This is exactly the idea that Isaac Asimov had in 70's, I highly recommend you to read his short essays The New Teachers and Whatever You Wish (just google them, it's just few minutes reading really). For me it was mind-blowing that he wrote them some 40 years ago and yet they were so accurate! The future is now, baby :-)
Step 1: Embrace Mathematics. Step 2 : Master Python & Computer Science Fundamentals. Step 3 : Start Mathematical Modelling and Take up Projects. Step 4 : Be Patient and Above all Dont loose hope.
Alongside 2, learn to use numpy. This allows you to do things like matrix multiplication a lot faster using their API because they implement it in C, which is compiled versus interpreted. Otherwise, your training could take hours
You know what's wild man. I just watched this video again a few months into my learning ML to see if this made any more sense. It does. Love the content.
I don't really know how to describe what I just watched. You are brilliant. It's like you were a tour guide who was given, ahead of time, all of the questions your audience was going to ask. Then, without missing a beat, you remixed and weaved like 7 or 19 super intricate songs into a melody for your audience's consumption. I consider a lot of the tutorial content I consume black and white, or more black black black white white white black black white etc etc. A lot of content producers fail to realize that the grey area, between the spaces on the black and white comfort checkerboard...is where the real opportunity to teach exists. I'll watch 5 videos on the same exact thing and during each video I'll be helpless not to ask the same question. You, my friend, are very fucking good at answering those types of questions. It's like....you are a machine.
I've got to respectfully disagree. Calculus and linear algebra are A. Boring. and B. Really hard for most people to "just learn." (doubly so for people like me with discalculia.) Instead I would say "A really good way to learning lin-alg and calc is to apply it to a problem you are interested in, like machine learning." Your point reminds me of people who say the best way to learn to program is to start with transistors and slowly work your way through first principles to assembly language and then C. For most people this is super-hard mode with a lot of frustration and very little pay off.
My experience was pretty much the same. Except it rapidly became clear that I’d forgotten nearly everything about maths - so I had a two-year detour to teach myself maths again before coming back to ML. I’m so glad I did because I now have a much richer understanding of the algos. Love the channel mate.
I have a learning dysfunction, I had to drop out of school (twice) because I just didn't understand how to use what I was being given to actually learn. This video is amazing. Thank you for putting 2 and 2 together for me on like, how to even learn! I know this was meant as a really specific thing, but like, I get it now. I understand the steps I can take to learn! 100% sincere
I started with basic, then worked my way up to data exploratory analysis (numpy, pandas, seaborn, matplotlib, etc. ), then I started going into machine learning and found myself going back and forth between statistics and coding. Yeah.. You need strong foundation in math, otherwise, you wouldn't know how to optimize your model to get the best possible prediction. That's my 2 cents.
I started with AndrewNg (Best Course for me!) then go for Jeremy Howard - Deep Learning for Coder (Taught me how to code neural net) and the read Kaggles Notebooks written by others...
Aasim Baig Kaggle is a pretty cool site, I haven't explored it completely but I subscribed to their newsletters and they have pretty interesting data manipulation contests. You can see the community's submissions and some explain their Python code pretty well. It works to see real use cases which is more helpful than watching someone just keep writing code nonstop on video.
Bro! You are amazing! Noob to computer science and programming here, with a lil background in maths and physics, so I m literally starting something COMPLETELY new with lil to no background on the topic, and your attitude is just amazing... People like you always jump-start my motivation to keep going regardless of the objective complexity of the subject! Thanks and you just earned a new subscriber! Thanks for taking the time to do this! Big up :)
Literally had no idea how MC worked earlier today but after watching 3B1B's videos on Neural Networks I now understand the concept and the math behind it. Now I just need to reinforce that idea and trandlate it into code. This might only work for me since im a very fast visual learner but yeah I think he did a really good job at explaining it so props to 3B1B.
There has been multiple times where I found my self not thinking I could learn machine learning because of all the new and different things. It is great to see i am not alone and that if I push through I can make it.
Mr. Jabrils. I did what you said and I create a NN to detect contrast on an RGB input. Took me a few hours once I understood concepts of feed forward and backward propagation. First issue I had was the inputs I gave it (0-255) did not work with a sigmoid so I had to normalize the inputs x/255 and everything worked out!
hey jabrill first year computer science student LOVE your channel btw. I'd love for you to get into the code, maybe as a side series or even another channel when you inevitably get big. much love keep it up.
You did a great job introducing the "how to learn" aspect of what I think stumps a lot of intimidated curious learners. Basically, when you explained "When I saw this I didn't know what it meant, so I googled it." That is crazily enough, a big lesson to learn. Just hearing that at least is encouraging me to take my first independent steps to learning this stuff. So of course, thanks for the video!
Nice, you make me feel more confident about this stuff.. Yes it still looks like a mountain, but your whole message is that anything is doable one bite at a time. Thanks for that! So encouraging!
