AI learns to play multiplayer game; beats humans at it

  • With the stupendous development seen in silicon technology as well as rapid advancements in the field of machine learning, capabilities of Artificial Intelligence (AI) based systems have grown multi fold over the years..
  • After besting humans in complex board games such as Chess and Go, DeepMind's AI has now emerged as a competitor for humans in a version of Quake II, a virtual battle field game.
  • As opposed to rather simpler games such as Go and Chess, this particular version of Quake III Arena is a multiplayer game like PUBG and thus requires human skills such as cooperation.
  • The game revolves around the idea of capturing the enemy's flag while protecting your own flag from them. As such,what's being demanded of the AI to win the game here is more than what's been asked of algor in the past.
  • As for experts, they claim this is more than just an AI winning against humans in a game. The expectation is that AI can now be taught to learn from this and then adapted to improve automated systems in the real world.
  • For example, the learnings from this can be used to help self-driving cars in the future better interact with each other and humans.

Don't forget to share this pointer!

View more comments +