Tech & Science Google's AlphaGo AI can teach itself to master games like chess

16:36  07 december  2017
16:36  07 december  2017 Source:   Engadget

Google offers real-time search data for news, images and video

  Google offers real-time search data for news, images and video Finding out what people are searching for in real time is a trivial matter these days, thanks to Google Trends. Online since 2015, the service is useful for discovering what the world is paying attention to, like why Donald Trump was elected last November or which Halloween costume will be the most played out. Heck, Mac users can even get a screensaver to visualize searches as they happen. Now Google is exposing even more of its real-time data to include news, images, video and even shopping results in your topical search area.The example Google gives is a search for Taylor Swift.

Google ' s DeepMind team has already advanced its AlphaGo AI to dominate Go without human input, but now the system is clever enough to master other Chess and Shogi were relatively easy tests, as they're simpler than Go. It'll be another thing entirely to tackle complex video games like StarCraft II

Although a computer beat the world chess master two decades ago, experts thought that it would be at least This new self- taught AI absolutely decimated last year' s algorithmic champion AlphaGo 100 games to 0. Rather than studying expert human moves, AlphaGo Zero only games against itself .

Lee Sedol standing on a stage holding a microphone © Provided by Engadget Google's DeepMind team has already advanced its AlphaGo AI to dominate Go without human input, but now the system is clever enough to master other board games without intervention. Researchers have developed a more generalized system for AlphaGo Zero that can train itself to achieve "superhuman" skill in chess, Shogi (a Japanese classic) and other game types knowing only the rules, all within less than a day. It doesn't need example games or other references.

This doesn't mean that DeepMind has developed a truly general purpose, independent AI... yet. Chess and Shogi were relatively easy tests, as they're simpler than Go. It'll be another thing entirely to tackle complex video games like StarCraft II, let alone fuzzier concepts like walking or abstract thought. There's also the question of speed: less than 24 hours works for board games, but that's too slow for situations where AI needs to adapt on the spot.

Even so, this is a major step toward AI that can accomplish any task with only minimal instructions. Robots and self-driving cars in particular may need to learn how to navigate unfamiliar environments without the luxury of pre-supplied training material. If nothing else, chess champions have one more reason to be nervous.

ArXiv.org

Google won't show news from sites that hide their country of origin .
Google's ongoing quest to curb fake news now includes sites that are less than honest about their home turf. A spokeswoman speaking to Bloomberg explained it as a matter of adaptation. Google has to update its policies to "reflect a constantly changing web," she said, and that means ensuring that people can "understand and see where their news online is coming from.

—   Share news in the SOC. Networks

Topical videos:

This is interesting!