1. AI: AlphaGo European champion : Fan Hui A feat previously thought to be at least a decade away!!!

2. 1997, AI named “Deep Blue” beat chess world champion.

3. Mastering the game of Go with deep neural networks and tree search In Nature, Januray, 2016 Google DeepMind

4. Key Points Given current game state, – how to decide next action “a”? – How to evaluate each legal action?

5. Monte Carlo Tree Search (MCTS) A popular heuristic search algorithm for game play – By lots of simulations and select the most visited action. Evaluation Node value: Wining rate Key point: how to calculate this value One playout to simulate the game

6. Monte Carlo Tree Search (MCTS) AlphaGo – Key point: how to calculate the value for each action (path)

7. The visited number for this edge The mean value of all simulations passing through it. Winning rateWinning rate predicted by fast rollout playoutWinning rate predicted by deep reinforcement learning Action probability predicted by supervised and reinforcement learning

8. 30 million games 30 million self-play games

11. Now AlphaGo will face its ultimate challenge: a 5-game challenge match in Seoul against the legendary Lee Sedol, the top Go player in the world over the past decade. Lee Sedol March 9 th, 10 th, 12 th, 13 th, 15 th