This “trick” of you is really useful but a lot of college teachers don’t do that, after i started doing that more and thinking more about everything math and algorithms became so much simpler
Awesome content, I really appreciate all your videos, helped me a lot because I just was asking myself what I need to learn on math to get really good in ML, I currently search for clinical datsets and try to solve problems and create models with sklearn. Thanks and keep up the good work!
I like your way of explaining formula your one example made me search your channel for all over the youtube Please please start a series explaining all ml formula with statistics 🙏 ❤ Hope you reach more and untill its not gonna struggle searching you
5:50 That's a great video. I'm really curious as to how understanding various math concepts are used like random variables, probability distributions, matrix rank,etc. Would love to see a follow-up video on the math and use cases
Well I’m glad I researched this field before i started my undergrad in data analytics and will do my MSc in data science. Longer route and math used to give me anxiety but it’s a must have, also a competitive edge against my boot camp competitors.
Hey infinite, I am currently in my last year of high school and am very stressed on what I should do next year. I have 4 bachelor's in my mind each with its own benefit. I want to be person with diverse skillset and be able to develop software but also be a AI/ML engineer. - CS: overall great covers all things well, but I feel like it doesn't have enough math for Ai/ML (and I also feel like I am going to lose my somewhat talent for math) - CE as more math and overall gives flexibility, but this lacks some raw theory which CS has more of. - Math: I mean math is math but also doesn't contain enough CS. - Physics and astronomy(most unlikely): math in a different form fr + space, quantam computing is cool aswell I am sure that I am going to Ugent next year (belgium), but js not sure which of these 4 to choose.
Point you are trying to make via Linear regression example at 5:08 which I first look Linear regression equation its look like exactly the same as straight line equation from geometry class which I learn in high school. And rest all other equations that are part of this topic of linear regression was going over my head. So, I sat on night with whatever silly doubt coming to my mind I give it to chat GPT and validating my understanding. After 3hrs of chatting and lot of digging with Chat GPT I get each and every part of this equation cleared. Like why I was comparing this model equation with geometry taking me to wrong direction, what is sense of doing square at RSS step, etc. Even I saved my whole chat with chat GPT in word document..... I am still in the process of learning but one thing which I totally agree is that strong "Mathematical Intuition" will help us more in understanding the things better. Just shared my little experience.... Thanks for this video.
You mentioned only z-score. Depending on actual distribution and the way unknown parameters are linked to standard deviation, you might need t-distribution or chi-squared.
basics and fundamentals are king in almost any domain, and even here in ML, math/bias-variance is still king despite the insanely fast evolution of AI.
And ya before realizing that I need to develop some foundation in probability theory concept and shifting my self to book "First course in Probability by Sheldon Ross". I saw few starting lectures of Statistical Learning playlist on Stanford learning channel. OMG author of "Elements of statistical learning " teaching themselves .... One thing I tell you guys those guys know their stuff .... let me say it again they know what they are talking about.... I highly recommend if you know little bit about linear regression just saw first few lecture... You got amaze..
I have heard dsa is also important for data science what do you think?If you think it is can you provide a few resources because i really liked your explanation and resources because it really saved me from burnout...
Thanks for the comment! I'm assuming you mean data structures and algorithms? While I think its an important CS concept and definitely cant hurt to know, if you are close to burnout I think you can skip it for now :) Knowing how to work with numpy arrays and pandas dataframes will be more important to you on a practical level. I still think it's good to know once you are further from a burnout because it will give you a good intuition for efficiency in Computer science. I am actually working on a script for a video that should help you a lot, won't be the next video but the one after that (on how to learn data science efficiently and not waste time)
There are three stages of truly understanding the math behind machine learning: 1. The Unfamiliar Stage At first, the math can feel cryptic and disconnected. It’s hard to find the intuitive meaning, but persistence is key- don’t give up. Keep pushing through! 2. The Familiar Stage After some struggles, things start to click. You begin to grasp what each formula represents because you’ve caught onto the ‘vibe’ of machine learning. 3. The Mastery Stage Now, you don’t just understand the intuition, you play with the formulas. You have an instinct for them, a sense that lets you not only understand but also anticipate new insights. True understanding means not just knowing but sensing it deeply. Give time for your neuroplastisity.
This is all very basic stuff any student could learn within half a year and I have a really hard time believing this isn't taught at bootcamps, especially since it is where I am from
I believe there is a difference between being familiar with the concepts and truly understanding them intuitively and being able to apply them in practice. For instance, any high school kid knows the chain rule of calculus, but I don't think they will be able to backpropagate though any layer in a NN by hand without a good deal of exposure and practice first.The concepts might seem simple enough, but leveraging them in the real world is usually not straight forward.
This “trick” of you is really useful but a lot of college teachers don’t do that, after i started doing that more and thinking more about everything math and algorithms became so much simpler
desculpa, eu nao pude entender seu comentario em inglês. acho que vou ter que ver o video inteiro
Awesome content, I really appreciate all your videos, helped me a lot because I just was asking myself what I need to learn on math to get really good in ML, I currently search for clinical datsets and try to solve problems and create models with sklearn. Thanks and keep up the good work!
Glad to help! Thanks for the nice comment and support!
statquest is also very good for understanding specific concepts
Double bam
BAAAAM
triple bamm
Underrated Fr 🙌...... such quality content needs more attention.
