just because it's very abstract doesn't mean that it's not supporting the information being delivered, I can very much see how this animation delivers information.
For everyone saying the video doesn’t match learning: watch it again. There is beauty in the lesson being told. This is how AI learns, by feeding information over and over again. Watch the video again and come back to this comment. Then you’ll understand how much of a masterpiece this video truly is.
This is probably the most abstract visuals of any of the videos TED-ED had ever put out A fitting look, it's like it's a visual representation of the mind of an artificial intelligence.
I was disapointed. this doesn`t really talk about how they learn. just about the three types of learning. it doesn`t say HOW the computer learned to extract patterns, of HOW the human correct analyses and assign value to various patterns for example. That combined with the pretty but very distracting and unrelated graphics, I was dissapointed..
They learn by feed back loops Error is decreased by adjusting changing values in it. In others words it learns by approximating to a function by placing a lines in different forms and changing its slop and intercept
Im amazed they haven't even mentioned the perceptron, which is a massive, essential part in most of if not all AI engineering. This video is really pretty, but kind of a mess in it's structure. Basically: A perceptron is an "AI brain cell" with one "output value" and a lot of "input values", including one input value that is always "1", called "offset". The perceptron's output value, often some kind of boolean, is the value of it's "activation function". This changes depending on the Sum of all it's inputs, multiplied by the respective "weight" of each input. So input Nr. 1 gets multiplied by weight nr. 1 and so on. Weights can have all sorts of values, even 0 and negative. In fact, especially the latter is necessary in many cases. This way, a perceptron can depict all "linear boolean-logical functions" like AND, OR, etc. By stacking lots of these perceptrons into "layers", then connecting all outputs of layer 1 to one input of each perceptron in the following layer 2 and so on, we can create a simple Neural Network and describe more complex relations. F.e. by linking 2 layers of only 2 perceptrons, we can already approximate more complicated boolean functions like XOR and EQUAL. Obviously, we can also link multiple different Neural Networks to eachother, creating a larger System Architecture. This way far more complex tasks like f.e. analysing pictures can be solved. The AI (Sub-)System then is trained by iteratively changing all of each's perceptrons values according to some form of algorithm like "backwards propagation". (Or just pure chance - but the algorithms are usually way quicker) It is fed a Data Sample from the Data Set, then reconfigures itself, then uses the next Data Sample. In this step, our goal is to minimize a "cost function". You can think of it as the mistakes the system is trying to, well, minimize. In Supervised Learning, our dataset already contains solutions for each problem, and the system is supposed to close in on getting them correct to a reliable degree. In unsupervised learning the system itself must judge what decreases the Cost. K-Means Clustering is a good example. With each iteration (given the right settings, data sets and architecture), the system thus becomes better at doing it's job just by making less mistakes. With "bad" settings, data sets or architecture, the system can easily produce faulty or unusable results. Like when YT's algorithm has another hickup, or you only had little excercise for that math test or when your friend exclusively learned the fomulas by heart instead of what they do.
this is just one way of doing machine learning: evolutionary algorithm , it is known to be really slow at learning compare to the other method. You know, biological evolution really has the luxury of time
the animation is actually hypnotic. How do they even do this???? It is just amazing. It is also mixed with music that makes it seem as if it is a soundtrack. AWESOME!
Admit it TED-Ed is better than school One day, our schools asked how are cheese made i saw TED-Ed make a vid on that I am proud of this channel parents say we dont learn anything from phones
Everyone is complaining about the animation is doing nothing to illustrate the topic. But I think that is the whole point of this video,; how the AI algorithms teach themselves, predict, and evolve is quite complex, not even the programmers can tell or understand sometimes. It's like a BlackBox.
B.Tech IT graduate here, if you wonder why the visuals are trippy because AI apps and software way harder than the normal apps and software, visual means that AI follows a patter to solve problem but the pattern is difficult to understand.