I am really struggling with all these jargon right now, and was honestly starting to feel really discouraged. Listening to you talk about your own struggles with ML when you started out made me feel much better and made me hopeful that someday I'll be able to do it too. Now I'm ready to dive back in headfirst into the struggle!
For anyone that wants to start learning, there are lectures on youtube on deep learning and neural networks taught by professors at MIT that I've found very useful to understand the concepts and the lingo used
Thank you. I think my education goes well with ML or AI. I got a degree in psychology (neurobiology focus), math and CS. Just started pursuing python once more to go into data science
I'm so happy i found this video and your channel! This was so useful for me, I thought I was the only one who felt Andrew Ng was speaking Zarglar after the first couple of modules. You've inspired me to try again with the try-get stuck-research and repeat method. Thanks!
Please have a Zarglar desk figure made, so we can buy it to support you and remind us that it's simply something we will have to go to battle against. Your videos are a masterpiece. Thanks!
Woah, this video was epic! I think the work you're doing here is really important. I try to stop myself from slipping into the ol' Zarg langauge on the daily, glad to know I'm not alone in our fight! I think they have some sort of brain infiltration device or something.
Whoa. The LogicGod himself giant_neural_network. Thanks for stopping by man, without your beautiful breakdown series, I don't think this video would exist! 😀 - Jabril
Jabrils No way, I’m glad to be just a small part of it. This stuff is perpetually interesting and you make it even more so... glad to have found your channel! With that said I’ll end this sickening love session and just say I look forward to more of your ml projects!
I remembered this comment when I came across this article en.wikipedia.org/wiki/Curse_of_knowledge Cool to know there's a whole wiki page devoted to it lol.
This may be the best TH-cam tutorial ever. Making the process of learning this entertaining is fantastic. I had to go about it in the same way, and I *wish* there had been a video like this years ago... Sending this to my son. Thanks!
Your the reason I’m pushing myself to finally learn to program man You make it look easy and simple and your style of doing things is inspiring to me! Thank you Jabrils!
@@bierzuip7169 OpenAI have chosen dota because its waaaaaay more complex than lol. If you try to argue which one is better like a 12 y old then maybe you are not mature enough to understand that nowadays these are two totally different games.
@@Kolano666 They aren't... The greed consumes Dota, that's it, i played the game from the start for five years, and in Second seasson of league i quit dota... League constantly grab the players of moba, in certain way, dota becames footsal, and lol football, but just in terms of popularity, cause the main function is the same... Diferent moba is the blizzard one, Smite, Dota and Lol are even... Just because they are lazy and stuborn, the icefrog don't give, blizzard don't give, so neither of them win (Just a few who quit icefrog and start league project).
Dude I found out about your channel about a week ago and I'm already hooked on Machine Learning! Keep up the good work and keep that chill vibe you got going on.
Machine learning is extremely dense in calculus (calculus 2 and 3 mostly) and matrix/ linear algebra. If you’re taking a computer science in college, taking the extra step to take those classes on the side if not already can def help you grasp the fundamentals.
Great video Jabrils! I am also learning machine learning online right now. But I am coming from a heavy math background, but not as strong programming background. Personally for me learning by teaching works best. So I am putting up tutorial detailed jupyter notebooks on my Github for the topics I learn.
Yeah I agree with you there, & obviously I have never walked the path, but I Think you should have an easier time coming from Math > CS. Best wishes! - Jabril
Yahh! Hamad Coming from Math > CS is relatively easy. If you are passionate, you are definitely gonna rock. I am from CS > Math background, trust me it is really harder than that. But in the end all struggle pays out, now I am a Data Scientist, hoping to learn new things from every one of you near future. Jabril, you are one of the best youtuber for your unique way of presentation. Good luck buddies, keep shining. - Shaon
Basically to sum up how to learn anything within the domains of Math and CS: -Use textbooks on relevant topics along with online MOOCs that many of the textbooks work together with. There are many specific books that are good/ great, but the absolutely best CS book for anyone with a basic understanding of Calculus (Derivatives, Integrals, DE) and not much else is Guttag's textbook as an introduction along with the accompanying free course on edX. Most high school kids here will try to jump into ML without even knowing Python, so definitely learning CS rigorously from this book before trying that. Apart from just Calculus, basic knowledge of Linear Algebra is important. For this I would recommend either Strang for a complete beginner or Axler's Linear Algebra done right for people with more theoretical aspirations. With the standard Calc course, a CS course, and a Linear Algebra course, you now have the prerequisites to dive into ML. The books I'd recommend for beginners level ML are: -Machine Learning (Mitchell) -Pattern Recognition and Machine Learning (Bishop) -Machine Learning: A Probabilistic Perspective (Murphy)
I’ve taken cs courses for my major and have a math minor now looking into machine learning. The “jargon” seems easy to me and I’m super excited to start my ML journey!
