Intelligentization is an important topic of AWE2018. Many exhibitors have also displayed their own innovation results around intelligence. However, when it comes to intelligent exploration, it may be more effective to walk in the front line of artificial intelligence. Since its founding in 2012, Yun Zhisheng has become the industry's leader in six years, and many of the products on the show floor will often see voice solutions provided by Yun Zhisheng.
In addition to smart home, Yun Zhisheng also has its own layout in the areas of networking, health care, education, etc., and contributes to the smart landing of these areas. At present, what stage is artificial intelligence facing? What challenges does the industry face? Why does this identify? The artificial intelligence will break out. In the face of a series of questions thrown out by reporters, Kang Heng, vice president of cloud-known IOT business division shared his views.
Yun Zhisheng IOT Division Vice President Kang Heng
What is the reason for determining this time?
The AI's 'Fire' was well-known to all. It even was selected as the 'Top Ten Trendy Languages for the Chinese Media in 2017'. In fact, the concept of artificial intelligence was proposed as early as 1956, but it has evolved over the past half century. The process was tepid until AlphaGo defeated Korean professional nine-step player Li Shishi. Artificial intelligence has been noticed by the public. The main working principle behind AlphaGo is 'deep learning'.
Kang Heng said that the development of artificial intelligence requires three core elements, namely, algorithms, computational power, and data. The key technologies corresponding to these three elements are deep learning, cloud computing, and the Internet. Now that these three factors have matured, , so he judged that the artificial intelligence would not stop again. “In terms of algorithms, the emergence of deep learning has a huge impact on the artificial intelligence industry. Whether it is voice recognition or image recognition, it has greatly promoted the accuracy of artificial intelligence algorithms. With the improvement, the industry believes that the speech recognition rate of 90% is only available. Deep learning has reduced the error rate of speech recognition by more than 30%. This is an extremely significant improvement. To. '
In addition, Kang Heng also stated that not only technology itself promotes the development of artificial intelligence, a new technology is applied in life, but also needs the combination of industry, capital, and academia to promote this power more than purely. Academic research is much larger, and relatively weak links in this process of promotion can also be remedied.
AI's 'fire' has its advantages and disadvantages
From a technical point of view, artificial intelligence is divided into three stages: perceived intelligence, cognitive intelligence, and general intelligence. Kang Heng believes that current cognitive intelligence is relatively mature, and the accuracy of speech recognition has surpassed that of humans in certain specific situations. , And cognitive intelligence is still at a very early stage. There are still needs to be explored in terms of 'understanding', 'decision', 'judgment', etc. Although in some scenarios it can be done, but from machine to 'human' The distance is still very far away, and general intelligence is far from reaching.
From the products exhibited by AWE exhibitors in recent years, it can be seen that intelligence has become an irreversible trend in future development, but the 'fire' of artificial intelligence has advantages and disadvantages for the industry. Kang Heng said that artificial intelligence “Fire” is conducive to the promotion of technology development across industries and even the landing of products. It has great value for both companies and users, but at the same time it also has some negative impact on the industry, because once capital entry is involved, there will be 'For profit,' 'short-sighted,' or even 'sturdy seedlings', some less mature technology will be exaggerated, which makes everyone's expectations far exceed the actual capabilities of product technology, so there will be 'pseudo-intelligence'.
Kang Heng stated that although Yun Zhisheng was born in 2012, only six years have passed, the development of each step of the company has a clear context. 2012 to 2014 is the first stage, at which time the entire industry is still immature. When everyone did not pay attention to artificial intelligence, Yun Zhisheng did a lot of technology accumulation and exploration; the second stage was from 2013 to 2014, during which some technical innovations were tried, and also technical products During the process, a mobile phone voice input plug-in launched at that time achieved a good market response; the next step was to explore new directions that would allow the technology to bring about significant improvements to the user experience in some relatively loose areas.
Kang Heng told reporters, 'Each before entering an industry, Yun Zhisheng will have some judgments. The first is to judge whether the industry can form a large-scale application; Second, we must consider the next two to three years or three to five years. The commercialization of this industry. '
In the field of artificial intelligence, in addition to cloud knowledge, companies such as HKUST, BAT and others are also exploring. For this competitive relationship, Kang Heng said: 'We have a good word for 'competition,' competition. Cooperation, the positioning of Yunzhisheng and BAT are not the same. Yunzhisheng develops in the direction of verticalization. BAT expands the existing business. There is certain competition but not all areas are completely competitive. Competitive cooperation is Long-term status. '
For the future development strategy of Yunzhisheng, Kang Heng stated, 'First of all we are still in the direction of the Internet of Things, artificial intelligence to do positioning B2B2C, we hope to provide a complete artificial intelligence solution for some B2C companies, we will not do 2C The product, but it will not only be the product of the technology engine, we also hope that the technology-driven industry leading to some value, the gap between technology and user value across the past. '