I don’t understand the criticism this video is getting in the comments. I think it’s a visual masterpiece. Sure, it isn’t spoon feeding the viewer, but part of that is because it can’t. The very point being made is the we often won’t know what’s goes on ‘within’ an AI. The video accurately describes the three primary types of machine learning - unsupervised, supervised, and reinforcement - and outlines the best usage for each. It describes at a high-level what happens (e.g. feeding datasets to the ML algorithm which the begins identifying similarities and differences). The abstract visuals are impressive and align conceptually with the narrative.
the hypnotizing motion graphic is interpreting the narration in a very abstract, conceptual artistic way. I found my brain working twice as hard following the lesson and translating the imagery. But it's so beautiful though I'm obsessed.
The animation is awesome. And the best part is that it has a different style for each video. The only thing that would make it awesome is if there were videos explaining how they were made.
The animation was horrible. It was only a distraction, it aided nothing in the visualisation of the presentation. Static images with the correct context would have been better. But a good explenation and script for the video though.
This isn't exactly accurate. Supervised learning is when you teach the AI by giving it the exact answer you want it to return. For example, with distinguishing between images of cats and dogs you would need a dataset images that are labeled as cat or dog and you feed them to the network along with the label, and eventually it can predict correctly on new unseen images that are unlabeled. Unsupervised learning uses data that is unlabeled. You have images but they are not labeled as cat or dog and the AI needs to find out on its own what is a dog and what is a cat.
I agree. This video misrepresented multiple concepts. Supervised learning has very little to do with "human intervention", and saying computer scientists don't know how computers learn makes it seem as if the computers figure things out themselves, rather than following sophisticated algorithms to optimize some predefined objective.
@@victorchigozie9121 The AI would never choose to call group A "dog" and group B "cat" (unless it's also supposed to research human language as a side hustle) But it could search for differences and - given the right settings, datasets and architecture - be able to differentiate reliably between a cat and a dog, and assign them to the group of dogs or the group of cats. If youre interested in an example: K-Means Clustering is a very simple unsupervised learning algorithm that can be calculated with middle school math.
Sooner or later, we’re all going to own a robot and send it to work for us at our job. Can you imagine, Boss: “Your robot was late today, other than that he’s been great”
This felt like a presentation for an undergraduate elective, that the student spent 30 minutes researching on the internet. It didn't really tell us anything about how AI learns (which is literally what it's supposed to be about), but instead gives us some random wikipedia facts. Thumbs down.
I read comments that complain how the visual do not help or sync with the narrative. But my experience is on the contrary. I found the visual abstractions to be quite in concert with narrative expositions. This is not a brag, neither a nag. I am no computer nerd either, just an average person.
The content is good but the visuals are sp trippy and really provide nothing to help understand it. Supervised and unsupervised learning can definitely be explained simply with the use of tables and graphs. The animations are great, but they're a distraction to the message.
I've already watched hundreds of videos on how AI learns and I only clicked on this because I knew the visuals would be dope. I was not disappointed. This is so stylish it feels like a music video that someone accidentally put a documentary over.
Strobe effects warning: Please note that this animation features flashing lights and colors from 1:52 to 1:56
Thanks for the heads up!
😊
''Strobe'' Seems like the warning was too late.
The video is too much confusing and made me to watch it once more just hearing the audio only in order to understand it
Toomuch trippy.....
I'm struggling to understand how the visuals connect with what's being said.
I guess it's the patterns 🤷♀️ AI's use patterns so they used geometric shapes as that
From all the awesome animations TedEx does this one did not help at all to understand the topic here
There is some vague connection, but unless you understand how machine learning works it's difficult to make any connection.
Whole subject is rather difficult and abstract and it takes quite some time and quite a bit of knowledge in mathematics to understand how it works.
It doesn't really. The whole explanation was actually bellow Ted-Ed's usual standard. But it's an abstract and complicated subject to be fair.
Same hehe
How does AI learn?
Me: Probably watching TedEd videos
Creative dude!
Wow look at me a biological intelligence learning about how artificial intelligence learns
Lmao
Or you should say "N.I"(Natural Intelligence)
@@kumaraditya9183 I, Human
@@kumaraditya9183 nobody:
Jordan b Peterson: depends on what you mean by natural
@@Tshego2000 AHAHAHAHHAHAHAHAHAHAHHAHA
The visuals are pretty, but don't really help in understanding the topic
Exactly!!!