I had learned Matrix multiplication in grade 7 and finished linear algebra and statistics in the 12th grade. Four years later I graduating with a CS degree and a minor in mathematics. I opened a book on machine learning later and saw the formulas and gave up. It sounds crazy, but it took Andrew Ng telling me that it is just linear algebra to get me to realize I could learn the stuff. Still super hard. And debugging is so much harder that any other coding I've done but Jabril is right, you just have sit down and learn to read the crappy mathy pseudocode because all the good learning materials use it.
Brilliant video man, you pretty much laid out a great moethod of how to learn things independently rather than getting it bashed in to your head with school or courses. I've had that same problem in university where I just stopped understanding all of the jargon and couldn't digest the information anymore. But actually spot on with what you said, today I realized I didn't know almost anything about consensus protocols and found a good resource to explain it to me. Awesome job in bringing out a well structured approach to independent learning that would have been otherwise completely missed.
Im already familiar with python but never was interested in ML until I discovered your channel. I just wanna say you’re such an inspiration because of you I started learning more, so thank you. Your channel is so underrated! Would love to see your channel get bigger. ❤️❤️
So, let's wrap it up : Step 1: Learn math as you go (With the right resources and practice it's a lot easier than you think) Step 2: Learn programming starting preferably with Python or R Step 3: Eat a lot of sweets (candies, chocolate bars, etc...) Step 4: Practice with tangible and useful projects from day 1 Step 5: Teach what you've learned to others to keep it sharp and make sure you are actually understanding what you do Step 6: Subscribe to Jabrils to stay informed, challenged and keep leveling up your ML Skills
This video just justified all the mathematics that schools teach for Comp Sci degrees. Matrix multiplication, Summation = College Algebra. Derivatives = Calculus. Nice video, thanks!
There are unis in my country called ' preparatory school ' in which you study math and physics for 2 years and then take a hard exam to get into engineering schools ....so andrew's zanglar is understandable for me , besides that if anyone should learn machine learning you have to learn it from scratch and not like a lot of people who jump on kaggle and just copy paste codes ....if you are intrested start with andrew NG course then migrate to kaggle to practice , you should also learn python and tensor flow ...keep trying and never give up
I created my own machine learning network from scratch through a 12-byte code tester and it sequentially will figure out every code that is valid or invalid within a few seconds through a basic algorithm that many people use to solve Sudoku puzzles.
I think you want to be very specific about what it is you want to achieve with machine learning. I don't condone starting learning machine learning from a general and strictly statistiscal/mathematical based perspective if you are doing applied work. You always start with something specific so over the time you build up intuition, an intuitive understanding of how things work. If you want to apply machine learning to industrial and applied finance/investment and asset management, then you're mostly likely to try to use it to forecast return. Then you want to start from learning tradional linear regressioin models, the standard stuff you learn from any graduate econemtrics class. Once you're comfortable with that, you are ready to tackle machine learning and deep learning stuff. I have a feeling machine learning is an optimization toolbox in a forecast setting, which can surely be applied to other AI related subjects. In one word, asking how to get started with machine learning is like asking what programming language to learn for someone who has not learned any before. Well, it depends upon what it is you want to acoomplish with this language. Having an intuitive understanding is far more important than being able to derive stuffs, for most of us who are concerned with applied work. You can always, at a later time, indulge yourself with a bit of exploration in a maze of maths. In industry, maths work is left for PhD in Statistics, Maths and CS to do. They get paid very well, but not nearly as well as those who are in charge of them, usually someone with a PhD in Finance or any subject other than Maths and Statistics. Leave the dirty job to the mathematicians. Try to understand what they do, a little bit of how they do it, mostly importly, be able to build your work based upon theirs. With that being said, it never hurts to understand quite a bit of maths and programming. In fact, the first job you land in industry is not being a manager, but an analyst like job where you'll spend a lot of time coding and debugging.
Just stumbled upon one of your videos on the homepage and I'm so lucky it showed up because this so relatable. One of my courses this semester is Intelligent Systems (AI/ML) and I'm having a tough time and I REALLY want to get into it but the vocabulary and topics are just hard to understand ahaha. I feel more driven to study more after watching a couple of your videos. I hope your channel grows more and gets the attention it deserves. I really like your style. Keep it up man I'll be following along! :)
This video makes me so happy! Last year i was in highschool (i studied languages, so my math knowledge was like 0) and everytime i tried studying something software related i always got frustrated. Now after the firsts six months of university (and after calculus 1 and linear algebra) i can finally understand this mathematical mumbo jumbo! I dunno i feel happy
I just found your channel. Definitely dig it. I'm an engineer in the field starting my own ML journey. Thanks for sharing! I should add that I can totally relate to the Zarglar thing. I don't have a CS degree or even a formal education. I'm 100% self-taught.
Iam so happy,I just found u r video this is really helpful to me,I watched many videos of neural networks & machine learning,how to begin ,how to learn, everything is bored me but u r video is osum ...,when u started ml ..I did same. As u ,but now I can do .. I know how &where to start ..I know but it's take more time ..I can't leave without like/sub ...u r rocking man..
