[Introduction to Xin Zhiyuan]Intel and Alibaba are both aware of the importance of data in the AI era and are carrying out data one after another.
Intel, Ali two big factories have given such a clear definition. In the era of artificial intelligence, tapping the potential of data has become a new generation.
As early as 2014, Ma Yun said in his public speech that:The world is entering the DT (Data Technology) era from IT (Information Technology).
Kit HoChee, vice president of Intel, also said:Intel has transformed itself into a data-centric company.
And Alibaba and Intel also revolve aroundIntel signed a memorandum of strategic cooperation with Aliyun to bring innovative technology to the Tokyo Olympic Games in 2020 and the Beijing Winter Olympic Games in 2022.
September 25, at 2019At the Round Table on "Cloud Intelligence Accelerating Data Potential Release", chaired by Yang Jing, founder and CEO of Xin Zhiyuan,Jia Yangqing, Vice President of Alibaba, General Manager and Senior Research Fellow of Aliyun Intelligent Computing Platform, Min Wanli, Founder and CEO of Beifeng Capital, Dai Jinquan, President of Intel Big Data Technology Global CTO, Big Data Analysis and Artificial Intelligence Innovation, and Guo Nan, Co-founder and COO of Huiying, discuss data together. In the era of dividend, the development of AI and the new trend of science and technology will create the future ecology.
AI is data-based while Pratt-Whitney AI is people-oriented.
Although this round-table dialogue is based on technical trends, the whole process has been full of humanistic flavor. As we all know, Alibaba has his own flower name culture.At the beginning of the dialogue, Ms. Yang Jing helped everyone ask about Ali.RadicalThree points of water, that is,: Yang Qing.Min Wanli, a former chief scientist of Ariyun, told Ms. Yang Jing that her flower name was
Not long ago, Intel released China's first documentary "Hello AI", which shows the application of AI from a social and humanistic perspective, in order to record the real impact of AI on human beings, in the round table dialogue.YangCollect enough data to retain fast disappearing languages with smart voice, and AI helps more people with surgery。
First, how should the potential of data be released for the benefit of mankind? Data Gold Rushing in AI Era: Data will Drive Intelligent Decision-making in 5 Years
The essence of artificial intelligence is data technology, which forms intelligent decision through data processing, and it is the frontier field of DT.
At this round table discussion,Min Wanli, former chief scientist of Aliyun and founder and CEO of Beifeng CapitalIntelligence is the general trend of the times. After the construction of IT, today's informatization, or the group behavior of human society, the tagging, standardization and labeling of individual behavior have been completed.Nowadays, whoever can find a way to create intelligence and value can enjoy the dividend of the industry.
In this direction, we will find that all data-intensive industries and industries, such as the Internet, smart cities, health care, are gold mines and oil fields to be reclaimed today.
The use of cloud computing, algorithms, development platforms and other tools will lead to the trend of oil mining and gold mining in the new era.
Unlike the real mineral energy, intelligence is just beginning, the potential space is huge, more importantly, the raw material used for intelligence is data computing power, and data is not non-renewable energy:Data are generated every minute and every second, and the release of value requires quick-sighted technical means.
Min Li said that the data potential was only thought ten years ago, and now there are technologies such as cloud computing and AI, so there is little cost for value creation.
This is completely different from the way capital emphasizes capital in the original industrial revolution. All actions are generating data, and people can afford supercomputing. Whoever has creativity, industrial experience, who has the courage to find out. Like gold diggers in California in the 19th century, today's bets on data potential may be the next.
Jia Yangqing, Vice President of Alibaba, General Manager of Aliyun Intelligent Computing Platform, Head of Alibaba Open Source Technical Committee and Senior Research Fellow
Alibaba vice president, Ali cloud intelligent computing platform division general manager, Alibaba open source technology committee head, senior researcher Jia Yangqing also expressed strong belief:
Nowadays, when AI solves all kinds of problems, the computer system itself will produce a lot of data. Today, many of our system designs are based on choices. For example, when there is a large amount of data storage, we need to consider which data is stored in cold storage and which in SSD. These decisions are made by observing simple data statistics through our brains.
If we believe in the power of AI, then in a large number of data statistics and processing process, there will be a more intelligent, algorithm-driven decision-making process.
2. Jia Yangqing: The Chinese algorithm has been in the forefront of the world, but its application needs to advance with the algorithm.
In this round-table forum, Jia Yangqing, as the guide of algorithm and open source framework, also talked about the development status of artificial intelligence in China.
In response to a question from Ms. Yang Jing, founder and CEO of XinzhiyuansolutionAn Analysis of what stage China has reached in the Arena of artificial Intelligence and what challenges it will face in the FutureJia Yangqing said that in terms of algorithm and engineering ability, the contribution of Chinese engineers and scientific researchers has been able to bring this aspect into line with the world, even at the forefront.
Jia Yangqing himself is one of the three most popular open source in-depth learning frameworks Caffe,TensorFlow and PyTorch.,amongCaffe and TensorFlowThese two world-class frameworks directly established Jia Yangqing's AI status in Silicon Valley:
Caffe is the first general deep learning framework in history and one of the most widely used deep learning frameworks in the world.Microsoft, Yahoo, Nvidia, Adobe and other world-class companies are using it.
TensorFlow, had to mention the robot AlphaGo, which became famous for go (AlphaGo), TensorFlow is the powerful support behind AlphaGo, known asGoogle brain
In addition, Jia mentioned Amazon's chief scientist and one of the major contributors to MXNet's in-depth learning framework in the round table dialogue.Li MuAnd the new assistant professor of CMUChen TianqiAnd other Chinese scholars have made outstanding achievements in algorithm.
