DP-Reinforcement Learning
The Brief
Reinforcement learning is how Deepmind learns. Basically trial and error! We wanted to create a game that children would migrate to, to bring them to the stand. This was all about fun and engagement!
The Solution
Build a fun game that perfectly demonstrates the idea of reinforcement learning. We wanted to create an authetic old-school arcade experience so we decided to use an original Atari 2600 controller!
The Outcome
"A homage back to the Atari 2600 8-bit titles, an interactive display like nothing else! To showcase how Deepmind uses Reinforcement Learning for AI we created this game, with our friends at Gramfilm, to create an interactive wall at the Cheltenham Science Fair. The idea is you play the game without instructions, then play again once you have learnt what’s good and what isn’t. Trial and error!"
Overview
Great design, fun, addictive game play and a retro Atari controller came together to create a unique experience for people of all ages.
We just had to make sure we time limited the game to ensure everyone got a go!
Team Thoughts
