how much math do you NEED for machine learning?
Vložit
- čas přidán 25. 07. 2024
- In this video, we talk about how much math you will need for machine learning. It is a lot less than many will make you believe. I will cover the essential statistics, linear algebra, and calculus and courses where you can learn them if needed. Additionally, I will share my approach and favorite lookup book for all other questions. Finally, I reveal how much Mathematics I use in my daily life as a Machine Learning Engineer.
Resources Mentioned:
Mathematics for Machine Learning, book: mml-book.github.io/book/mml-b...
Statistics course Stanford: www.coursera.org/learn/stanfo...
Linear Algebra: www.coursera.org/learn/linear...
If you enjoyed this video, I would be excited to connect on Twitter or LinkedIn.
Twitter: / datawithsandro
LinkedIn: / sandro-luck-b9293a181
Medium: / datawithsandro
Intro: 0:00
Why Know Math in ML: 2:49
Favorite Book: 3:56
Statistics and Probability: 4:38
Linear Algebra: 6:38
Calculus/Analysis: 8:19
All Other Math: 9:00 - Věda a technologie
This was pretty helpful, tanks
One of the most informative videos in my lives, thx
Glad you think so!
That production was amazing in the beginning 😂😂😂🔥🔥🔥
I don't think math by itself is frightening. If you're frightened you should probably not be making this career change. However, the scary thing is the time required to sink INTO math so you can get into this field effective is what's scary. It's the opportunity cost, really.
Well put! Understanding that it won't happen overnight is a must
Hello, can you make a video on how to build various data loaders? I feel like everyone skips this part of ML teaching and whilst I go into a project knowing how I am going to build a model, i always spend hours sort of mindlessly figuring out how I am going to load my data in (Particularly images - E.G my most recent project uses Siamese network). I guess my question is, do you have any recommendations or a more 'engineered' technique to going about data loading? Thank you.
Great idea, I will have a look at what I can put together👌
Which type of math did you [hate/struggle with] the most? And do you still use it today? Share with others and let me know if you have any questions!
Thanks
great you enjoyed it
My God ! His notes !!
Well I was allowed to bring one page per exam, better fill that space I guess😂
Do a video about what do you think is a machine learning engineer.
Great Idea will look into it, spoiler alert the answer heavily depends on the company you ask😉
I procrastinated to break into this field for 2 years because if Math anxiety. I don't know whether I should start learning ML without math or I will study math alone before that.
You definitely should do it! Do both at the same time it is far more motivating to know what you can use it for in my experience, best of luck
@@datawithsandro2919 I will take this advice. Thanks
I've got a Junior ML Developer job recently. I only have 6 months of experience in Python Backend. I have economics degrees (bachelor's and master's). I studied statistics and math during university but nothing crazy just general stuff. So, you think, I can be an ML developer without insane math skills?
People say I probably won't build ML algos from scratch and just use TensorFlow / Keras and basically implement stuff.
Is this true? I'm scared to accept this offer since my math is kinda rusty...
I would say the truth is somewhere in the middle. A) yes you can be a ml developer without insane math skills, it really depends a bit which sub area of ml im general B) yes you probably wont be building completly new ml architectures from scratch, but really only ai/ml researchers/teams do this om a regular basis and even then I would say it's rarely more then 5-10% of the job, also in the commerical normal ml dev scene development/small scale research happens( but heavily depends on where you start and work)
get the badge in
Thank you very much for your suggestion. I really like ML and wished to be a Machine Learning Engineer , but I thought perhaps I should not choose this career as I never took any statistics course and took calculus and algebra long time ago (forgotten by now). Now this video given me some hope and inspiration. I hate integral calculus though , do Machine learning Engineers require integral calculus ability ? Thanks for the link of the book . Is there any book "statistics for machine learning " ?
You are very welcome and happy you found the video inspiring. There is a lot of hope and it is never too late if you are willing to learn. As mentioned in the video would not think of it as a hard requirement, in the end you should be fine with some basic knowledge and the ability to read formulas in papers, if you want to do applied ML. Additionally i really think that you can learn many of these aspects while programming/using other models. The book mention is mostly a combination of statistics and algebra(putting the two together at many points). If you feel like you have to relearn some of the concepts maybe first do a short course on statistics and linear algebra from then on you really can learn as you go( math is not more then 10% of most jobs in the field, very useful and important to understand but nothing that should really hold you back from bringing ML models out into the 🌎
@@datawithsandro2919 Thank you for your response. Actually I always got A in calculus and algebra courses. Learning statistics wont be problem for me. However, I heard organizations want Masters degree from machine learning engineer candidates. I have bachelor degree in computer science and I have no intention to do graduate degree. Does this mean it will be hard for me to get a job as ML engineer without master's degree despite taking some certifications ?
@@yrysf777 Sounds great, honestly here in europe Master or higher is very common, however as mentioned it is changing. I think you should simply go out and try, maybe you won't get the job in the dream company right away but something in either a start up or a related job that has a lot of exposure to ML Engineering should be very achievable. Building from there will be much easier once you have some relevant work experience in the field, and as the field is constantly growing jt should become a lot easier over time. Also if you have a job that sometimes gives you a chance to do some ML projects, will boost your chances a lot for your next job
Took ML course at my uni and failed the exam because it was all about proving formulas by hand (bruh). Why do unis overcomplicate everything :(
you need to know the names of esoteric equations, not necessarily how you use them. Chances are most of the guys in the field have very little real math skills, most of the math happens when you are shitting out academic papers and when trying to sound smart in discussions.
1:01: 📚 Don't be intimidated by math in machine learning, as you can get away with doing the bare minimum.
4:38: 📚 Understanding statistics, probability theory, and linear algebra is crucial for machine learning.
7:25: 🧮 Learning mathematics in machine learning is important, but focus on the basics and have fun building models.
Recap by Tammy AI
LoL scared of math
10th Class Maths Is Enough
Almost 4 minutes without answering the question, I’d recommend that you just go straight to the point or people will lose interest.
I actually did immediately after I read your comment, thanks for the heads up, I’ll look for another video 🙏