Function Learning Neural Net

Inspiration

I was inspired by those cool evolution ai videos on youtube

What it does

It trains on data from a function ( in this case sine) and it uses an evolutionary algorithm to slowly converge on the current patterns by mutating over generations and only keeping the best solution so far.

How we built it

I built it using c++ and its standard library functions (also used claude to figure out how to get the display working on the exe and some bugs)

Challenges we ran into

Its a struggle to work on something abstract like this as I would have to think through it alot and test cases on paper to see what was wrong.

Accomplishments that we're proud of

I am very proud of it working and I coded for so long in one sitting.

What we learned

I learns a lot more about matrix multiplication and neural nets in general through all my failures implementing it.

What's next for Function Learning Neural Net

Adding the ability to choose the function being tested on, better mutation methods.

Built With

• c++