AIR 004 kingsoft CEO Zhang Hong Jiang intelligent big data cloud

Lei feng’s network (search for “Lei feng’s network”, public interest) by: sorted CCF-GAIR artificial intelligence and robots in the World Summit, kingsoft CEO Jin Shanyun CEO Zhang Hong-Jiang and speech. Starbucks iPhone 6 Plus Case

Why artificial intelligence with such rapid progress in the past few years, Zhang Hong-Jiang believe that it ultimately comes down to large progress and the evolution of cloud computing.

Support of AI: the data explosion

Progress in artificial intelligence largely due to progress in data, data on global consumption, 2013, 2020 will increase 10 times, 40% a year, app Exchange pictures in the circle of friends is 1 billion a day.

Zhang Hong-Jiang, for example, if users have saved to millet millet on the cloud photos, more than 500 million words, needed 1.2 million stones, 100P of storage space is required, and today China stock photos, this is data from 2014, when 300 billion, requires more storage space and calculation of how huge it is.

The dream of 20 years ago

Applications on the phone, can do to face a series of processing, such as face recognition and age determination. It’s not a phone, but the phone’s back-end running on a cloud, when phones took a picture, then, in the cloud, and then get the results back, Zhang Hong-Jiang their dream team 20 years ago.

1997 Zhang Hong-Jiang patent of a team at HP Labs, which is a framework for distributed, data received from the client, and then passed on to the back-end system, do a series of processing, and pushes the result onto the front end. Today, after 20 years, and finally implemented on this matter on a cell phone, and more importantly, it can be realized on cloud and end. Mobile phone into a data acquisition Terminal, carrier cloud computing as a whole, the whole data carrier.

Face recognition technology face the history history of the expansion of the database, is to face database + capacity expansion history

Academic years, face life in the early 90 ‘s when all academic databases that are used in the era of only hundreds of face images, even in this case the recognition ability of the robot is still below the person’s ability to identify, and light requirements are very high, the need for positive images. In the industrial age, in 2014, the face recognition rate of 97.35% in large quantities, and in 2015 to reach 99.63%. Domestic standard of companies in more than 70 million photos, plus millions of other photos, people use the deep Web has more than 30 levels, training 20 NVIDIA, accuracy is far more than humans.

AI+ data + cloud: enterprise standard

Companies today do not have the data it is difficult to survive, and data to be supported on the cloud. Include quality data acquisition and analysis of data, and in application of its resolution has become the corporate standard. Of course, this fact must be supported on the cloud, no clouds, no data, no large data acquisition, processing and data management.

Cloud computing scenarios Starbucks iPhone 6 Plus Case

Zhihou, Zhang Hong-Jiang about cloud computing and large data in application of relatively mature in the following areas: medical, Government, gaming, video, cloud, hardware only cloud and hybrid cloud.

Cloud computing trends

Zhang Hong-Jiang said cloud computing has become the mainstream IT technology, Internet industry matured fast company, up to $ 10 billion sales company, c-side is Google, most of the Facebook,B side of the Amazon AWS. When the b side achieve such high sales, indicates that the market has matured.

Threats to human and artificial intelligence 2.0

Question for artificial intelligence problem of the future will pose a threat to humans, as well as artificial intelligence 2.0.

The first issue, Zhang Hong-Jiang said, there is a curve, no people are invisible to the singularity, unable to judge. For the problem of artificial intelligence 2.0, he thinks maybe artificial intelligence 2.0 is not necessarily to drive, strong signs of artificial intelligence can be the problem of inference. Humans may not need to wait for the future has been very understanding of the human brain to make artificial intelligence, or talk about 2.0.

Finally, Zhang Hong-Jiang, summed it up:

Artificial intelligence we do, when we see a rapid breakthrough as artificial intelligence, we should pay more attention to its back-end, not only is the algorithm itself by leaps and bounds, and more importantly supporting the algorithm data and cloud computing by leaps and bounds, so we see in the future is to front a series of intelligent devices, but its brain is actually inside the cloud.

Like this:

Like Loading…