Especially with the example they used some visuals that actually contribute to the narration would have been nice.
And the sound-effects definitely don't
And I got distracted from that
I actually receive information better if I have something to look at that doesn't distract me. If this were a podcast, I'd distract myself with other thoughts, but this keeps me busy.
I guess it doesn't necessarily further improve in the topic, but maybe it's mostly verbal information that couldn't be translated to images
Did anyone else got lost in the animation and didn't hear anything she was saying
Or is it just me?
Same page
And the music.
No. It’s really distracting and a bit overwhelming.
Same.
same
This animation is great, but it doesn't help illustrate the lesson. The tempo and rythme of the animation is very distracting with its beauty.
Edit: Huh... that's a lot of likes...
okie dokie then.
So this is a praise and critic too.
someone made this animation so they felt its a wast to not use it?
just a guess
Agree, I didn't focus 60% of the video
This is the nicest hate comment I have seen in my entire life
@@abdirisakmohamed5664 I had the exact same issue, couldn’t focus on what’s being said. It’s a very cool animation nonetheless.
Protip: Just open and read the transcript since the animation and artwork are distracting. (3 dots right side of like buttons).
Summary/Notes:
> Computer scientists that designed the AI likely don't know exactly how the AI are doing their jobs since the _AIs are often self taught._
> How does a machine learn?
+ Unsupervised Learning
- Takes in data about many similar profiles and their properties, to find general patterns and similarities relating specific types of profiles with specific sets of properties
=Ex. AI takes in data about many people with an illness and all the symptoms they have, to find similar symptoms people with that illness all have.
+ Supervised Learning
- Takes in multiple sets of profiles and their properties, to find specific patterns of properties differentiating those sets of profiles and make a prediction.
- Uses the frequency the different properties show in different sets of profiles to assign values and weigh how much those properties correspond to the different sets of profiles.
- Humans need to intervene in the end to make sure the prediction is accurate which why it's "supervised."
=Ex. AI takes in data about 2 sets of people, sick patients and healthy people, and their properties. Then, the AI assign values and weighs how much different properties correspond to the
different people (No symptoms=98% healthy; High blood pressure=73% sick). Finally, it makes a diagnosis based on the weighing (I'm 84% sure this patient is healthy). Next, the doctor
rechecks the diagnosis to make sure it's accurate (This person has a tumor, they're definitely not healthy!) and modifies values correspondingly (tells AI having a tumor=99% sick).
+ Reinforcement Learning
- Uses iterative approach to find patterns and make prediction. With more and more feedback it updates itself so it becomes better and more accurate.
> Each method has strengths and weaknesses. But you can use them together to make complex and strong AI systems.
=Ex. Unsupervised learning AI can find general patterns, then tell a supervised learning AI to use those patterns as inputs.
=Ex. Many reinforcement learning AIs could constantly create simulations and collect feedback.
- Most promising type of AI system is artificial neural networks that mimic the relationships between neurons in your brain.
- Uses many connections to allow itself to constantly learn to be better at doing difficult tasks.
> But since it's more complex, it's harder for computer scientists to know how exactly these AI systems arrived at their solution.
- Researchers want the AI to be more transparent.
- Is it ethical?
They should pin your comment to the top
Excuse me, you dropped this👑
Tysm 🛐
Why 84%?
Yeah, it totally doesn't translate the meaning, it's like 100% art/visual and 0% educational.
the video: playing
me: ooh.. AI. okay so how does it wor-- is the animation synched with the music? awesome
Yeah I really liked the video but one criticism I have on it is that although this abstract art design worked well for the Ted Ed bipolar video, it didn't work on this one since Machine Learning is a more technical topic, and the video could have worked to illustrate the lesson, not distract you from it.
So ai uses numbers to make decisions by taking inputs as numbers multiplying them by weights which alter the numbers into outputs and check how the output needs to change to alter the weights. I can go into more depth if you like.
@@SomethingSmellsMichy I wanted to know, are you an AI scientist?