I am 15 year old. I learned programming from CS50x Intro To CS. Then I did a basic AI course named CS50AI. Then I learnt discrete maths, some basic linear algebra, some basic multivarianle calculus. Now I am doing ML by Andrew Ng, Competitive Programming, Complex Analysis, 3 books.(Art and Craft of Problem Solving, Intro to Algorithms by CLRS. After I will properly finish linear algebra and multivariate calculus. This progress is of 3-4 months. The lingo thing this guys says is true if you don't have a math background. You should watch some Linear Algebra 3b1b, multivariate calculus videos from Khan Academy (taught by 3b1b) If you don't know single variable calculus(MITOCW is good) . You should learn it before linear algebra and multivariate calculus.
You don't exactly need to be from a math background, matrix multiplication is basics 10th grade math and statistics and regression is taught in every engineering college
Subscribed. I'm a Web Developer, but this stuff has always been interesting to me. The description, personality and editing hooked me. Well done! Round of Applause!
I can't help but feel that the way Jabrils types things on his computer is by just making general gestures towards his keyboard, and the words appear on the screen.
![The moment we stopped understanding AI [AlexNet]](/img/n.gif)
Chillest programming channel i've ever stumbled upon.
Good shit bro, keep it up.
Agreed. So smooth, functional
It's good, why do you call it "shit" ?
Indo memek?
@@noopyx3414 in this context, it's slang for "stuff" ;)
also dani
Step 1: don't be afraid of math
But I am so trash at math what do I do
Same question here.
@@slimshady1006 Practice a lot, a lot, a lot. Math is a language, and like any spoken language it takes time and regular practice.
@@slimshady1006 MacDonald's
@@slimshady1006 Math is a language, the same way you understood how to speak English. I am very suspicious of people who say theyre bad at math.
If machines are so good at learning, maybe they can start teaching soon!
nice to see you here!
Hey there Niels, likewise :D
Machines can't use guns though
youll see
This is exactly the idea that Isaac Asimov had in 70's, I highly recommend you to read his short essays The New Teachers and Whatever You Wish (just google them, it's just few minutes reading really). For me it was mind-blowing that he wrote them some 40 years ago and yet they were so accurate! The future is now, baby :-)
Step 1: Embrace Mathematics.
Step 2 : Master Python & Computer Science Fundamentals.
Step 3 : Start Mathematical Modelling and Take up Projects.
Step 4 : Be Patient and Above all Dont loose hope.
on point.
What are cs fundamentals ?
Sai Ram algorithms, data structures, etc there are great videos on TH-cam explaining cs fundamentals
Alongside 2, learn to use numpy. This allows you to do things like matrix multiplication a lot faster using their API because they implement it in C, which is compiled versus interpreted. Otherwise, your training could take hours
Sai Ram open any technical university program and you will see
You know what's wild man. I just watched this video again a few months into my learning ML to see if this made any more sense. It does. Love the content.
I don't really know how to describe what I just watched. You are brilliant. It's like you were a tour guide who was given, ahead of time, all of the questions your audience was going to ask. Then, without missing a beat, you remixed and weaved like 7 or 19 super intricate songs into a melody for your audience's consumption. I consider a lot of the tutorial content I consume black and white, or more black black black white white white black black white etc etc. A lot of content producers fail to realize that the grey area, between the spaces on the black and white comfort checkerboard...is where the real opportunity to teach exists. I'll watch 5 videos on the same exact thing and during each video I'll be helpless not to ask the same question. You, my friend, are very fucking good at answering those types of questions. It's like....you are a machine.
best way to start learning calculus and linear algebra. there is little point in doing anything else before you have a solid grasp of the basic math
agree. after learning calculus and vectors in school it started making so much more sense
I was surprised people want to start Machine Learning without even knowing what a matrix is, must make the journey a lot harder
I've got to respectfully disagree. Calculus and linear algebra are A. Boring. and B. Really hard for most people to "just learn." (doubly so for people like me with discalculia.) Instead I would say "A really good way to learning lin-alg and calc is to apply it to a problem you are interested in, like machine learning." Your point reminds me of people who say the best way to learn to program is to start with transistors and slowly work your way through first principles to assembly language and then C. For most people this is super-hard mode with a lot of frustration and very little pay off.
Simon I agree. You don't need to be an expert in calculus and linear algebra, but learning at least the basics is vital.
Simon check out 'euprime vector space' on TH-cam
My experience was pretty much the same. Except it rapidly became clear that I’d forgotten nearly everything about maths - so I had a two-year detour to teach myself maths again before coming back to ML. I’m so glad I did because I now have a much richer understanding of the algos.
Love the channel mate.
I love your way of simplifying verbose and dense topics into concise understandable videos
I have a learning dysfunction, I had to drop out of school (twice) because I just didn't understand how to use what I was being given to actually learn. This video is amazing. Thank you for putting 2 and 2 together for me on like, how to even learn! I know this was meant as a really specific thing, but like, I get it now. I understand the steps I can take to learn!
