But How Does The Perceptron Actually Work?

Artificial Intelligence For Everyone: Episode #6
What is Neural Networks in Artificial Intelligence and Machine Learning? What is a linear classifier and how does linear classifiers work? How does the perceptron work?
This video is an beginners guide to neural networks, and aims to help you understand how the perceptron works - somewhat of a "perceptron for dummies" video - explained in a sense so that everyone can follow.
Artificial Intelligence and Machine Learning shapes the world around us more than ever, and understanding the basic concepts is an useful asset for any person, regardless of their walk in life or profession.
www.kodkompis.com
PEACE!
----------------------------------------------------------------------------------------------
JOIN NO PMO NATION 👬:
----------------------------------------------------------------------------------------------
👬 Instagram: / nopmonation
----------------------------------------------------------------------------------------------
JOIN THE ARMY OF HAPPIER AND STRONGER PEOPLE 👬:
----------------------------------------------------------------------------------------------
🎓 SUBSCRIBE ON KZbin: goo.gl/JDWLKZ
🎓 JOIN US ON SLACK: goo.gl/srBTka
🎓 JOIN MY EXCLUSIVE MAILING LIST: eepurl.com/di4dNj
----------------------------------------------------------------------------------------------
POPULAR EDUCATION SERIES 💝:
----------------------------------------------------------------------------------------------
🎓 MASTER NOFAP: goo.gl/z6E6HU
🎓 BECOME HAPPIER: goo.gl/DZ4cps
🎓 ATTRACT WOMEN: goo.gl/MKxdeS
🎓 MACHINE LEARNING: goo.gl/hULpKQ
🎓 ARTIFICIAL INTELLIGENCE: goo.gl/pzCWpU
----------------------------------------------------------------------------------------------
HOW TO ASK OSCAR QUESTIONS 🎤:
----------------------------------------------------------------------------------------------
👬 MESSAGE ME ON INSTAGRAM: / oscaralsing
👬 ASK ME ON SLACK: goo.gl/srBTka
Linkedin: / oscaralsing
Facebook: / oscaralsingcom
Website: www.oscaralsing...
----------------------------------------------------------------------------------------------
PRODUCTS I LOVE ❤️:
----------------------------------------------------------------------------------------------
LIFE-CHANGING BOOKS: goo.gl/MMH4XG
MY CAMERA/PROGRAMMING GEAR: goo.gl/WPCkZr
----------------------------------------------------------------------------------------------
ABOUT OSCAR 💝:
----------------------------------------------------------------------------------------------
Oscar is a leader, educator and programmer specialised in Artificial Intelligence and Machine Learning who strives to build a world where all leadership spawns from an intrinsic compassion for others.
He is heavily interest in mindfulness and meditation and is a daily Brazilian Jiu-Jitsu practitioner. Furthermore, he Loves lifting heavy things and reads a lot of books and believes in a world where compassion and mutual understanding and respect permeate all of our actions.
🎉 Leader of the Year (2017, All Swedish Students)
🎉 10/100 @ Sweden's Top Future 100 Leaders 2018
🎉 37/100 @ Sweden's Top Future 100 Leaders 2017

[@OscarAlsing]

Are you ready to become a MODERN ALPHA MAN? 👇 🙌🏼 Subscribe to HOW TO ALPHA: kzbin.info/door/4sGrtzlpt2fLSsfQ9NrpbA

[@jenelleanstey6860]

Explained very well. I'm a little confused about how the line actually gets adjusted. I understand that you change the weights, however, how does this cause the line to change?

[@Titan9731]

But the real question is: Why is Elon musk in the thumbnail?

[@rashmir2534]

To see this video and to become like Elon Musk. LoL

[@BhargavChinta16]

Gosh, I came here to see why he was in the thumbnail. :/

[@31media97]

I stop view this video, at minute 0.34 ,

[@RomanQrr]

I was sent here by a cat.

[@garibaldyzhu6679]

me too

[@gaurishmali3706]

Me to in 2021

[@rebeccatorbergsen]

same

[@zRedPlays]

same

[@veshuee003]

same

[@antipov2]

Really messed me up when you labeled X as vertical and Y as horizontal.