I like your way of explaining formula your one example made me search your channel for all over the youtube
Please please start a series explaining all ml formula with statistics 🙏 ❤
Hope you reach more and untill its not gonna struggle searching you
Your channel is amazing this exactly what i needed. Thks man 😃
5:50 That's a great video. I'm really curious as to how understanding various math concepts are used like random variables, probability distributions, matrix rank,etc. Would love to see a follow-up video on the math and use cases
Well I’m glad I researched this field before i started my undergrad in data analytics and will do my MSc in data science. Longer route and math used to give me anxiety but it’s a must have, also a competitive edge against my boot camp competitors.
Super sir
This channel is just going to grow like crazy
This was very insightful! Thank you.
very helpful video 🙂💕
Hey infinite, I am currently in my last year of high school and am very stressed on what I should do next year. I have 4 bachelor's in my mind each with its own benefit. I want to be person with diverse skillset and be able to develop software but also be a AI/ML engineer.
- CS: overall great covers all things well, but I feel like it doesn't have enough math for Ai/ML (and I also feel like I am going to lose my somewhat talent for math)
- CE as more math and overall gives flexibility, but this lacks some raw theory which CS has more of.
- Math: I mean math is math but also doesn't contain enough CS.
- Physics and astronomy(most unlikely): math in a different form fr + space, quantam computing is cool aswell
I am sure that I am going to Ugent next year (belgium), but js not sure which of these 4 to choose.
Ritvikmath is also a good channel for explaining statistic and ml concepts.
Would very much like you to cover, bias, varience
Thank you❤ keep uploading video like this, easy explanation
Point you are trying to make via Linear regression example at 5:08 which I first look Linear regression equation its look like exactly the same as straight line equation from geometry class which I learn in high school. And rest all other equations that are part of this topic of linear regression was going over my head. So, I sat on night with whatever silly doubt coming to my mind I give it to chat GPT and validating my understanding. After 3hrs of chatting and lot of digging with Chat GPT I get each and every part of this equation cleared. Like why I was comparing this model equation with geometry taking me to wrong direction, what is sense of doing square at RSS step, etc.
Even I saved my whole chat with chat GPT in word document.....
I am still in the process of learning but one thing which I totally agree is that strong "Mathematical Intuition" will help us more in understanding the things better.
Just shared my little experience.... Thanks for this video.
You explain the main goal of minimising rss. Could you please make videos on other equations from statistics and linear algebra
Please make the video on Bias and Variance and the biase variance tradeoff
You mentioned only z-score. Depending on actual distribution and the way unknown parameters are linked to standard deviation, you might need t-distribution or chi-squared.
Please make videos on those topics.
Sehr gut
basics and fundamentals are king in almost any domain, and even here in ML, math/bias-variance is still king despite the insanely fast evolution of AI.
I would love a video about bias, variance, an the bias variance trade off
3Blue1Brown is gem ❤
I don't see any point in memorizing ML algorithms.
Great video
Such a good video
3:41 so basically be a math major. Or more specifically, approach formulas like a math major or mathematician
the math in compsci also tries to achieve this type of understanding, at least in my university (germany)
Hey, I found your videos helpfull can you make video on multidimensional arrays?
And ya before realizing that I need to develop some foundation in probability theory concept and shifting my self to book "First course in Probability by Sheldon Ross". I saw few starting lectures of Statistical Learning playlist on Stanford learning channel. OMG author of "Elements of statistical learning " teaching themselves .... One thing I tell you guys those guys know their stuff .... let me say it again they know what they are talking about.... I highly recommend if you know little bit about linear regression just saw first few lecture... You got amaze..
Bootcamps will loss customers if they say that you will need some math.
I have heard dsa is also important for data science what do you think?If you think it is can you provide a few resources because i really liked your explanation and resources because it really saved me from burnout...
Thanks for the comment! I'm assuming you mean data structures and algorithms? While I think its an important CS concept and definitely cant hurt to know, if you are close to burnout I think you can skip it for now :) Knowing how to work with numpy arrays and pandas dataframes will be more important to you on a practical level. I still think it's good to know once you are further from a burnout because it will give you a good intuition for efficiency in Computer science. I am actually working on a script for a video that should help you a lot, won't be the next video but the one after that (on how to learn data science efficiently and not waste time)
@@InfiniteCodes_ thanks i watched your recent video too before this comment and will watch that one too thanks for guidance :)
...I'm going to stick with web dev 😂
🤣
Good for you 😂
I am thinking this too 😢😂, will also try this too
Man you made me laugh after many days of depression 😂😂
There are three stages of truly understanding the math behind machine learning:
1. The Unfamiliar Stage
At first, the math can feel cryptic and disconnected. It’s hard to find the intuitive meaning, but persistence is key- don’t give up. Keep pushing through!
2. The Familiar Stage
After some struggles, things start to click. You begin to grasp what each formula represents because you’ve caught onto the ‘vibe’ of machine learning.
3. The Mastery Stage
Now, you don’t just understand the intuition, you play with the formulas. You have an instinct for them, a sense that lets you not only understand but also anticipate new insights. True understanding means not just knowing but sensing it deeply.
Give time for your neuroplastisity.
I made a similar video on my channel. But this one’s better! Hopefully we can collaborate someday. 😅
Most of the people in that field are bad at maths and not bright. Frameworks are what made them Data Scientists / ML Engineers etc.
This is all very basic stuff any student could learn within half a year and I have a really hard time believing this isn't taught at bootcamps, especially since it is where I am from
I believe there is a difference between being familiar with the concepts and truly understanding them intuitively and being able to apply them in practice. For instance, any high school kid knows the chain rule of calculus, but I don't think they will be able to backpropagate though any layer in a NN by hand without a good deal of exposure and practice first.The concepts might seem simple enough, but leveraging them in the real world is usually not straight forward.
Are you German?