Learning to See [Part 7: There is no f]
Вставка
- Опубліковано 12 чер 2024
- In this series, we'll explore the complex landscape of machine learning and artificial intelligence through one example from the field of computer vision: using a decision tree to count the number of fingers in an image. It's gonna be crazy.
Supporting Code: github.com/stephencwelch/Lear...
welchlabs.com
@welchlabs - Наука та технологія
This teaching is absolutely incredible. I teach High School Math and Computer Science, and this channel is perhaps the best educational videos I have seen. Your teaching is RELEVANT. I can't imagine how time consuming it is to create, but it is mind-opening, fascinating, and addicting. I found this channel yesterday, and I've now watched every video on the channel.
best christmas present
Excelent. You really should do a collaboration with 3blue1brown
I second this!
yes!!!!
Sound fun!
Yes!!
Completely agree, your way of explaining things and the material you both cover seems like it would make a really good mix ^^
One of the best channels on UA-cam!
One of the? How could you?! It is *the* best!
I can't even put into words how good these videos are.. keep it up and don't give up!
I love this series, but I will always miss the complex numbers one
Me too! That series was tons of fun. I'm thinking the next series will be a little more in that direction.
I would love it, I have seen some really intresting topics, like complex analysis, complex derivatives and complex exponentiation that I wish had shown up in the series, but one cannot cover all topics like that in a series. I like that your serieses often are an atempt to solve a problem, that makes them more focused
Man, I had a completely different question in my research that I was looking everywhere for an answer to, and nothing turned up that was particularly helpful... until I thought to look back at this series by some vague notion that maybe the answer lies in this episode. I was right, this video answered my question. This series and you’re series on imaginary numbers, I always end up coming back to for any number of reasons, and I never leave disappointed. Thank you for that.
Brilliant, just brilliant!
Every time I come back to watch it again... the epiphany hits every time
Can't wait for this update
These videos are amazing! They explain the core ideas of machine learning.
I originally thought this would be a beautifully made but outdated series focusing on 15-year-old methods, but then I saw "stop all this madness and just use deep learning?" in your video.
Yes, please, use your visualization magic on deep learning. And never stop explaining the "why"s of everything.
Channels like this really inspire me to keep on learning!
How do you only have 50K subs ?
because when he's done with the series, machine learning will be obsolete xD
probably because most of the people don't even understand what he is saying?
Probably that is not the answer, we should agree on that hahaha.
But I`m curious: what is up to date, then (I`m a mathematician, I just find all of these series very fun, being far from a professinal on the subject here)? And do you think Welch could jump straight to it (pedagogical question)?
He has such GREAT content, but only a small % of the human population cares about this subject that WE all love!
Dude ... Your videos are really really REALLY great. I can't thank you enough! Please keep it up!
Well done. Your videos demonstrate the process of science, not just machine learning.
Just subbed. Your videos are fabulous and I appreciate all the work you must be putting in.
just found your channel and binge watched all of these videos. Now I have to wait like everyone else :( . Amazing content!
Thanks for watching!
That's perfect :D Even machine learnining turns out to be an exercise in sorting through errors and mistakes :D
Love your style of video. The topic is over my head but still interesting.
on a different note...these videos are beautifully edited...good music
This series is AMAZING! Keep it up :)
Outro music in your videos' is awesome and so are your videos !
This channel is amazing.
This channel is simple beautiful!!
Thanks for watching!
This is awesome. Most of the machine learning classes mention about the statics concepts such as regression, mean, variance e.t.c. Any plans on doing the video like the complex number series for statics. That would be very cool. Man that complex numbers video was damn thorough. The first time i felt video gave me more understanding than working out with paper and pen.
Ngl, that "there is no f" was quite powerful
your videos are awesome, please keep doing them
wow, this makes machine learning better
my prof was ugh, hated the class
Thanks a lot, a lot. Please do next squeals. I beg you. Until then I'll start from Part 1: Introduction and come here every possible times i can.
I clicked to 'liked' before watching it :D
Edit: Watched the video, double liked :D
your doing great job man!
OMFGGGG. I love this guy haha. Cant stop watching your videos !!!
thank you, that was a great explanation!
amazing facts and I urge you to make more machine learning videos from scratch description, it'd really help.
"There is no spoon. It is only in your mind" That takes on a whole new meaning when you realist your brain is doing the same thing as this neural network when you recognizing a spoon!
Great material!
Thanks!
damn good videos
The way the outro was edited, sounds soo much like This American Life.
I'm a musician...when I first saw this I thought it said there is no forte
THERE IS NO FORTE.
Welch Labs
Hey Welch Labs! Love your videos and your amazing style of teaching. I know myself and others were wondering what songs you play in the background of your videos...especially the piano one (the outro of the imaginary numbers video series). I've tried shazaming it but no luck. I'd love to be able to play it on my piano...could you help me out?
Hey Richard, thanks for watching! Here's the piano song: www.premiumbeat.com/royalty_free_music/songs/language-2
Can't wait for the next one!
Is there any other youtuber similar to Welch Labs at explaining things? I mean, he's awesome, sort of unique in this matter...