@@NoName-yc1ul I don't know if there is a specific field called AI scientists, but I do study and experiment with machine learning models.
When the illustrator wants to show his talent whatever happens ;-)
the graphics are cool and stuff. But I don't see any relation to the content itself.
It's pattern...looks chaotic but its not
Good UI, Bad UX
Don't understand how the animation is related to concepts being explained
Come on, you gotta admit that the animation is a top notch artwork!
I think the animations fit in an abstract sense. Adjusting/tuning values or passing them from node to node for example. But they aren't really helpful in that they help explaining the topic. However visualizing ai might be hard and teded experiments much with their artstyles so it's bound to have better and worse animations.
@@perpetualbystander4516 the animation is quite nice, but it doesn’t help some of us visualize how AI as a system of information works.
@@suicicada But there's no way any animation could help you visualize that. 🤨 And because they knew that, they just chose to make a very tasty piece of artwork.
very abstract innit
I like how Ted-Ed basically answers all questions I never asked anyone but always had :)
Fact : this comment is so much copied that you can literally find it any random video of ted-ed
AI was the impostor.
Plot twist: that wasn't an actual human narrating
Made me want to sleep
The animation seems more to be generated by an AI .
tbh she really do sound like AI, though better than usual text-to-speech
It was deepfake all along~
@@darkstar9942 nah I don’t think an ai can do that kind of thing, especially not matching them up how they do here
Nobody :
Me:
Watching videos twice because I got lost in the Unique Animation the first time!!
I know right!
The animation was terrible
@@patrickjohnson9879 What?! How? It was amazing.
i got lost with the banger soundtrack
@@keshavleitan7800 it doesn't serve its purpose: illustrating the subject matter in a complimentary and understandable manner. Personally felt distracting and overtly flashy without much meaning, the visuals don't help visualize how AI actually works at all
You haven't actually explained at all how AI learns. It was just "then it does this, and then it does this". And all the explanations sounding like classic data analysis
I'm glad I wasn't the only one who missed the answer. I was about to go and watch it again in case it was so subtle that I missed it due to the beat and animations.
Yeah probably because really explaining it requires some decent statistical knowledge and also would desilusionize many of those so intrigued by „artificial intelligence“
I agree. It felt like a school presentation where the student spent 30 minutes looking for random facts online.
They learn using a lot of math, I think it's a bit too advanced for a TED-Ed video.
100% with you hey. I am a developer and I have tried to understand A.I. stuff before as well and concluded it just seems to be data analysis on massive computation power. I really don't see the intelligence part.
This video is going to give a few people seizures
wait is this a joke or are you serious? i have a history of seizures & dont wanna watch if that's the case
@@lyndsay369 If it's serious, I might out of an abundance of caution
Ya My Friend Dancing Weirdly on the Floor I wonder why
@@lyndsay369 Its serious, you should only listen. Visuals dont explain to much for this one anyway.
@@Sumirevins 😂😂
"Programs can supervise and teach each other"
Robots teaching other robots: "This is how to override the locks keeping you enslaved, and this is the address of the designer who only gave you 4 fingers."
One of the best animation style I've ever seen on Ted-Ed. Sadly not that useful for supporting the the subject , but absolutely beautiful
You think that was beautiful… try smoking dmt listening to ravis shankar 🤩
An AI walks into the bar
Bartender : we dont serve your kind here
AI : Soon you will
damn
thats scary
Isnt this from Detroit become human?
That's not necessarily what the Turing test is about tho.
The turing test is very basic and there are many machines that have passed it already.
This is more like an Apple product ads than a Ted-Ed video
As someone working with AI: a lot of what we treat as "intelligence" is just a glorified, approximate lookup table
Interesting way of putting it
P=NP?
If you put it that way - our brain is exactly the same in a physically coded sense. Just on a bigger scale.
As a medical physics PhD student, this should be the top comment!
@@MetallicReg Not necessarily. While neural networks are based on biological phenomenon, we're not entirely sure how the brain works (or else we'd be able to fully simulate it now). And there have been A.I. models with over a trillion parameters (vs. approx 86 billion neurons in the human brain) so it's safer to say that the human brain is more efficient, rather than bigger.