[@OscarAlsing]

haha, I'm sorry my friend!

[@MelinaMelina80]

me too, my math neurons are trully deeply disturbed by this : )

[@ondetv7405]

I came here by a referrence from the game "while True: Learn()". Great video! Just goes to show that it is still both relevant and good teaching 4 years after you created it. One thing though: Please do try to learn to pronounce "linear" correctly. That confused me a bit for a while thinking you were referring to some concept i was unfamiliar with until i realized later in the video that you were "just" talking about lines :) Once again, great work!

[@dwikonugroho6868]

same, epic games really give us amazing "underrated stuff".

[@danielwilczak7833]

Wx+b is really just your basic slope formula y=mx+b. You just substitute in W "weight" as your slope and b "bias" as your y intercept. I don't know why he said its from the origin, that doesn't make sense. This is why detail matters.

[@Camiliostation]

Well in some sense Oscar is right, since the y-intercept is basically the point where the y-axis is intercepted and meaning that that value of y (or b=bias in this case) is the distance from the origine from where the line should be drawn. So seeing it as a distance from where the line is set. Hope this helps?

[@yashwantdhole7645]

Dear sir, I am very happy to see teachers like you. I am 21 year old Indian guy in final year I.T Engineering. Teachers themselves at our college do not know anything about this subject. Tommorow is my paper and you came up like an angel. Thanks a lot. I wish India had more teachers like you.

[@hammadshaikhha]

We need more teaching videos like these Oscar! Sometimes simple and informative works best.

[@OscarAlsing]

Thank you :)

[@jankabicek8272]

Hi i was brought here by a game actually ,called while True:learn() i was curious about this perceptron thanks for explanation. P.S:sorry for my bad english.

[@ashiktakvir2547]

a video on backpropagation and weight adjustment would be really helpful

[@DrRandyDavila]

kzbin.info/www/bejne/mYK5ZKZoirCeaZI

[@cashout8582]

Bruh, you educators got to make this way more entertaining. A video game brought me here to lose curiosity

[@quynhnguyenuc4719]

I come here 'cause I wanna know what's "meow meow" meaning...

[@nawrin6636]

Thank you so much... From Bangladesh 🇧🇩

[@OscarAlsing]

You're welcome!

[@rosaliebeaureguard5202]

THIS SAVED ME THANK YOU I have a cognitive science exam tomorrow and this cleared everything up thank you!!!!

[@OscarAlsing]

+Rosalie Beaureguard no worries! Glad I could help and good luck!

[@prnfl]

LIH-NEE-YER

[@mysticalearth4259]

Great tutorial. I liked how you explained the part where you adjust the linearity of the vectors if the output is incorrect - based on the supervised learning classifications. Clear and easy to understand!

[@OscarAlsing]

Thank you!

[@julianlee6542]

nice video but I really want to know "how" the weights and bias are updated in a quantitative way.

[@wall_k]

The right pronunciation is ˈlinēər, I think

[@zawsrdtygbhjimokpl6998]

a guy whose eyes look like machine ones is explaining an aspect of AI. Thought it would've taken them longer to take over

[@signature445]

could you please tell ,which writing software you are using,Thanks in advance.....

[@andrezaluko]

OMG! I had trouble paying attention 😏

[@Nigolasy]

I see the IKEA curtains I use - I like

[@472roy]

thank you for this video. i'm very new to this. but learned a lot

[@pavanvankadaru8965]

One of the best videos on this subject in a very succinct yet detailed

[@OscarAlsing]

Thank you Pavan!

[@unintelligentpieceofhuman9658]

am I the only one who got sent here by a cat?

[@Mr8999naga]

Good ones expecting more videos like these as in explaining w.r.t the real time examples

[@123liveo]

Excellent description. Helped me understand better. Thanks.

[@jmymlr]

Is the linear classifier perceptron using a linear support vector machine? If not, they seem to be attacking the same problem.

[@chkashyap98]

just what is required..!,Hope more videos come up...!

