DeepMind’s Artificial Intelligence Can Beat All StarCraft II Professional Gamers

Nowadays, computers are more and more sophisticated, and artificial intelligence systems are increasingly more complex so that they already do things like us, humans. In a recent demonstration, DeepMind’s artificial intelligence, AlphaStar AI, a deep learning machine that plays video games, beat all StarCraft II professional gamers who participated in the tests.

During ten head-to-head matches, AlphaStar AI beat both Dario “TLO” Wunsch and Greegorz “MaNa” Komincz, the best StarCraft II gamers in the world. Neither Wunsch nor Komincz could match DeepMind’s artificial intelligence which won every single battle.

The tests were part of the DeepMind’s research in the artificial intelligence industry, and Blizzard contributed as much as possible by giving StarCraft II to the computer researchers. StarCraft II is the best game for the specialists to research AI systems since it is a sophisticated-enough video game that helps researchers test and educate their artificial intelligence machines.

DeepMind’s Artificial Intelligence Can Beat All StarCraft II Professional Gamers

“We take many replays from pro and players, and we try to get AlphaStar to understand by looking at a situation that human player is in. And then we try to get it to imitate those moves,” explained DeepMind research co-lead Oriol Vinyals.

Like humans, DeepMind’s AlphaStar AI evolves by studying and practicing. The researchers make the artificial intelligence system watch and observe StarCraft II gamers while they play. The deep learning machine learns from what it sees and builds its own strategies in the game.

“Overall, AlphaStar uses considerably fewer APM (actions-per-minute) than a human pro. That indicates that it’s winning by not clicking insanely but by doing something much smarter than that,” said David Silver from DeepMind. “We measured how quickly it reacts to things. If you measure the time between when AlphaStar perceives the game. From when it observes what’s going on, then has to process it, and then communicate what it chooses back to the game. That time is actually closer to 350ms. That’s on the slow side of human players,”

You May Also Like

Leave a Reply

Your email address will not be published. Required fields are marked *