Spinning Up a Pong AI with Keras and OpenAI

Posted by whatrocks on Wednesday December 12, 2018 @02:40PM
from the blog.floydhub.com dept.

https://blog.floydhub.com/spinning-up-with-deep-reinforcement-learning

( 2 comments | 9 points | on HN )

Re:Spinning Up a Pong AI with Keras and OpenAI by minimaxir on Wednesday December 12, 2018 @02:51PM (#18667168)

You posted this last week, which already had a good amount of discussion: https://news.ycombinator.com/item?id=18620814

[ Parent | on HN ]

Re:Spinning Up a Pong AI with Keras and OpenAI by twillmas on Wednesday December 12, 2018 @02:51PM (#18667162)

If the in-game hard-coded AI just follows the ball, how do you ever beat it? I’d like to see the logic behind that 1970s opponent paddle AI, especially as it ramps up difficulty.

[ Parent | on HN ]