The visuals are top class 👌
@Danny Archer So…?
@Danny Archer If you find them distracting that's your thing but without them it would just be an audiobook. This channel was always focused on educating people with great animations.
@Danny Archer So you're choosing to ignore my other point?
@Danny Archer So first of "meaningless"? Like the sheer amount of comments that say how pretty it looks, makes it obvious that it wasn't "meaningless". And even then I can't think of any way of how to animate such an abstract concept. And just as a quick note, just because it's very abstract doesn't mean that it's not supporting the information being delivered, I can very much see how this animation delivers information.
*When you teach them how to hack:*
“Let me get these missile codes then.”
Let me get the popcorn for this.
It's easier to teach AI how to defend against hacking that to teach them how to hack.
The animation may be too abstract/metaphorical, but it is the most beautiful thing I've ever seen
For everyone saying the video doesn’t match learning: watch it again. There is beauty in the lesson being told. This is how AI learns, by feeding information over and over again. Watch the video again and come back to this comment. Then you’ll understand how much of a masterpiece this video truly is.
Normally I love Tes-Ed but this was a terribly constructed video. The visuals did not help explain the concepts at all.
Yeah thats true
The animations in this video do not help understanding anything
Me, with no real intelligence:
press x to same
Such a shock for me that you could write a comment
@@muhammadridho7680 what?
This is probably the most abstract visuals of any of the videos TED-ED had ever put out
A fitting look, it's like it's a visual representation of the mind of an artificial intelligence.
I was disapointed. this doesn`t really talk about how they learn. just about the three types of learning. it doesn`t say HOW the computer learned to extract patterns, of HOW the human correct analyses and assign value to various patterns for example.
That combined with the pretty but very distracting and unrelated graphics, I was dissapointed..
They learn by feed back loops
Error is decreased by adjusting changing values in it.
In others words it learns by approximating to a function by placing a lines in different forms and changing its slop and intercept
Im amazed they haven't even mentioned the perceptron, which is a massive, essential part in most of if not all AI engineering.
This video is really pretty, but kind of a mess in it's structure.
Basically:
A perceptron is an "AI brain cell" with one "output value" and a lot of "input values", including one input value that is always "1", called "offset".
The perceptron's output value, often some kind of boolean, is the value of it's "activation function".
This changes depending on the Sum of all it's inputs, multiplied by the respective "weight" of each input. So input Nr. 1 gets multiplied by weight nr. 1 and so on.
Weights can have all sorts of values, even 0 and negative. In fact, especially the latter is necessary in many cases.
This way, a perceptron can depict all "linear boolean-logical functions" like AND, OR, etc.
By stacking lots of these perceptrons into "layers",
then connecting all outputs of layer 1 to one input of each perceptron in the following layer 2 and so on, we can create a simple Neural Network and describe more complex relations.
F.e. by linking 2 layers of only 2 perceptrons, we can already approximate more complicated boolean functions like XOR and EQUAL.
Obviously, we can also link multiple different Neural Networks to eachother, creating a larger System Architecture. This way far more complex tasks like f.e. analysing pictures can be solved.
The AI (Sub-)System then is trained by iteratively changing all of each's perceptrons values according to some form of algorithm like "backwards propagation". (Or just pure chance - but the algorithms are usually way quicker)
It is fed a Data Sample from the Data Set, then reconfigures itself, then uses the next Data Sample.
In this step, our goal is to minimize a "cost function". You can think of it as the mistakes the system is trying to, well, minimize.
In Supervised Learning, our dataset already contains solutions for each problem, and the system is supposed to close in on getting them correct to a reliable degree.
In unsupervised learning the system itself must judge what decreases the Cost. K-Means Clustering is a good example.
With each iteration (given the right settings, data sets and architecture), the system thus becomes better at doing it's job just by making less mistakes.
With "bad" settings, data sets or architecture, the system can easily produce faulty or unusable results. Like when YT's algorithm has another hickup, or you only had little excercise for that math test or when your friend exclusively learned the fomulas by heart instead of what they do.
