Mathematics fundamentals for ML AI (for dummies like me)

[kevinUsa]

17 hours ago, Spartan said:

local community colleges lonmath classes offer chestaru they are also good

i did one for statistics last year 

Fee enta untadi colleges kipart time aitay ??

For International students and locals..

Or college name cheppu

[yslokesh]

18 hours ago, csrcsr said:

Cool bro asalu subjrcts enduku petaro ipudu ardam avutudni and beauty is phython packages are there we should know when to use what , was amazed maths influence on all programming data structures ,modern day functional programming paradigms 

Haaa...Tenth class lo SET theory lo: What is a function? 

Ani adigithe, andharoo chebutaaru: "No two ordered pairs will have the same first co-ordinate" ani..

kaani computing theory lo idhi yekkada vaadataaru ante, yevvariki theliyadhu..

.

This is the same concept from which Functional Programming Languages are built. ---> For it to be a function, for the same input,  this should always return the same output ...

.

This is so important in memory & Run-time optimizations -- Asalu mind block aipothundhi aalochisthe..

def f(x : Int) = { return  x*0.0002*4.567 / (3.45 * 6.78 * 0.9898) }

That f(..) will be cached and when the same input comes, it will not repeat those heavy FLOATING Point calculations.

This optimization happens at numerator and denominator levels ..

for ex:  The denominator (3.45 * 6.78 * 0.9898 ) is automatically converted to a function and stored for later purposes...

 

This is the basic theory behind Super Computing...

 

[kevinUsa]

8 minutes ago, yslokesh said:

Haaa...Tenth class lo SET theory lo: What is a function? 

Ani adigithe, andharoo chebutaaru: "No two ordered pairs will have the same first co-ordinate" ani..

kaani computing theory lo idhi yekkada vaadataaru ante, yevvariki theliyadhu..

.

This is the same concept from which Functional Programming Languages are built. ---> For it to be a function, for the same input,  this should always return the same output ...

.

This is so important in memory & Run-time optimizations -- Asalu mind block aipothundhi aalochisthe..

def f(x : Int) = { return  x*0.0002*4.567 / (3.45 * 6.78 * 0.9898) }

That f(..) will be cached and when the same input comes, it will not repeat those heavy FLOATING Point calculations.

This optimization happens at numerator and denominator levels ..

for ex:  The denominator (3.45 * 6.78 * 0.9898 ) is automatically converted to a function and stored for later purposes...

 

This is the basic theory behind Super Computing...

 

Marchi poyina bro asalu ee functions mottam

[a4apple]

completed just now.

[socrates]

best explanation 

 

[socrates]

On 2/1/2025 at 5:45 PM, yslokesh said:

Haaa...Tenth class lo SET theory lo: What is a function? 

Ani adigithe, andharoo chebutaaru: "No two ordered pairs will have the same first co-ordinate" ani..

kaani computing theory lo idhi yekkada vaadataaru ante, yevvariki theliyadhu..

.

This is the same concept from which Functional Programming Languages are built. ---> For it to be a function, for the same input,  this should always return the same output ...

.

This is so important in memory & Run-time optimizations -- Asalu mind block aipothundhi aalochisthe..

def f(x : Int) = { return  x*0.0002*4.567 / (3.45 * 6.78 * 0.9898) }

That f(..) will be cached and when the same input comes, it will not repeat those heavy FLOATING Point calculations.

This optimization happens at numerator and denominator levels ..

for ex:  The denominator (3.45 * 6.78 * 0.9898 ) is automatically converted to a function and stored for later purposes...

 

This is the basic theory behind Super Computing...

 

not only that, we didn’t understand the power of integrations and differentiations in maths while learning it as subject

but we only understood the real usage (applications) when we were  introduced to computer graphics subject like creating a 3 dimensional space object and moving that object here and there with some speed

[Tngr]

On 2/1/2025 at 4:20 AM, socrates said:

best one but not for beginners as you mentioned already 

 

 

Did you understand the intuition behind the dot product of vectors is the summation of product of their respective axis value. I didn’t understand, please help

[socrates]

2 minutes ago, Tngr said:

Did you understand the intuition behind the dot product of vectors is the summation of product of their respective axis value. I didn’t understand, please help

are you referring the timestamp 3:41 ?

[Tngr]

47 minutes ago, socrates said:

are you referring the timestamp 3:41 ?

My request is regarding the video in 3blue1brown linear algebra chapter 9( video 9) about dot product and duality. I cannot understand the intuition behind why dot product of vectors (a1,b1) and (a2,b2) is a1a2+b1b2

[l0cal]

unna konchem confidence dobbindi lol

[socrates]

25 minutes ago, Tngr said:

My request is regarding the video in 3blue1brown linear algebra chapter 9( video 9) about dot product and duality. I cannot understand the intuition behind why dot product of vectors (a1,b1) and (a2,b2) is a1a2+a2b2

 

37 minutes ago, Tngr said:

Did you understand the intuition behind the dot product of vectors is the summation of product of their respective axis value. I didn’t understand, please help

i didn’t watch that video 

but in general terms  , dot product is always a single number, vectors are linear regression values

so in simple terms dot product is getting the single value from two linear regressions

correct me if i am wrong or completely wrong 

[socrates]

thanks @csrcsr for this thread, i got it where to reboot myself 

i always focused on learning new tech, this thread is reminder like go back to your basics first 

and @Tngr @yslokesh too

[Mr Mirchi]

Dummies ga

[Thokkalee]

On 1/31/2025 at 10:08 PM, csrcsr said:

for people who are looking for some Math refresher who did not understood the concept in school/college/uni like me (just passed for the sake of exam) or people who are not from math back ground or you just want to refresh as hobby where you can teach your kids after a decade found some interessting sources free unapdu chudandi and also share if you find any thing interesting

For people who are comfortable with math just ignore this might be too basic for you

 

Also math is not needed as we are not going to be data scientist and we are not going to develop a new LLM that is correct but to use which algo to use in what scenarios you need some fundamentals (at least i am lacking) no harm in learning as said earlier worst case mee kids ki chepochu when they grow up

 

if you want deeper study this book is good but i am getting lost nenu only linear algebra start chesanu 

https://mml-book.github.io/book/mml-book.pdf

geekforgeek which is usually basics but they covered the topics which you need to be comfortable 

https://www.geeksforgeeks.org/machine-learning-mathematics/

 

if you want to learn deeper https://www.youtube.com/@3blue1brown best channel hands down but it can get advanced and more time

for some fundamentals interestingly not popular channel this guy explained very well

 

Anna, ML nerchukuntunnava?? What is your study plan? Plz share 

[csrcsr]

18 minutes ago, Thokkalee said:

Anna, ML nerchukuntunnava?? What is your study plan? Plz share 

Pedda plan emi ledu anna antha xinema ledu but trying to.learn math 

And reading supercficial models trying ro see if i can train gpt4 with custom data and emina prompta ki response vache laga 

cheyale maa friend valu edo chetsunaru vala company lo trying to understand

edo resding bro now and then no concrete plan but trying to learn bits may will get some picture at least jargon rofl 

But getting scared roju rojuki e generative ai agentic stuff

Math just to undrstand (we are not going to build llm but undrrstand how to use it)

Phython programming with latwst packages

Inka chudadame edina wrapper vadi relvanat ga undadaniki