Superb quality content at every corner in this channel :)
I highly recommend 3blue1brown! ua-cam.com/channels/YO_jab_esuFRV4b17AJtAw.html
Beautiful, pretty, makes sense, what more do you bloody want?
awesome with some Christmas...tks
this video was put up on my birthday!
happy (late) birthday!
I learned truth tables in school in a really dry, non-utilitarian and even non-artistical way. This series of videos would have been AMAZING to make me not hate them (or well, not hate the person who was supposed to teach us this topic, because I am well aware that there are not un-interesting mathematics, just un-appealing explanations)
Thank you for these videos. I usually watch them silently, but I really like them and appreciate them and the effort you put in them. THANK YOU
Thanks for watching, good catch!
I really enjoyed your videos about imaginary numbers. Could you make more videos about math?
Great video!!
What's the music in the end?
Getting flashbacks to my digital circuits course. Inb4 Karnaugh maps.
MOAR PLEASE. :)
great!
nice video! what is the name of the song at the end ?? It's pretty nice
Nice video man, just noticed that at 2:27, you put the 0110 lego on the + side, but on the grid 0110 is -. 11+'s on legos, 10+'s on the grid.
Also at 6:40 I squeezed my buttcheeks so hard. I know it's true and required, but I don't wanna...
Good Catch, yes there's an error in my table. I've added this to the description.
3:08 got it! (x1 or x2) and (x3 or x4)
3:17 yes!
nice job!
Thanks!
What's the music at the end tho? It's really nice
How do you animate and stuff? Great series
Thank you! Mostly just overhead time lapse images, and then I edit for timing.
one day I will understand this
So, I'm still not getting this - what is the utility of the Boolean function?
"There is no f" Press F for respect
May we choose the function which has the simplest reduced-form boolean expression?
Hm, this criteria also may not lead to a single answer :(
that "make assumptions" is more or less how I learned minesweeper. assume square 1 is a mine, meaning one of square 2or 3 is, meaning square 4 is definitely not a mine. but I have other information saying square 4 is definitely a mine, rendering my assumption of square 1 false
to the question of which rule do we use: it depends on which practical situation we need the rule to fit, in this case: "which ones are fingers?"
(X1 and X2) or (X1 and X3) or (X2 and X4) or (X3 and X4)
for which has 2 next to each other
*audible gasp* Not assumptions! Anything but that!
Seriously though, I think a good assumption we could make is one of "minimum complexity". Sure, we could make 2048 different rules, but some of those rules will be every tightly defined to fit the data. If we assume that simple is best, then we could make a lot of progress (so here's hoping that I'm right, and that's what we do next.)
:)
That's Occam's Razor which has some rational basis, but is still an assumption nonetheless.
+Welch Labs (or anybody.) What song is at the end?! I've tried 3 different programs and nothing can identify it.
www.premiumbeat.com/royalty_free_music/songs/benny-s-clarinet?auto thanks for watching!
Welch Labs Thanks! Love it
How the *f* is it possible?
Do you a link to any further reading on the Futility of Bias-Free Learning?
I'll make some specific recommendations in the next video on January 7. Thanks for watching!
Awesome, well done for an excellent video series!
There is no spoon :D
I laughed when I saw "stop all this madness and just use deep learning?"
finaaaaaly :D
None of this makes sense to me and I don't know why or how to start having it make sense :(
I heard 'assumptions' and my stomach dropped. Jesus Christ
Haha awesome.
What aboiut using NAND logic only? It will not solve the problem thoguh. One thing is interesting: we need bias in order to learn. This would explain a lot about our cognitive biases. I always had a feeling that in order for machines to learn (more or less like we do), they would need to be biased and also full of noise and chaotic. So was my biased expectation.
@2:29: What does that mean, "If we did we would be done" ?
he said that we dont know which set the gray squares belong to and that "if we did we would be done"
if we [did know what set they belonged to] we would be done
Isn't not(not a and not b) equivalent to a or b?
Now we must wait until next week can you make more than one video per week?
One day when I have more time! but not yet unfortunately :/
Welch Labs i hope that day will come soon :))))
we have to make.... ASSUMPTIONS! *audience gasps*
:)
dude i want to sleep. but its too fun :(
g4(x) = (x1 and not x3 and x4) or (x1 and not x2 and x3 and not x4) would also work.
There is no spoon, i mean f
First, I used an XOR to make g4 waaaay simpler, second, I'd use a neural net and have it learn by making mistakes like a human
good call on the XOR - I didn't think of that.
why the caption of video is written in russian?
Data science as a drug
I was right! Woo!
Woo!
Hey man, great videos. Really fits with my hobbies but on a different scale.
2:48 watch later
There is no kyser soze.
The f is a lie. :O
Can we get some Fs in the chat
Brrrr assumptions
first comment
i assumed that learning how to teach is easy. that's a bad assumption.
this video makes me realy Bad... Not sad but just thinking about thé fact that I can't realy define anything, just generalize...
This series isn't interesting anymore.... It's not about machine learning at this point , it's following your own train of thought, which is a little flawed and bland.
That's just like... your opinion, man...
Mrjarnould Haha I know :D