Linear Algebra - Math for Machine Learning

In this video, W&B's Deep Learning Educator Charles Frye covers the core ideas from linear algebra that you need in order to do machine learning.
In particular, we'll see how linear algebra is not like algebra -- it's more like programming! And then we'll build on that intuition to understand why linear algebra is so central to machine learning.
Slides here: wandb.me/m4ml-linear-algebra
Exercise notebooks here: github.com/wandb/edu/tree/mai...
Check out the other Math4ML videos here: wandb.me/m4ml-videos
0:00 Introduction
1:29 Why care about linear algebra?
5:15 Linear algebra is not like algebra
7:53 Linear algebra is more like programming
14:31 Arrays are an optimizable representation of functions
18:01 Arrays represent linear functions
22:34 "Refactoring" shows up in linear algebra
25:19 Any function can be refactored
28:16 The SVD is the generic refactor applied to a matrix
33:51 Using the SVD in ML
38:15 Review of takeaways and more resources

[@xa9ax]

You never fail to impress me as an educator. This is such a good refresher. Kudos!

[@yunwang1243]

Who is this guy? It’s the best Linear Algebra in ML I could find! Better than all my professors

[@QuantumLegal]

Great course. It never ceases to amaze me how many so-called machine learning videos never tell them how much math you need to actually building neural networks or genetic algorithms etc.

[@honkhonk8009]

I have ADHD, but you managed to captivate me for so long holy shit. Goated video.
Im in first year rn and im tryna learn Linear Algebra.
The hardest thing to do in life, is to learn something off a textbook, and not even know HOW your gonna be using it.
You dont know what information is important, you dont know why somethings like that, and you basically end up stuck.
This really helped teach me linear algebra imo.
I find it impossible to learn stuff without first knowing the motivation and application of it lol.

[@Harduex]

Awesome! That's the first time that I actually get the logic of using matrices in the ML. Keep up the good work!

[@JJ_eats_wings]

always good to refresh my linear algebra!!

[@antopolskiy]

This was great, can't wait for more. I love your explanatory style, for me it threads the ideal boundary between too detailed and not detailed enough. Thank you!

[@charles_irl]

Thanks Sergey! That's exactly the boundary I try to walk, so it's really gratifying to hear that I did it right.

[@ninjaturtle205]

I think you have such a new way of presenting these ideas and concepts. This is insight that some people acquire through ages of learning and experience. But I still feel that these ideas need to be expanded upon, and fleshed out more for the average or advanced student. Please consider providing a further in depth series, going into each of LA, calculus, and prob/stats portions of the MATH4ML series.

[@proprius]

Thank you, this was very well explained!

[@Moiez101]

solid! can see and feel your passionate through the screen bro. Excited to go through this playlist. I just got hired as a junior data scientist but struggle with the math portion of machine learning especially linear algebra and calculus.

[@MCMelonslice]

Very helpful insight. Thanks 👍

[@asuzukosi581]

Wow, I'm 11 minutes in and this is the best explanation of linear algebra I've ever seen

[@tomthanhswe]

Thank you a lot for this math playlist

[@santiagonoya5702]

Good video! Really insightful

[@jjpp1993]

so exited!

[@albertog2196]

Loved this

[@giovannimazzocco499]

Brilliant explanation, very nice interpretation of matrix multiplication as a form of function composition.

[@cambridgebreaths3581]

Fascinating. Happy to subscribe

[@WeightsBiases]

Thanks for watching!

[@johnnovotny4286]

Charles impatient to let you know: you can get this too. Pure magic.

[@tam31433]

This is so coooool!!💪👍💪👍💪👍

[@rohitkundu2120]

Thanks for this

[@benjaminfindon5028]

5:30 ur right and i love it

[@mertbozkir]

Perfect lecture, compherensive explanation... I fall in love with W&B 💖💖

[@TechWithRushabh]

12:09 the matrix X can be named transformation_matrix ?

[@muhammadsuleimanhussain-jf9nz]

GOOD ONE

[@guzh]

The descriptions of matrices A, C matrices are very unclear. Hope you can add some examples.

[@udyan_upal]

this channel is epic

[@martinsanchez-hw4fi]

In 21:36 you say that elements outside the kernell remain outside under linear combination. That is not necessarily true, that is why we work with linear independence.

[@anormalperson3121]

I'm just starting this course, to anyone who has completed it; is it enough for me to get started with the actual machine learning content? Or will I need more math after this course?

[@_shery.]

Frye can you please state prerequisites for this series. I am starting my journey in machine learning

[@WeightsBiases]

Hello! I think basic knowledge of math and Python should be enough.

[@_shery.]

@@WeightsBiases ok, thanks

[@surendranmurugesan]

Is it "optimisation by programming" or "programming by optimisation"?

[@ran_domness]

Hmm lots of assumptions on prior knowledge. Would be good to spell out prerequisite knowledge necessary to understand. Thanks, good video.

[@robloxboy-sk2in]

I understand now

[@TheBlueHutch]

Do you need a calculator for Linear Algebra? or not

[@ikechiakwara6058]

8:09

[@web_resource]

15:39

[@stanlee3810]

Certainly not for beginners. Still good though

[@hueydo3522]

Horrible explanation on the SVD not gonna lie. So convoluted what you just say makes a complex problem even more complex when a convoluted concept really doesn't have an easy answer. I can see why you draw the similarities of code factorization but again the idea is not as nuance as that.