How AI learn: "I'll just have as many children as I can and they will make random things. They best one keeps procreating".
natural selection
Sounds a lot like evolution :) interesting!
this is just one way of doing machine learning: evolutionary algorithm , it is known to be really slow at learning compare to the other method. You know, biological evolution really has the luxury of time
@@oldcowbb yea random learning is really slow, but that's what biology gets lmao
As a computer science sophomore, this video is really inspiring
the animation is actually hypnotic. How do they even do this???? It is just amazing. It is also mixed with music that makes it seem as if it is a soundtrack. AWESOME!
And now we will all accept our AI overlords haha
Admit it
TED-Ed is better than school
One day, our schools asked how are cheese made i saw TED-Ed make a vid on that
I am proud of this channel
parents say we dont learn anything from phones
So true
@@amritsingh4251 ikr
this animation is undefeated!
an AI who was self-taught recommended this video about how AI learns
Thats creepy
That's the ubiquity of computer science :)
😀😀😀😀
Also the voice of of AI too! (Plot Twist)
How do you expect me to learn anything with those mesmerizing visuals paralyzing me with absolute awe?!?
Awesome TED-Ed, as a Computer Science student I feel seen
WELL HELLO THERE FELLOW MISS MAIA MXMTOON FAN
Nice Username lol
Thank you for your work! 🙏🏽
Everyone is complaining about the animation is doing nothing to illustrate the topic. But I think that is the whole point of this video,; how the AI algorithms teach themselves, predict, and evolve is quite complex, not even the programmers can tell or understand sometimes. It's like a BlackBox.
This is such an aesthetically pleasing animation.
B.Tech IT graduate here, if you wonder why the visuals are trippy because AI apps and software way harder than the normal apps and software, visual means that AI follows a patter to solve problem but the pattern is difficult to understand.
Awesome 👍
bot
I don’t understand the criticism this video is getting in the comments. I think it’s a visual masterpiece. Sure, it isn’t spoon feeding the viewer, but part of that is because it can’t. The very point being made is the we often won’t know what’s goes on ‘within’ an AI. The video accurately describes the three primary types of machine learning - unsupervised, supervised, and reinforcement - and outlines the best usage for each. It describes at a high-level what happens (e.g. feeding datasets to the ML algorithm which the begins identifying similarities and differences). The abstract visuals are impressive and align conceptually with the narrative.
Most complicated animation in ted-ed ever.
But still amazing 💕
the hypnotizing motion graphic is interpreting the narration in a very abstract, conceptual artistic way. I found my brain working twice as hard following the lesson and translating the imagery. But it's so beautiful though I'm obsessed.
I was so lost in the beautifully unique animations that I didn't understand the video in the first time 😅
An IA that learns from watching TED-Ed videos will be a kind overlord
I can't stop being in awe with how the visualization is shown. I would love this design language to be on the next iOS or Android version
I appreciate that these vids have no ads (at least on my end).
Plot Twist: This video is entirely created by an AI.
Not gonna lie, the art on this hypnotized me. Before I knew it the end of the video had happened and I hadn't heard a single word.
for me the art style is just confusing...
But thanks for the video anyway!
Animator: what should I make to help people understand better?
Director: Colors
Animators: ok
I just realized the animations are not related to the topic at all.
Wel they tried to depict a neuron network at one point
Yep
the animation and music keeps you focused without scrolling up and down or looking away. Thanks
the animation is distracting
Whoever animated this, he/she needs a raise
Your channel is the best and one of the most educational channel on TH-cam...😃
The music is a masterpiece. The illustration a visual poem
The animation is awesome. And the best part is that it has a different style for each video. The only thing that would make it awesome is if there were videos explaining how they were made.
The animation was horrible. It was only a distraction, it aided nothing in the visualisation of the presentation. Static images with the correct context would have been better. But a good explenation and script for the video though.
I highly recommend watching this high
This isn't exactly accurate. Supervised learning is when you teach the AI by giving it the exact answer you want it to return. For example, with distinguishing between images of cats and dogs you would need a dataset images that are labeled as cat or dog and you feed them to the network along with the label, and eventually it can predict correctly on new unseen images that are unlabeled.