[@OscarAlsing]

Glad you liked it!

[@chkashyap98]

Oscar Alsing can you also make a video on regression?

[@OysterNoise]

Why have random numbers for the weights? If they're going to be adjusted anyway, why not start with the same values for the weights every time?

[@OscarAlsing]

To minise the risk of getting stuck in local minimas

[@Mrloulouof44]

If you have the same weights at the beginning, then each neuron produces the same error. Hence the backpropagation algorithm will modify all the weights in the same way. By doing so, the model will not learn properly. What it means is that if you start with the same weights everywhere, you are unable to see which weight is too low and which weight is too high. Have a look at backpropagation to understand exactly why this is happening: kzbin.info/www/bejne/iXStmXWdrsSgl9U

[@goldenagestreams]

Very helpful, thank you Oscar!

[@OscarAlsing]

Glad I could help 😊

[@ghettomusic132]

hottest ML teacher

[@OscarAlsing]

haha, thanks!

[@kindeep]

That one line that didn’t get straightened...

[@mcasualjacques]

we could say that if there's intelligence in A.I. it's in the data, not in the mathematical device called neural network or perceptron though since the data can contain items where the same input values may produce different output values .... maybe the data isn't that intelligent and it's not necessary noise in the data acquisition i guess it just means the data has more inputs than that set of inputs one of those undeclared inputs can be randomness, or something not-random i wonder if adding a random input in the training data ..... nah it would just produce noisy outputs

[@AutomaticPizza]

Doesn't explains the conection between the line and how the waights are chosen.

[@mmanuel6874]

I too was sent here by a cat and his master

[@OscarAlsing]

Say hello to the cat and the master from me

[@nullciphero]

I thought Elon Musk was gonna explain it

[@smokabababylon]

Wow, thank you for sharing your knowledge. Brillant ! Greetings from France.

[@OscarAlsing]

Cheers ❤️

[@influenciaindebida]

10:57 mechine, categorize a platypus

[@ytubeanon]

perhaps this was noticed during the last 5 years... but, 'linear' is pronounced 'lin-ee-ear', rhymes with 'gin-ee-ear'

[@apreasher]

CatOverflow

[@TheXxBananeroxX]

hi i need your help, i had my exam one month ago and it didnt go very well because one of all the questions was "What type of space hypothesis does the perceptron use?" i answered "well it is a linear classifier so it is like if it draw a 'line' for every possibile set of input" and the teacher didnt like my answer, i was sure about what i was saying and he continued saying that it is not correct, i got nervous from that moment so he offered me 23/30, i rejected it because i studied and i feel i can do it better, well i need help of answering this question because i will have my exam again in 2 weeks, how would you anwser do that question? thank you for your time

[@OscarAlsing]

Hi Steven, First of all, thanks for taking your time and asking the question. I'm sure a lot of people have been wondering the same thing. The hypothesis space is the set of linearly bounded half-spaces as described here: courses.engr.illinois.edu/cs440/fa2011/lectures/Lecture21.pdf Hope this helps!

[@TheXxBananeroxX]

yeah it helps thank you

[@saileshpatra2488]

Perfect man. This is awesome 🤘✌

[@zoumanakeita8016]

Thank you for this great explanation Oscar !

[@OscarAlsing]

Glad I could help Zoumana!

[@roxanabrown5306]

Very helpful! Thank you.

[@OscarAlsing]

Glad I could help Roxana!

[@crazymexicandope]

That is a beautiful shirt. Where did you get that?

[@OscarAlsing]

Thank you :) It's from Zara, I think. Unfortunately it had a bit of a rough time in the laundry and is now resting in the shirt graveyard.

[@prayagcbose3862]

hay i might seem silli for asking this but in terms of supervised learning, we allreadt know the desired output then why do we have to train a perceptron in the first place....i know it sounds doump but it's a genuine doubt of mine..

[@OscarAlsing]

Yes, you train your perceptron on a set of data where you know the answer. The reason for doing this is so you can predict on data you haven't seen before, and want to predict the desired output.