100% sincere
wait, you're telling me all the calculus and linear algebra I learned. I can finally apply in life?
I mean all the math you learned was born to be used in gaming
@@melonman1276 makes sense
I started with basic, then worked my way up to data exploratory analysis (numpy, pandas, seaborn, matplotlib, etc. ), then I started going into machine learning and found myself going back and forth between statistics and coding. Yeah.. You need strong foundation in math, otherwise, you wouldn't know how to optimize your model to get the best possible prediction. That's my 2 cents.
I've learnt python well and have a good grasp on the math. Now what
I started with AndrewNg (Best Course for me!) then go for Jeremy Howard - Deep Learning for Coder (Taught me how to code neural net) and the read Kaggles Notebooks written by others...
Aasim Baig Kaggle is a pretty cool site, I haven't explored it completely but I subscribed to their newsletters and they have pretty interesting data manipulation contests. You can see the community's submissions and some explain their Python code pretty well. It works to see real use cases which is more helpful than watching someone just keep writing code nonstop on video.
Bro!
You are amazing!
Noob to computer science and programming here, with a lil background in maths and physics, so I m literally starting something COMPLETELY new with lil to no background on the topic, and your attitude is just amazing...
People like you always jump-start my motivation to keep going regardless of the objective complexity of the subject!
Thanks and you just earned a new subscriber!
Thanks for taking the time to do this!
Big up :)
3Blue1Browns has a very good tutorial on machine learning
Annie Möff agree
He has a good breakdown and explanation...but not necessarily a tutorial one might suspect
The One And Nearly Only his series is more of something to watch after a tutorial to understand in a better way.
i know but he is more of a visvual but not any real code or way to write the code.
No, he briefly explains neural networks and backpropagation.
Literally had no idea how MC worked earlier today but after watching 3B1B's videos on Neural Networks I now understand the concept and the math behind it. Now I just need to reinforce that idea and trandlate it into code. This might only work for me since im a very fast visual learner but yeah I think he did a really good job at explaining it so props to 3B1B.
Now i finally understood "Silence speaks better than words...." 😂👌❤
There has been multiple times where I found my self not thinking I could learn machine learning because of all the new and different things. It is great to see i am not alone and that if I push through I can make it.
Mr. Jabrils. I did what you said and I create a NN to detect contrast on an RGB input. Took me a few hours once I understood concepts of feed forward and backward propagation. First issue I had was the inputs I gave it (0-255) did not work with a sigmoid so I had to normalize the inputs x/255 and everything worked out!
Yo I am yell at the wall YES! I am proud of you DDSTAR! KEEP GOING!
- Jabril
how did u train it?
So are we not gonna talk about how he searched google IN google ? 2:25
Edit: I know it's a joke, I'm just pointing it out
Lol, makes me cringe when I see people do that at school.
Photonic Tech HAHAHAHAHAHA
@@PKNproductions he is doing it while he sayin "quick google search" and thats the joke bro
that's the joke
meta
hey jabrill first year computer science student LOVE your channel btw.
I'd love for you to get into the code, maybe as a side series or even another channel when you inevitably get big. much love keep it up.
I'm working on something. Hopefully I can get it out soon. Thanks for the support john.
- Jabril
You did a great job introducing the "how to learn" aspect of what I think stumps a lot of intimidated curious learners. Basically, when you explained "When I saw this I didn't know what it meant, so I googled it." That is crazily enough, a big lesson to learn. Just hearing that at least is encouraging me to take my first independent steps to learning this stuff. So of course, thanks for the video!
You summed it up perfectly Dan!
Nice, you make me feel more confident about this stuff.. Yes it still looks like a mountain, but your whole message is that anything is doable one bite at a time. Thanks for that! So encouraging!
I am glad to help adam!
- Jabril
I am really struggling with all these jargon right now, and was honestly starting to feel really discouraged. Listening to you talk about your own struggles with ML when you started out made me feel much better and made me hopeful that someday I'll be able to do it too. Now I'm ready to dive back in headfirst into the struggle!
Those gloves though! 😍
Your videos keep getting better and better - keep it up man! :)
Sykoo no worry about the gloves hes been eating laffy taffy the whole time
Sykoo my Unity guy
Yaay unity
This guy reminds me of Tyler the Creator, he acts dumb and wears stupid gloves but is low key a genius!
This guy is a real fraud.
For anyone that wants to start learning, there are lectures on youtube on deep learning and neural networks taught by professors at MIT that I've found very useful to understand the concepts and the lingo used
look for: MIT 6.S191
Thank you. I think my education goes well with ML or AI. I got a degree in psychology (neurobiology focus), math and CS. Just started pursuing python once more to go into data science
I love how you act what you talk, this format is so unique. LOVE IT, also the content is gold. Thank you Jabrils!
I'm so happy i found this video and your channel! This was so useful for me, I thought I was the only one who felt Andrew Ng was speaking Zarglar after the first couple of modules. You've inspired me to try again with the try-get stuck-research and repeat method. Thanks!