Speaking about future trends, Jia said: from the perspective of the development trend in 2019, the development speed of the algorithm will slow down, or become stable.When we talk about two steps, the algorithm and the application should each take a step forward. Today's biggest opportunity lies in how technology can land in the application.In a scientific research field, a single shot is OK, but in practical application, such as how to integrate data and business logic very closely, is the current trend, and also the point where AI can create value.
When it comes to the development trend of technology in the future, Jia Yangqing also said that on the basis of finding data, using data and managing data, we need to have the ability of comprehensive data management in the future.
Min Wanli also stressed that there may be a growing demand for the use of hybrid technology in the future.
Intel Big Data Technology Global CTO, Big Data Analysis and Artificial Intelligence Innovation Dean Dai Jinquan
Well, if the data is oil, then the calculation is the oil dredger. Dai Jinquan, president of big data Institute for Analysis and artificial Intelligence Innovation, global CTO, of Intel big data Technology, believes:In different computing architectures, the collaboration of hardware and software brings about the improvement of computing power.
Intel and Ali are betting on a big move: unleashing the data potential of the AI Era
Nowadays, cloud intelligence accelerates, and cloud computing brings computing power to artificial intelligence, big data, the Internet of things and other fields.As Ms. Yang Jing said:The intelligence of the cloud has indeed become a nuclear weapon. It has solved many painful points for all walks of life in tapping the huge potential of data and moving towards intellectualization.
Talking about the power of Alibaba's nuclear weapons in this regard, Jia Yangqing said: now that people are using the power of data more and more, this is becoming more and more demanding for us, especially in terms of hardware.
Jia Yangqing explained that this seems to be a law, the ability of hardware is always far from enough, because the software will always put forward higher requirements.From AlexNet to AlphaGo Zero, the demand for computing power has increased by more than 300,000 times.
Alibaba's own data has been exponential growth, but the infrastructure and all the system optimization processes are very much in need of the ability of hardware collaborative design.
When building a large data and AI system, we use Ali's observed different computing models to optimize the end-to-end from the work scenario, which reflects the value of hardware and software collaboration.
Jia Yangqing stressed that there will be more possibilities in the field of hardware chips in the future, there will be competition and cooperation, internal players need to exchange demand, share experience, and make the cake bigger.
Software and hardware collaboration has always been Intel's strong point. With the support of Intel's software and hardware collaboration, Aliyun has made new progress in large data computing capability.
On September 26, at the 2019 Yunqi Conference, Intel and Alibaba jointly announced thatInternational Big Data Performance BenchmarkingTPCx-BBinIn 2005, Ariyun MaxCompute, supported by Intel, set a world record on the scale of 100TB data.TPCx-BB is the most comprehensive end-to-end large data standard test set in the industry. Aliyun has become the first cloud service provider to announce TPCx-BB results.
In addition, Ariyun's EMR (Elastic MapReduce) big data service in another test TPC-DSTwoChina also set a record.As a cloud service provider, compared with the results released by OEM, Ariyun big data's performance not only measures the capital investment of software and hardware, but also includes the consideration of operating costs.
These two records show that MaxCompute, based on Intel-to-Strong Extensible Processors, not only has leading end-to-end overall performance, but also takes the lead in achieving measurable cost-effectiveness advantages in the cloud service market.
Intel and Ariyun sign a memorandum of strategic cooperation to bring innovative technology to the 2020 Tokyo Olympic Games and the 2022 Beijing Winter Olympic Games
In addition, Intel signed a memorandum of strategic cooperation with Aliyun and announced that:For the 2020 Tokyo Olympic Games and the 2022 Beijing Winter Olympic Games, close cooperation will be carried out in the fields of 3608K VR,3D digital twinning, cloud broadcasting and so on.The two sides will also cooperate in the field of artificial intelligence 3D athlete tracking for the Tokyo Olympic Games in 2020.
Intel will provide support for high-performance processors and AI technology platforms in these innovative areas, and Ariyun will play a strong role in cloud infrastructure and services.
IV. AI Medicine, excellent release of data potential
Intel big data technology global CTO Dai Jinquan also believes that,Next, how to apply AI model and deep learning model to production environment is the key.Intel, Ali, Huiyi Huiying are constantly trying to achieve a series of results.
AI Healthcare is a typical data-intensive area, a production environment that takes full advantage of data potential. Intel's strategic investment in Huiyi Huiying is in the medical field, using AI to dig out data.
Guo Na, co-founder and chief operating officer of Huiyi Huiying, said:
In addition, Huiying, together with Intel, has launched a full-cycle health management cloud platform for breast cancer AI, providing a full-data link monitoring and management platform and a full-cycle breast disease solution.
People's pursuit of health will not stop, and the field of AI health care is closer to people's bright expectations for the future scientific and technological world.
In fact, when we plan not only from the technical point of view, but also from the technical point of view.It is easier to stand in the waves than in the valley when you move away from the scope of technology and look for the future course from a wider industry space.
At this round table forum, Min Wanli emphasized this point:Today's R & D process is increasingly dominated by non - technical factors. 。 When planning for the future, we need to consider what kind of problems the development of this technology is to solve for human beings, whether it has landing scenarios and possibilities, and finally return to the closed-loop of any technology to ultimately create value.
Use vernacular and non-technical language to clarify the logic of technology and business logic closed-loop. With such ability and awareness, the combination of AI, cloud computing and industry will accelerate, and the digital transformation of the industry will be upgraded in an all-round way.