Google’s DeepMind AI Research News & Update: Company Works With Blizzard To Use ‘Starcraft II’ As Study Environment! Can Artificial Intelligence Win?
Google has partnered with Blizzard to use its game "Starcraft II" for its Google's DeepMind Artificial Intelligence (AI) improvement. The team up was announced at the recent concluded BlizzCon 2016 in Anaheim, California.
DeepMind Technologies Limited is recently acquired by British artificial intelligence company that is also obtained by Google in 2014. The company has been doing its research and developing Google's DeepMind AI, which targets to "solve complex problems without needing to be taught how."
Blizzard announced that it will be releasing tools that Google's DeepMind AI team needed to teach its AI on how to play "Starcraft II." The team will be using its DeepMind platform to create an environment and simulate it with the game.
According to Value Walk, Google's DeepMind AI team is optimistic that the Blizzard's "Starcraft II" will be able to help its research about the real world. The gameplay that "Starcraft II" offers the world of "messiness" in the game that the team can simulate in a useful environment for their research.
The "Starcraft" game series has been played widely for almost 20 years and was recognized as 1v1 best video game of all time. The game was also a pioneer in eSports, showcasing battle championships played by top players around the globe. This status has made Google's DeepMind AI team to choose "Starcraft II" over other video games available to help its research into the artificial intelligence.
Google's DeepMind AI team sees an opportunity to "Starcraft" since the nature of the game has been a real-time strategy, whereas a single player needs to decide and use their resources effectively. This scenario has the same effect on the real-world that will help AI research compared to other games with massive online battle arenas (MOBAs).
Meanwhile, Google's DeepMind AI team has also partnered with the researchers at the University of California to focus on awareness of physical laws. The idea is to make AI machines evaluate object's properties even without the knowledge of physical laws. This only confirms that despite the AI machine was successful in superhuman performance, like resolving complex problems, it still does not have a human intuition even like a child's.