Glad to see other poc in the ML/DS space. Keep it up ✊
We're coming up✊🏽🙌🏽all love here
Please have a Zarglar desk figure made, so we can buy it to support you and remind us that it's simply something we will have to go to battle against. Your videos are a masterpiece. Thanks!
What's Zarglar?
"Have you considered taking a course in machine learning at udemy?" These ads man... Smarter every day.
this channel should have more subscribers
try that
This is the best video for someone that knows nothing about machine learning. Thank you
Woah, this video was epic! I think the work you're doing here is really important.
I try to stop myself from slipping into the ol' Zarg langauge on the daily, glad to know I'm not alone in our fight! I think they have some sort of brain infiltration device or something.
Whoa. The LogicGod himself giant_neural_network. Thanks for stopping by man, without your beautiful breakdown series, I don't think this video would exist! 😀 - Jabril
Jabrils No way, I’m glad to be just a small part of it. This stuff is perpetually interesting and you make it even more so... glad to have found your channel! With that said I’ll end this sickening love session and just say I look forward to more of your ml projects!
I remembered this comment when I came across this article en.wikipedia.org/wiki/Curse_of_knowledge Cool to know there's a whole wiki page devoted to it lol.
th-cam.com/video/-SdyFenqT7c/w-d-xo.html
Easy way to understand what is neural network
This may be the best TH-cam tutorial ever. Making the process of learning this entertaining is fantastic. I had to go about it in the same way, and I *wish* there had been a video like this years ago... Sending this to my son. Thanks!
Wow, and I thought my math teacher (on the high-school) talks about something what I will never need! xO
All of this math stuff we have handled.
Your the reason I’m pushing myself to finally learn to program man
You make it look easy and simple and your style of doing things is inspiring to me!
Thank you Jabrils!
“The League of Legends Open AI” Triggered!
True man xd
triggered because dota can never be as good as league? or, the other way around lol...
@@bierzuip7169 OpenAI have chosen dota because its waaaaaay more complex than lol. If you try to argue which one is better like a 12 y old then maybe you are not mature enough to understand that nowadays these are two totally different games.
@@Kolano666 They aren't... The greed consumes Dota, that's it, i played the game from the start for five years, and in Second seasson of league i quit dota...
League constantly grab the players of moba, in certain way, dota becames footsal, and lol football, but just in terms of popularity, cause the main function is the same...
Diferent moba is the blizzard one, Smite, Dota and Lol are even...
Just because they are lazy and stuborn, the icefrog don't give, blizzard don't give, so neither of them win (Just a few who quit icefrog and start league project).
@@davidchristenes9062 Dota>lol for all eternity, dota is just way ahead of lol in almost every category, no debate
Dude I found out about your channel about a week ago and I'm already hooked on Machine Learning! Keep up the good work and keep that chill vibe you got going on.
I love your video editing Style. Great video.
Machine learning is extremely dense in calculus (calculus 2 and 3 mostly) and matrix/ linear algebra. If you’re taking a computer science in college, taking the extra step to take those classes on the side if not already can def help you grasp the fundamentals.
arlanzotoxic knowing stats can’t hurt either, and linear algebra is good for any field- the applications are everywhere.
Great video Jabrils! I am also learning machine learning online right now. But I am coming from a heavy math background, but not as strong programming background. Personally for me learning by teaching works best. So I am putting up tutorial detailed jupyter notebooks on my Github for the topics I learn.
Yeah I agree with you there, & obviously I have never walked the path, but I Think you should have an easier time coming from Math > CS. Best wishes!
- Jabril
Yahh! Hamad Coming from Math > CS is relatively easy. If you are passionate, you are definitely gonna rock. I am from CS > Math background, trust me it is really harder than that. But in the end all struggle pays out, now I am a Data Scientist, hoping to learn new things from every one of you near future. Jabril, you are one of the best youtuber for your unique way of presentation. Good luck buddies, keep shining.
- Shaon
th-cam.com/video/-SdyFenqT7c/w-d-xo.html
Easy way to understand what is neural network
I like this approach. Isolate what confuses you, learn it, then move on.
good thing im a cse student so i know all that, meaning i can start learning ml
lucky me !
This has to be one of the first times I've seen somebody say "you're gonna need to learn some maths". Great video!
Love your channel, you are the first youtuber in a long time that has inspired me and at the same time made me hit the like/sub/notif bell 😂
Basically to sum up how to learn anything within the domains of Math and CS:
-Use textbooks on relevant topics along with online MOOCs that many of the textbooks work together with.
There are many specific books that are good/ great, but the absolutely best CS book for anyone with a basic understanding of Calculus (Derivatives, Integrals, DE) and not much else is Guttag's textbook as an introduction along with the accompanying free course on edX. Most high school kids here will try to jump into ML without even knowing Python, so definitely learning CS rigorously from this book before trying that. Apart from just Calculus, basic knowledge of Linear Algebra is important. For this I would recommend either Strang for a complete beginner or Axler's Linear Algebra done right for people with more theoretical aspirations. With the standard Calc course, a CS course, and a Linear Algebra course, you now have the prerequisites to dive into ML.
