A game of Go. Photo: Baona / iStock

DeepMind's latest iteration of AlphaGo — the artificial intelligence that beat world champion Go player Lee Sedol in 2016 — can learn to play the ancient game without feedback from humans or data on their past plays, researchers report today in Nature. Instead, the new AlphaGo Zero started with just knowledge of the rules and learned from the success of a million random moves it made against itself.

The score: After three days of training, the AI beat the original AlphaGo 100 to 0 — and was also able to create new moves in the process. This demonstrates a decades-old idea called reinforcement learning, suggesting that "AIs based on reinforcement learning can perform much better than those that rely on human expertise," writes computer scientist Satinder Singh in his accompanying article.

What it means: If AI can utilize reinforcement learning, that could be important in cases where large amounts of human expertise isn't available. But, it isn't clear how much this strategy will generalize to other applications and problems, says the University of Washington's Pedro Domingos. Go, though more complex than chess, offers a problem with defined rules unlike a busy street with unpredictable pedestrians and ambiguous shadows that a robot-controlled car might operate in.

What's new: AlphaGo's initial iteration was trained on a database of human Go games whereas the newer AlphaGo Zero's artificial neural networks use the current state of the game as input. Through trial and error and feedback in the form of winning, the AI learned how to play.

It then used that same network to choose its next move whereas AlphaGo used a separate network. This reinforcement learning strategy, which was used extensively by AlphaGo as well, has its roots in psychology: the neural network learns from rewards like humans do.

The DeepMind researchers wrote: "the self-learned player performed much better overall, defeating the human-trained player within the first 24h of training. This suggests that AlphaGo Zero may be learning a strategy that is qualitatively different to human play."

How they did it: AlphaGo Zero uses less computing power than earlier versions but Google's immense computing power was still key. The sheer number of games the AI can play against itself is an advance, says Domingos, who is the author of a book called The Master Algorithm.

He points out though that the roughly 5 million training games of self-play it took for AlphaGo Zero to beat AlphaGo is "vastly more" than the number of games Sedol had played to become a champion.

Recent work suggests simpler forms of learning could achieve similar goals. A paper published earlier this year by OpenAI showed how a technique similar to hill-climbing — in which the AI basically starts with a solution then makes small tweaks to optimize it — can solve Atari games, albeit simpler than Go.

Go deeper

Trump commutes Roger Stone's sentence

Roger Stone arriving at his sentencing hearing on Feb. 20. Photo: Drew Angerer/Getty Images

President Trump on Friday evening commuted the sentence of his longtime associate Roger Stone, according to two senior administration officials. Stone in February was sentenced to 40 months in prison for crimes including obstruction, witness tampering and making false statements to Congress.

Why it matters: The controversial move brings an abrupt end to the possibility of Stone spending time behind bars. He had been scheduled to report to prison on July 14.

Updated 7 hours ago - Health

Which states have set single-day coronavirus records this week

Data: COVID Tracking Project and state health department data compiled by Axios; Map: Danielle Alberti and Naema Ahmed/Axios

13 states this week surpassed records set just last week for their highest number of coronavirus infections in a single day, according to the COVID Tracking Project and state health department data. 16 states in total reported new highs.

The big picture: The United States' alarming rise in coronavirus cases isn't just due to increased testing — particularly where the number of cases has grown fastest over the last month, Axios' Andrew Witherspoon and Caitlin Owens report.

Updated 8 hours ago - Politics & Policy

Coronavirus dashboard

Illustration: Sarah Grillo/Axios

  1. Global: Total confirmed cases as of 6:30 p.m. ET: 12,389,660 — Total deaths: 557,279 — Total recoveries — 6,830,596Map.
  2. U.S.: Total confirmed cases as of 6:30 p.m. ET: 3,169,611 — Total deaths: 133,885 — Total recoveries: 983,185 — Total tested: 38,856,341Map.
  3. Public health: The reality of the coronavirus bites.
  4. Trade: Trump says he's no longer considering phase-two trade deal with China because the pandemic damaged the two countries' relationship.
  5. 🎧 Podcast: Rural America has its own coronavirus problem.