analysis- looking at the future of artificial intelligence from alphago- where is the next emerald of ai-（分析-从alphago看未来的人工智能,是下一个翡翠的ai -）.docVIP

Analysis: Looking at the Future of Artificial Intelligence from AlphaGo: Where is the next emerald of AI? In March 2016, AlphaGo turned out to play against the chess player Ji Shishi, which has a great impact on the social intelligence of artificial intelligence. What did AlphaGo bring to us? In the AlphaGo search, the Google DeepMind team introduced a new concept - the combination of deep learning and enhanced learning to determine the two tasks, that is, to determine the current board is good or bad, and to predict the future benefit To the direction. This can be seen AlphaGo algorithm and the future business model of the association, that is, through the analysis of large data, let us on the ‘current state’ has a clear understanding. This state can be a board, can be the status of the two teams confrontation in football, it can be a state of the current marketing. Go one step, can be understood as the predecessor of the future to predict, in business activities, marketing activities in the next step. The two tasks of business behavior in the judgment of reality and the prediction of the future of business are equally important and require the support of large data. Go is a closed game, there is no external factors of interference, in order to get more data, AlphaGo introduced a self-game. The so-called self-game is to play their own game, you will continue to get feedback, and then to update the strategy. After countless times of this game, the final will get a good strategy, the final output is a behavior strategy. So AlphaGo tells us that in a closed scene, we can use self-game simulation method to get more data. From AlphaGo to Artificial Intelligence If you go along the next go, you will be faced with some questions: What is the goal of the artificial intelligence algorithm? Is there any data? Where is the data? Is the boundary of the problem clear? What is a reasonable way to go? What is a foul Where is the feature? How do I get