The books I'd recommend for beginners level ML are:
-Machine Learning (Mitchell)
-Pattern Recognition and Machine Learning (Bishop)
-Machine Learning: A Probabilistic Perspective (Murphy)
Wow, many of the the maths he was talking about is being taught to us in intermediate School in my country...
I’ve taken cs courses for my major and have a math minor now looking into machine learning. The “jargon” seems easy to me and I’m super excited to start my ML journey!
You're awesome and inspiring. I love your style. Keep it up!
You just made machine learning feel like front-end development... THANK YOU
Thanks for this video! :)
This guy is a real fraud.
salve FV
I had learned Matrix multiplication in grade 7 and finished linear algebra and statistics in the 12th grade. Four years later I graduating with a CS degree and a minor in mathematics. I opened a book on machine learning later and saw the formulas and gave up.
It sounds crazy, but it took Andrew Ng telling me that it is just linear algebra to get me to realize I could learn the stuff. Still super hard. And debugging is so much harder that any other coding I've done but Jabril is right, you just have sit down and learn to read the crappy mathy pseudocode because all the good learning materials use it.
i love your attitude towards learning new things! taught me a lot
Brilliant video man, you pretty much laid out a great moethod of how to learn things independently rather than getting it bashed in to your head with school or courses. I've had that same problem in university where I just stopped understanding all of the jargon and couldn't digest the information anymore. But actually spot on with what you said, today I realized I didn't know almost anything about consensus protocols and found a good resource to explain it to me. Awesome job in bringing out a well structured approach to independent learning that would have been otherwise completely missed.
i just discovered your channel, u r great dude!
Im already familiar with python but never was interested in ML until I discovered your channel. I just wanna say you’re such an inspiration because of you I started learning more, so thank you. Your channel is so underrated! Would love to see your channel get bigger. ❤️❤️
So, let's wrap it up :
Step 1: Learn math as you go (With the right resources and practice it's a lot easier than you think)
Step 2: Learn programming starting preferably with Python or R
Step 3: Eat a lot of sweets (candies, chocolate bars, etc...)
Step 4: Practice with tangible and useful projects from day 1
Step 5: Teach what you've learned to others to keep it sharp and make sure you are actually understanding what you do
Step 6: Subscribe to Jabrils to stay informed, challenged and keep leveling up your ML Skills
Wow! I love how you picked learning one thing after the next wherever it was needed, critical skill for the 21st century. Respect!
i'm a computor
help is on the way.
- Jabril
This video just justified all the mathematics that schools teach for Comp Sci degrees. Matrix multiplication, Summation = College Algebra. Derivatives = Calculus. Nice video, thanks!
2:35 he was the first to make it to Area 51! :O
There are unis in my country called ' preparatory school ' in which you study math and physics for 2 years and then take a hard exam to get into engineering schools ....so andrew's zanglar is understandable for me , besides that if anyone should learn machine learning you have to learn it from scratch and not like a lot of people who jump on kaggle and just copy paste codes ....if you are intrested start with andrew NG course then migrate to kaggle to practice , you should also learn python and tensor flow ...keep trying and never give up
This video is super helpful I really appreciate you putting the time into making it.
I created my own machine learning network from scratch through a 12-byte code tester and it sequentially will figure out every code that is valid or invalid within a few seconds through a basic algorithm that many people use to solve Sudoku puzzles.
My first time seeing him moving his lips as he talked
when?
Hands down best computer programming TH-cam channel ever.
2:36, TH-cam's AI is better than I thought. It knows zarglar!
I think you want to be very specific about what it is you want to achieve with machine learning. I don't condone starting learning machine learning from a general and strictly statistiscal/mathematical based perspective if you are doing applied work.
You always start with something specific so over the time you build up intuition, an intuitive understanding of how things work. If you want to apply machine learning to industrial and applied finance/investment and asset management, then you're mostly likely to try to use it to forecast return. Then you want to start from learning tradional linear regressioin models, the standard stuff you learn from any graduate econemtrics class. Once you're comfortable with that, you are ready to tackle machine learning and deep learning stuff. I have a feeling machine learning is an optimization toolbox in a forecast setting, which can surely be applied to other AI related subjects.
In one word, asking how to get started with machine learning is like asking what programming language to learn for someone who has not learned any before. Well, it depends upon what it is you want to acoomplish with this language. Having an intuitive understanding is far more important than being able to derive stuffs, for most of us who are concerned with applied work. You can always, at a later time, indulge yourself with a bit of exploration in a maze of maths.
In industry, maths work is left for PhD in Statistics, Maths and CS to do. They get paid very well, but not nearly as well as those who are in charge of them, usually someone with a PhD in Finance or any subject other than Maths and Statistics. Leave the dirty job to the mathematicians. Try to understand what they do, a little bit of how they do it, mostly importly, be able to build your work based upon theirs.