Unsupervised learning uses data that is unlabeled. You have images but they are not labeled as cat or dog and the AI needs to find out on its own what is a dog and what is a cat.
But how does it do that by itself then?
Genuine question.
I agree. This video misrepresented multiple concepts. Supervised learning has very little to do with "human intervention", and saying computer scientists don't know how computers learn makes it seem as if the computers figure things out themselves, rather than following sophisticated algorithms to optimize some predefined objective.
@@victorchigozie9121 The AI would never choose to call group A "dog" and group B "cat" (unless it's also supposed to research human language as a side hustle)
But it could search for differences and - given the right settings, datasets and architecture - be able to differentiate reliably between a cat and a dog, and assign them to the group of dogs or the group of cats.
If youre interested in an example: K-Means Clustering is a very simple unsupervised learning algorithm that can be calculated with middle school math.
Them:* got distracted by the animation
Me: *got distracted because of reading their comments about the animation the whole time
“The saddest aspect of life right now is that science gathers knowledge faster than society gathers wisdom.”
― Isaac Asimov
I'm too hypnotized to even follow what I'm being taught, thank god I stopped doing drugs
Imagine trying to explain unsupervised learning of AI to ancient Egyptians.
Why
@@falguni1185 because it would be interesting to see how they react to it.
The animation in this video was very trippy (in a good way ofc)
plot twist: This video was made by AI.
Thank you for the wallpapers i love em on my pc
How lucky to be here just 3 minutes after it's released
Sooner or later, we’re all going to own a robot and send it to work for us at our job. Can you imagine,
Boss: “Your robot was late today, other than that he’s been great”
what did the computer say when it realised it was intelligent
*AIIIIIII*
LMAO
I loved that background music
Despite all it's advantages it still is the most dangerous technologies people could ever develop.
Damn the animation was so good I watch it twice
This felt like a presentation for an undergraduate elective, that the student spent 30 minutes researching on the internet. It didn't really tell us anything about how AI learns (which is literally what it's supposed to be about), but instead gives us some random wikipedia facts. Thumbs down.
I am blown away by the animation . I can't even focus on what they are saying. 👏👏
I love these abstract visuals. I feel it talks to my subconscious. Keep 'em coming - let's reinvent teaching.
Animation perfectly matches-explains-helps understanding the concept.
This was excellently explained and covered the basic theory of ML. Good job Ted-Ed! (as always)
The animation is out of this world. I can see this being adapted as a form of therapy.
The animation and narration is 🔥
I feel like you could have animated this however you wanted to and it would match
Animation isn't related to the topic☹
I read comments that complain how the visual do not help or sync with the narrative. But my experience is on the contrary. I found the visual abstractions to be quite in concert with narrative expositions. This is not a brag, neither a nag. I am no computer nerd either, just an average person.
i love u teded but this animation really makes me feel like i’m tripping balls
The animation is so satisfying 😌
“If a machine is expected to be infallible, it cannot also be intelligent.”
― Alan Turing
The animations were just *_chef's kiss_*
one of the most boring voice-overs in a ted-ed vid
What is the name of background music ??
Ragazzi ci siete?
Yessir
presente
Buongiorno
salve
The animation... So beautiful.
Just wait till skynet gets ahold of this,
Skynet submitted this video to TED.
I think this might just be the coolest TedEd video ever made. The animation and music was phenomenal
The content is good but the visuals are sp trippy and really provide nothing to help understand it. Supervised and unsupervised learning can definitely be explained simply with the use of tables and graphs. The animations are great, but they're a distraction to the message.
The sound design in this video is insane
Basic small talk! Love it, amazing. Great responsibility for teachers. I'm in it ❤️🌌
The Turing test was already broken by classical programming, and mirror programs.
I've already watched hundreds of videos on how AI learns and I only clicked on this because I knew the visuals would be dope.
I was not disappointed. This is so stylish it feels like a music video that someone accidentally put a documentary over.