Since the State Council issued the "New Generation Artificial Intelligence Development Plan" this year, the strategy of developing the artificial intelligence industry at the national level to promote the people's livelihood economy can be said to be very clear.

With China’s heavy investment in the artificial intelligence industry, more and more voices are beginning to compare the strength of artificial intelligence between China and the United States. Numerous think tanks and data agencies have successively issued reports on the analysis of artificial intelligence engineering, commercialization, and scientific research in the two countries. Different conclusions have been drawn, but it is an indisputable fact that the level of artificial intelligence development in China has reached the level of internationalization.

Although we talked about cutting-edge artificial intelligence technology, we talked more about Deepmind, OpenAI, and the AI ​​labs of universities. However, we must know that although the scientific research and industry in the AI ​​field are closely related, the academic level is ultimately not equivalent to the industry level.

Perhaps we can use another perspective to see why China’s artificial intelligence opportunities are valuable today: Among the most critical elements of AI industrialization, Chinese companies have not missed one.

Let us sort out from a few dimensions, the brief history of AI and the performance of Chinese companies in the third golden cycle of AI featuring big data + machine learning. From a broader perspective, we may agree that AI is a technical inflection point that China has not missed and cannot miss.

The source of ecology: deep learning framework

Artificial intelligence is a standard and invasive technology. Every clever algorithm and every kind of logic that embodies light may solve the problems that plague humanity. Therefore, it is one of the most important tasks for AI companies to bring developers, researcher tools and environment to allow them to display their talents.

When we first introduced the deep learning platform, we were mainly based on the Caffe launched by the University of California, Berkeley. Caffe creatively incorporated convolutional nerves into the development environment and built a deep learning framework for more efficient cleaning.

At the end of 2015, Google opened up TensorFlow, which was previously used internally. With several version updates, Caffe broke the monopoly position and became the most active deep learning framework. Especially with the full use of DeepMind, its community resources have been widely recognized.

In order to compete for developers' eco-Facebook, Microsoft, OpenAI and other companies have successively launched a deep learning development platform, and re-try to subvert Google's dominance. At present, there are still more than a dozen major deep learning platforms in the world.

How important is the developer environment, as evidenced by the increase in the power of speech that Facebook acquired in the AI ​​community with PyTorch. Having a benign deep learning framework is the foundation of the entire AI technology ecosystem.

Domestic enterprises have not fallen behind this point. For example, Baidu first introduced the PaddlePaddle, announcing the entry of domestic companies into this field.

The PAI launched by Ali Cloud has become a representative of a new generation of deep learning framework with a friendly development environment and computing power. It also provides a good foundation for domestic developers to enter mainstream mainstream deep learning and research and development.

At present, the domestic deep learning framework is gradually colliding with the mainstream development platforms in Europe and the United States through the development environment, community resources, and corporate incentive programs. At least in the core aspect of AI R&D, China has regained key points.

Hard-powered answer sheet: The base of computing power of the chip representative

The artificial intelligence neural network task model is a completely new task form that is completely different from the traditional operation requirements. In other words, the traditional hardware computing foundation will no longer meet the needs of artificial intelligence.

In the PC era, there are arguments that the chip-makers have the world, and in the era of artificial intelligence, it still seems to work.

In 2011, IBM introduced a commercially available brain-like chip through the imitation of the human brain, which is arguably the basis of the artificial intelligence chip solution. However, this model quickly proved to be of low real value, and gradually it was shelved.

Since then, FPGAs, ACIS and other types of chips that handle artificial intelligence tasks have been introduced. Nvidia is said to have discovered in an experimental error that the GPU can handle deep learning tasks well, and it also opened the GPU and artificial intelligence hubbub.

The NVIDIA Tesla V100 is currently known as the most powerful AI-specific processor. This year, Google released a TPU chip specifically designed to deal with deep learning tasks, and it also showed great promise to AlphaGo.

