Dustin play flappybird

• 

  Changing the input

• 

• 

  This is the original game

• 

  A picture of how Dustin train himself

Inspiration

I have two days in Tiger hack, So I decided to start learning Machine Learning over youtube and Google

What it does

Simply play the game Flappy bird

How I built it

I clone a flappy bird game (via pygame) from Github (https://github.com/sourabhv/FlapPyBird) By simplify the game and re-editing the input. I create an AI (Dustin) and lets him learn how to play the game

Challenges I ran into

It was difficult to learn something over youtube in just 2 days. But it work out nicely. I test out Dustin and lets him train himself overnight. He took all of my processing power, and I have to force restart a lot of time

Accomplishments that I'm proud of

After 16 hours of learning, (from 0 points, Dustin make it to the 7 point his highest high score and no longer burning my potatoes computer when he learning. This means that I just learn how ML work, in just 2 days, that is an amazing achievement already

What I learned

that ML is not that hard and it does not require a powerful machine to run. Sometimes, without appropriate equipment, It is necessary to improvise and improve algorithm to make my program run smoothly

What's next for Dustin play flappybird

ML seems is challenging and I still have a lot of things to do. I wonder if I can help Dustin train himself to recognize fake new over the internet. That would be a great project.

Built With

• python