A Game Developer Learns Machine Learning

Hey Guys,

Im not sure if there is an ML sub-forum here or not yet, if not there should be

I just thought I would share my journey into ML using Unity ML agents incase anyone is interested in following along.

https://mikecann.co.uk/machine-learning/a-game-developer-learns-machine-learning-intent/

TLDR; I indent to write a series of posts detailing my learning process with the eventual goal of writing my own agent that can play my game Mr Nibbles Forever better than any human player

Do please let me know if you have any comments or thoughts below!

 

Okay so first up, I build and run the balance balls example, plus add a few tweaks of my own.

https://mikecann.co.uk/machine-learning/a-game-developer-learns-machine-learning-getting-started/

 

Do I need a high end PC to use tensor flow? I've a MacBook pro, will it work?

 

That should work just fine. TBH I havent seen my GPU actually being used, tho that could be windows incorrectly reporting. I need to investigate this.

 

I finally found time over the weekend to do some more ML work.

https://mikecann.co.uk/machine-learning/a-game-developer-learns-machine-learning-a-little-deeper/

This time I dig a little deeper and start from a blank slate and train up a simple model from there.

 

Spent a little more time on this over the past week or so and wrote up the results.

https://mikecann.co.uk/machine-learning/a-game-developer-learns-machine-learning-mr-nibbles-basics/

Not massively successful but very interesting nonetheless