For domestic related industrial chains, how to obtain the computing services that can support the R&D and deployment of artificial intelligence and obtain intelligent AI-based migration has become an urgent task.

In early September, Aliyun announced the launch of a new generation of heterogeneous acceleration platforms. Heterogeneous computing family covers 7 heterogeneous instances such as GPUs and FPGAs, and comes with the new high-performance computing example E-HPC. It can be said that AI computing power is supplied to the industrial environment through specific service connection. And Alibaba Cloud's solution seems to be more open and customized, more in line with the needs of Chinese companies.

In addition to cloud computing, the terminal's AI chip has also become a hot topic, and Tesla’s unmanned chips have become popular. On the phone, Apple introduced the A11 bionic chip. Huawei previously released the world's first mobile AI chip Unicorn 970.

It is not difficult to see that the AI ​​game based on chips and computing power has been launched in all aspects of China and the United States. In the future, the cloud-integrated AI calculation seems to be a more practical solution. The convergence between cloud computing and hardware manufacturers is also very exciting.

Enlightenment: About the wave of smart hardware

Using AI to add ordinary terminals and open a new era of Internet of Everything seems to have long been a consensus in the industry. A wide variety of AI hardware also appeared early.

But there is no doubt that Amazon's Echo has opened up a new window of opportunity for smart voice + popular hardware. For a time, the world's innumerable smart speakers have risen. It has been spectacular for a while.

What deserves our attention in this field is the keen and strategic investment by Chinese companies in the wave of AI consumption. Just as Amazon's greater ambition to invest in Echo is to open up the family ecology scene. It is also difficult for Chinese giants to invest in AI hardware to escape the temptation of ecologicalization.

From Jingdong’s sister-in-law to Xiaomi’s little lover, Smart Speakers became a giant strategic point for a while. What attracts people's attention is Ali's Tmall Elves, which opened up the boundaries between Ali's cultural ecology, e-commerce ecology and technical strength. It is clear that he has built a wider range of intelligent experience.

Intelligent cats such as Lynx are more likely to become the entrance to the AI ​​of home scenes. Through the subsequent development of more AI-enabled imaginations, a complete upgrade of consumer and home experiences was achieved.

In a situation where the IoM matrix is ​​more complete, it is reasonable to believe that AI technologies such as machine vision and image recognition will compete to enter the consumer intelligence hardware sequence. After starting to experience AI, the smart consumer market is likely to have just emerged.

From the laboratory to the physical world: the scene of AI landing

If AI technology can only stay in the laboratory, I am afraid that only one percent of today's people will pay attention to it.

After all, any technology needs to test its value in the real world. AI is also a very low-level technology and it is easy to connect with the efficiency, experience, and solution patterns of various industries, giving people unexpected results.

Although we are accustomed to the value brought by many AIs today, we would like to think about the initial contact with Google ML's smart recommendation, Apple Siri's voice interaction experience, and Tesla's vehicle collection data. Feelings are overwhelming.

Perhaps it is possible to lay the groundwork for the mobile Internet. Compared to the U.S. market, China has a larger AI market and a better AI acceptance environment. And Chinese companies have also created many “Chinafirst” in the field of AI's scenario landing.

For example, in the field of voice interaction, we have seen Baidu’s degree secret and HKUST’s two major giants fly together. In the field of machine vision, there have been five unicorns such as Shang Tang and Despise, which are truly amazing.

Here we must mention Alibaba Cloud's ET brain series. The deployment of ET city brains in Hangzhou is the first time in human history to use the big data + artificial intelligence method to direct real city traffic. Subsequently, the ET industrial brain with 1% increase in production capacity, the ET medical brain with higher accuracy of X-ray films than human doctors, and the ET environment with the ability to monitor soil and water pollution and the marine environment were successively on-line.

Alibaba Cloud's series of deployment apparently followed the AI ​​landing path with "Chinese characteristics." Combined with the capabilities of data sensing and large-scale computing, artificial intelligence can be exported to more practical social values. This model should continue to mature in the next Chinese market.

