I thought my class on Thursday was pretty cool and would like to share these with you!

***

Ok, imagine "Tic Tac Toe". If you don't know Tic Tac Toe, go play it now: https://playtictactoe.org/

Ok now, as a human, when looking at the board, it's easy for you to know if X or O that wins. But imagine machines - how would machine learn and be able to decide which player wins.

Imagine that your data only consists of the positions of the X and O at the board.

So how would you design the algorithm?

Would you look at the places where has X and 0?

Figure out the neighbors and relationships?

See if we have continuous neighbor?

If it's not continuous, take a break and try another path?

That makes sense. Since that's actually how our brain processes it when we look at the board.

But what's cool is the Sequential Minimal Optimization (SMO) algorithm - a linear-based algorithm.

I'm gonna give you a second before we dive into the answers. Take a look at what it is.

***

Done?

Instead of looking at it the way that we do, it counts the number of X. If X is more, X wins. Makes sense? Except in the case where there's a tie, this would definitely be true.

Isn't it cool? For machines to be able to learn and understand something from a totally different perspective?

I thought it was cool and have been excited to share these with you!

## Comments

## Post a Comment