A model of artificial intelligence from the Google DeepMind can play different open world games like No Man’s Sky. In such a way that it acts like a human just watching videos from the screen. Thus, it shows what could be a step towards generally intelligent artificial intelligences operating in the corporeal world.
Playing video games has long been a means of testing the progress of AI systems. For example, how does AI’s mastery of virtual chess and Go Google DeepMind. But these games have obvious ways to win or lose, making it relatively easy to train an AI to succeed in them.
Now, researchers from Google DeepMind developed an AI they call Scalable Instructable Multiworld Agentor SIMAwhich can play nine different video games and virtual environments that you haven’t seen before using just the game’s video feed.
“This is actually the interface that humans use to interact with a computer, it’s a very generic interface”, says Frederic Besse of DeepMind. When asked in natural language, SIMA can perform about 600 tasks, lasting 10 seconds or less, that are common across different games, such as moving, using objects, and navigating menus.
Besse and his colleagues used pre-existing video and image recognition models to interpret game video data, then trained SIMA to map what happens in the video for certain tasks.
Source: Atrevida
Earl Johnson is a music writer at Gossipify, known for his in-depth analysis and unique perspective on the industry. A graduate of USC with a degree in Music, he brings years of experience and passion to his writing. He covers the latest releases and trends, always on the lookout for the next big thing in music.