With that being said, it never hurts to understand quite a bit of maths and programming. In fact, the first job you land in industry is not being a manager, but an analyst like job where you'll spend a lot of time coding and debugging.
Wow, I didn't know I could get messages from telepathic people over the internet.
This dude doesn't move his mouth to talk once, He Only Chews!
Just stumbled upon one of your videos on the homepage and I'm so lucky it showed up because this so relatable.
One of my courses this semester is Intelligent Systems (AI/ML) and I'm having a tough time and I REALLY want to get into it but the vocabulary and topics are just hard to understand ahaha. I feel more driven to study more after watching a couple of your videos.
I hope your channel grows more and gets the attention it deserves. I really like your style. Keep it up man I'll be following along! :)
I’m so glad I found you!
I'm so glad to have your support!
- Jabril
This video makes me so happy! Last year i was in highschool (i studied languages, so my math knowledge was like 0) and everytime i tried studying something software related i always got frustrated. Now after the firsts six months of university (and after calculus 1 and linear algebra) i can finally understand this mathematical mumbo jumbo!
I dunno i feel happy
So i wanna know how to get started with ML...
Me:
*YES!*
Can I just say that your way to present your content is freaking amazing! I'm mad that I just found your channel. You’re great man!
you and me both had the same inspirations you went forward and i was laughing at Elon musk smoking weed
Updating resume to include Machine Learning 9 months experience, cheers mate.
"Mr Stark, I don''t wanna go" 6:07
I just found your channel. Definitely dig it. I'm an engineer in the field starting my own ML journey. Thanks for sharing!
I should add that I can totally relate to the Zarglar thing. I don't have a CS degree or even a formal education. I'm 100% self-taught.
Iam so happy,I just found u r video this is really helpful to me,I watched many videos of neural networks & machine learning,how to begin ,how to learn, everything is bored me but u r video is osum ...,when u started ml ..I did same. As u ,but now I can do .. I know how &where to start ..I know but it's take more time ..I can't leave without like/sub ...u r rocking man..
woah this video gave me a total new insight on this field
ive been trying to understand all this zarglar without actually searching it up
Zarglar looks like my sleep-paralysis demon ngl.
What's Zarglar?
I am 15 year old. I learned programming from CS50x Intro To CS. Then I did a basic AI course named CS50AI. Then I learnt discrete maths, some basic linear algebra, some basic multivarianle calculus. Now I am doing ML by Andrew Ng, Competitive Programming, Complex Analysis, 3 books.(Art and Craft of Problem Solving, Intro to Algorithms by CLRS. After I will properly finish linear algebra and multivariate calculus. This progress is of 3-4 months.
The lingo thing this guys says is true if you don't have a math background. You should watch some Linear Algebra 3b1b, multivariate calculus videos from Khan Academy (taught by 3b1b)
If you don't know single variable calculus(MITOCW is good) . You should learn it before linear algebra and multivariate calculus.
Aa someone with a math background I feel so lucky hoho
You don't exactly need to be from a math background, matrix multiplication is basics 10th grade math and statistics and regression is taught in every engineering college
@@sharathnair1702 yes but engineering has a strong mathematical background. Some other degrees don't
@@sharathnair1702 i don't think anybody knows what a sigma sum is before they enter university
@@gilfhunter42069 Asians does!
I don't typically like or dislike videos, but I felt like you did a phenomenal job at making learning ML approachable. Nice work :)
How you don't have diabetes by this point is beyond me...
I have no idea how to program but I was impressed with the quality of your videos. Here’s a sub man.
So.... 100k = speaking reveal? Is the person talking even the person on camera?
WE ask the questions that matter
speaking reveal at 100K!
- Jabril
Learning about learning machine learning can't get more fun than this! Keep it up bro :D
2:37 had me dying lmao
When writing a paper we have to include the formulas as well. Omg so much work 😩 and reading 📖 This Gave me hope hahah
@2:24 DID HE ACTUALLY GOOGLE "GOOGLE.COM"??
I just found a channel that I'll probably be addicted to for the month.
Open AI League of Legends Project?. Triggered. It was Dota 2 sir. Dota 2.
i was looking for this coment bruhh feels good i am not alone
Me too😅
Subscribed. I'm a Web Developer, but this stuff has always been interesting to me. The description, personality and editing hooked me. Well done! Round of Applause!
busyrand im a web developer too! But most of this AI/ML stuff flies right over my head, but i plan on commiting and becoming good at it!
tfw when you start talking with moving your mouth at the end of the video and i get anxiety
I can't help but feel that the way
Jabrils types things on his computer is by just making general gestures towards his keyboard, and the words appear on the screen.
What's that blue stuff you're eating?
Airheads, they're out of control.
This is the most entertaining, creative and pedagogical explanation of a deep mathematical topic.
Outstanding !!!