Top Ten Future Trends of Artificial Intelligence

With the maturity of deep learning technology, AI artificial intelligence is gradually becoming popular from cutting-edge technology. The game between AlphaGo and humans is not a video game as we have understood it in the past. The level of video games will never improve, but AlphaGo has the most critical "deep learning" function of artificial intelligence. There are two deep neural networks in AlphaGo, Value Networks and Policy Networks. Among them, Value Networks evaluates the chessboard point location, and Policy Networks selects the placement. These neural network models are trained through a new method that combines chess records learned in human expert games and reinforcement learning in playing chess with oneself (Self-Play). In other words, the existence of artificial intelligence can make AlphaGo's Go level continue to improve during learning.

The technical application of artificial intelligence is mainly in the following aspects:

Natural language processing (including speech and semantic recognition, automatic translation), computer vision (image recognition), knowledge representation , automated reasoning (including planning and decision-making), machine learning, and robotics. According to the technical category, it can be divided into two types: sensory input and learning and training. The computer obtains audio and video perceptual input through speech recognition, image recognition, reading knowledge base, human-computer interaction, physical sensing, etc., and then learns from big data to obtain a brain with decision-making and creative capabilities.

From the PC era in the 1980s and 1990s to the Internet era, what we have brought to us is the explosion of information and the decentralization of information carriers. After the network information acquisition channel shifted from PC to mobile, the Internet of Everything became a trend. However, technical limitations made it difficult for the mobile Internet to spawn more new applications and business models. Today, artificial intelligence has become the most exciting and anticipated technology of this era, and will become the focus of the development of the IT industry in the next 10 years or even longer.

The concept of artificial intelligence has actually been very popular in the 1980s, but the technical limitations of both software and hardware made it obsessed for a long time. Now, the development of the four major catalysts of large-scale parallel computing, big data, deep learning algorithms and human brain chips, as well as the reduction of computing costs, have made artificial intelligence technology advance by leaps and bounds.