Google DeepMind appeared before users with a new artificial intelligence tool. This artificial intelligence tool, called SIMA, was trained to play games with and like humans.
Games introduce characters called bots, controlled by artificial intelligence, to users. But these characters are quite boring, ordinary or too strong. SIMA, the new artificial intelligence developed by DeepMind, aims to solve this problem by playing games like humans.
SIMA will be Able to Play All Kinds of Games
Although this artificial intelligence, called SIMA (Scalable, Instructable, Multiworld Agent), is currently in the testing phase, it seems that it may appear frequently in the future. It is stated that SIMA can even learn to play games that have a clear linear gameplay and no ending, such as massive role-playing games or open world games. SIMA will have crossed an important threshold in this respect.
The aim of SIMA is not to replace the artificial intelligence that exists in games. Instead, it will act as another player who can join the game, understand game commands by speaking, and recognize three-dimensional worlds and objects around it. “SIMA is not trained to win the game, but to run the game and do what it is told,” said Tim Harley, DeepMind Researcher and SIMA Leader. He made a statement.
SIMA Tested on Many Games
According to the statement made by Google, it worked with 8 game producers, including Hello Games, Embracer, etc. SIMA artificial intelligence has tried games such as No Man's Sky, Valheim, Goat Simulator 3. Additionally, SIMA does not require a special API or source code to play the games.