Taken together, the use of artificial intelligence in Chinese companies to achieve specific industries to empower and innovate, the industry base in all areas required has already matured. Relying on the massive user network, the trinity value logic network composed of big data, cloud computing, and artificial intelligence may become a unique AI maturity mode in China.

In short, the creativeness in the scene will be the most desirable aspect of the next Chinese AI.

Deeper Future: Talent and R&D Convergence Effect

Many big cows have stated candidly that AI hits a war of talent in the end.

But this may not be accurate, because scientists and technologists who can influence the industrial situation are very few. After all, the actual situation of the war situation should be the maturity and innovation ability of the R&D team and laboratory.

From DeepMind's scientific madman culture, it is not difficult to see the strategic significance of AI talent in corporate experiments. The race of this kind of alternative track actually started early on our side.

Regardless of the AI ​​laboratory culture and engineer culture in Europe and America, it is not mentioned that in China, the AI ​​companies equipped with technical research teams are not new.

One of the earliest established Microsoft Research Institutes in Asia was the exploration of artificial intelligence. The local enterprise BAT's AI talent aggregation project is not backward.

For example, Baidu established IDL (Institute of Deep Learning) in 2013, which mainly tackles technical problems in the fields of deep learning, speech recognition, and intelligent training. In 2014, Ali also established iDST (Data Science and Technology Institute). It is said that iDST shoulders the era of big data, highlighting Ali's high level of responsibility in the underlying technology. This year, iDST has successively pocketed talents in the scientific community such as Golden Harvest, Hua Xiansheng, and Ren Xiaofeng. iDST's capabilities in forward-looking research and business scenarios have reached historical highs. The story after this is worth the wait.

Another widely watched phenomenon is that BAT has established AILab in succession.

The earliest established Baidu Silicon Valley AILab has been relatively active academically, and the results are also dominated by papers and open source algorithms, resembling the academic schools of large European and American companies.

In March of this year, Zhang Wei, who served as the vice president of Baidu Research Institute and the head of the Big Data Lab, was named director of Tencent's AIlab. Relatively speaking, Tencent AIlab pays more attention to the engineering and application of AI. The research field focuses on computer vision, speech recognition, natural language processing, and machine learning. It is relatively focused on the integration of Tencent's rich scenes with AI.

In July of this year, Ali AiLabs announced its appearance, and Wang Gang, a former professor at Nanyang Technological University, announced his participation. Compared to iDST, Alibaba AIlabs focuses on the development of consumer-level AI products. This is evident in the Lynx Elves that appeared in the same day as AIlabs.

According to the relevant report of Goldman Sachs, over 80% of the deep learning research results in the world will be produced in China. The AI ​​R&D team and laboratories among major companies will be the main driving force for these achievements.

The deeper technology future will always be triggered by people.

Write last

The above sections correspond to the development history of AI business from different angles. It is easy to see that China’s artificial intelligence has not fallen behind in all areas, and at least it has the ability to rank among the world’s leading voices.

The future development momentum of AI comes from integrating advantages, immersing academics, and developing ecology. Perhaps the simple comparison between China and the United States is actually meaningless. It is fundamental to know what advantages and weaknesses we have, and then release deeper creativity on this basis. For example, the upcoming Ali Yunqi Conference is said to also focus on artificial intelligence as a target, introducing new understanding and technological achievements of artificial intelligence.

Communication and the director of the exhibition are often the real motive force of technological evolution. As someone said, in the eyes of Silicon Valley developers can see the curiosity and romance that an Asian colleague does not have. However, in the eyes of practitioners and developers of Chinese AI, there is a sense of hunger that cannot be lost.

Copper Lugs

Copper Lugs,Copper Cable Lugs,Plating Copper Cable Lugs,Copper Tube Terminal Lugs

Taixing Longyi Terminals Co.,Ltd. , https://www.longyicopperterminals.com