AlphaGo Shows It Is Possible For Machine Learning to Beat Humans at Go


The game of go, invented in China more than 2,500 years ago, may be the world’s oldest and most complex board game. It’s a strategic game of infinite complexity that relies on more than just tactical skill. The sheer beauty of the game’s simple rules is a wonder to behold. Out of its simplicity, the game produces incredibly complex battles between Black and White armies that span the entire 19 by 19 grid.

The emergence of AlphaGo shows that even in the most profoundly difficult disciplines, it is possible for machine learning to surpass humans. This remarkable accomplishment has inspired a new generation of scientists to work on this remarkable art form, in the hope that they can use this technology to solve some of humanity’s most pressing problems.

Scientists working on artificial intelligence (AI) are finding ways to apply this advanced computer technology to a broad range of applications, including medicine, law enforcement and national security. One area that holds great promise is the field of decision-making, where AI can help in such vital areas as combatting climate change, detecting cancer and assisting with medical treatment.

For more than 30 years, NOAA’s Geostationary Operational Environmental Satellites (GOES) have been a vital tool for weather forecasting. The GOES-R Series, currently consisting of GOES-16 and GOES-17, is providing sharper and more defined images to help NOAA forecasters better see severe storms, hurricanes, wildfires and other natural hazards in the Western Hemisphere. The GOES-R series also carries the Advanced Baseline Imager (ABI) for multispectral imaging and the Geostationary Lightning Mapper (GLM), which provides real-time monitoring of lightning activity within severe storms across the United States.

GOES is unique among weather satellites in that it operates in geosynchronous orbit, which means it maintains a constant position above Earth’s equator. The GOES-R Series satellites, with their suite of sophisticated instruments, provide atmospheric conditions over the western hemisphere with greater clarity and frequency than any other NOAA satellites.

A key component of a successful satellite is the ground system that supports it, and NOAA’s Satellite Operations Center in Suitland, Maryland, manages GOES and its data. The GOES program is a collaboration between NOAA and NASA: the National Aeronautics and Space Administration builds and launches the satellites, and NOAA manages them and distributes their data.

The GOES-R series was designed to operate for three years; however, each satellite is equipped with sufficient fuel to continue operating beyond that time if needed. The GOES-R program continues to improve operational services and support atmospheric science research, numerical weather prediction models, and environmental sensor design and development.