[@prayagcbose3862]

@@OscarAlsing thanks for reply, but if we train a perception with AND logic function, it can only be used to predict that rite...? but we need all the 4 possibilitys of input output combination to train it in the first place. there comes my point, that if we know use all the outputs to train it, then what are we predicting...again correct me if I'm wrong....and thanks for reply...

[@OscarAlsing]

Please elaborate more 😊

[@MOhan-ur4ei]

hey oscar i just want to know how the XOR problem is solved, Loved to hear it from you

[@chasecolin22]

Look up Kernal Support Vector Machines. Its a method of drawing these separating hyperplanes, while morphing the space that these input features live on.

[@Roxor128]

The XOR function is (A AND NOT B) OR (B AND NOT A). If implementing with logic gates, you'd use two NOT gates, two AND gates, and one OR gate. A perceptron is effectively an AND or OR gate with optional NOT gates on its inputs (negative input weights). You could use three of them to implement a two-layer network which can compute XOR.

I guess you forgot to wear your glasses

[@rajeshdansena]

Hey, I found this very helpful to understand from basic.. Going forward, Can you please provide the algorithm and implemented code ?

[@OscarAlsing]

Hi Rajesh, I'm glad you enjoyed it. Great idea! Might do that in upcoming videos. Stay subscribed and you'll get notified.

[@rajeshdansena]

Oscar Alsing sounds great. Implementation in Python would be great :)

[@OscarAlsing]

Always Python

[@skyinnights]

OK I know this is a long shot. But if I send you a Presentation on Neural Networks would you be able to explain it? (just some text explanation would be great too, I don't need videos) I have exams and I really can't find any related videos or lectures. Thanks.

[@OscarAlsing]

Hey Sara. Would love to help but unfortunately I don't have the time for such vast explanations. Best of luck on your exam!

[@skyinnights]

Oscar Alsing Thanks 👌

[@OscarAlsing]

And if you have any specific questions dm me on Instagram @oscaralsing

[@skyinnights]

Oscar Alsing Sure. Thanks!

[@pawanjoshi3495]

great man

[@kendo3756]

i cant concentrate on the tutorial, i keep getting lost in to this mans eyes as well as his chest hair

[@ErezWeiss0]

Just Thank You!!!!!!!!!1

[@OscarAlsing]

Glad to help!

[@pavanvankadaru8965]

Is there a vedio you can make on when to use what neural network models?

[@OscarAlsing]

Not yet, great idea!

[@mistermiyagi6073]

solve some sample exam questions, this doesn't help at all in exam. i watched this video , then went to exam and the question was so different....

[@WilliamWesterlund]

Very well explained Oscar!

[@assortedtea902]

I down-voted this video due to clickbait Elon musk thumbnail.

[@OscarAlsing]

You're a true hero Mayuresh, let's save the world together

[@mohdamermohd6675]

You r awesome

[@jafelaviole8305]

at least he try to explain.....

[@joeallen2286]

you know whats a shame, the fact that you could have had some quality content but was misguided by the thumbnail. Now you know why your dislikes has +1...

[@iLorde]

so x1 = w1, x2 = w2 , and so on? (xn = wn?)

[@benoit1196]

no, it is more like: x1*w1 + x2*w2 = Perceptron input

[@jeanayme70]

Oe profe Ferdinand no sea pendejo ponga videos en español :v no entiendo ni mrd u_u

[@Game_Lab_Germany]

just meow

[@cypherllc7297]

U look like ultron

[@OscarAlsing]

Thanks, I guess!

[@cypherllc7297]

Oscar Alsing that's a compliment

[@Dassmonstret]

Lajneuer klassifyer

[@alexcipriani6003]

whats a "lainer" classifier ?

[@AliAhmad-ki2po]

XD....I was confused till the end of the video ...than I get that what he's saying is linear

[@johndalton4559]

la niere hahaha

[@sunshinebabe6203]

can't concentrate. Instructor too hot

[@anjovas]

Do you really have the impression your explanation was simple?

[@lukasdias9766]

CatOverflow