The top level AI battle between China and the United States may start in the medical field for a long time

In the name of the White House, the United States, which boasts of being the leader of artificial intelligence, has issued an executive order with the main purpose of maintaining the dominance of artificial intelligence The plan includes five dimensions: policy, objective, responsibility, government investment and AI resources, and explains the important issues concerned by the AI industry, such as R & D capability, resource allocation, approval process, moral standards, talent training, etc There is no doubt about the strength of artificial intelligence technology in the United States So far, no one has been on the right However, there are many R & D teams from all longitudes and latitudes of the earth, and any team has the possibility to catch up with this still advancing technology Since trump took office, similar precautions have not been common But this time, the political order directed directly at technology has undoubtedly injected a shot of heart power into AI enterprises It has the potential to regard AI as the same great title of "era" as "Internet Era", and the sense of urgency is conceivable Can AI shoulder the heavy task of this era and bring about another change of productivity? At least in the medical field, AI entrepreneurs firmly believe that, and are constantly working hard In this field, the battle between China and the United States has already begun What can the United States do through state power? For an emerging technology, government level influence is crucial to the development of AI The medical artificial intelligence, which broke out in the second half of the Internet wave, has been facing a mature regulatory system since it was active There is no wild development period of traditional equipment and medicine, and its every move has strict regulatory regulations Here, we think about the possibility of the future from the perspective of data and approval For AI technology, algorithm, computing power and data are its three core elements, and the core is data Any algorithm needs the baptism of data, and the quality, completeness and accuracy of data directly determine the performance of AI in actual combat So no matter China or the United States, holding the source of data is holding the throat of the development of artificial intelligence In the second part of this administrative order, it is mentioned: "enhance access to high quality and fully traceable federal data, models, and computing resources to increase the value of such resources for ai r & D, while maintaining safety, security, privacy, and confidence protections consistent with applicable laws and policies." (enhance access to high-quality and fully traceable federal data, models, and computing resources to increase the value of such ai r & D resources while maintaining security, security, privacy, and confidentiality protection in accordance with applicable laws and policies.) “ Enhance access to high-quality and fully traceable Federal data, models, and computing resources to increase the value of such resources for AI R&D, while maintaining safety, security, privacy, And confidence protections consistent with applicable laws and policies These two contents unreservedly reflect the efforts that the U.S government will make in the direction of data to promote the development of AI By contrast, China is not lagging behind On the contrary, under the leadership of the State Drug Administration, we are on the road of independent innovation In 2018, the database of pulmonary nodules and fundus has been initially established, many AI products have also passed the mnpa class II device certification, and the approval process of class III devices, namely the course, has also been published All kinds of behaviors show the determination of the state to promote technological innovation But under the new challenges, will the state make a new combination, open the open database or relax the approval to promote the development of AI commercialization, and then increase the talent and capital investment in the field of AI through the market? We don't know, but the promotion of approval policies and the formulation of industry labels such as data standardization and desensitization similar to those in the United States will become the main direction of the FDA's efforts, that is, "progress in stability" When it comes to approval, compared with the past CFDA in China, FDA has more experience So far, the FDA has approved a number of AI products, including the oncology AI products of atherys in the United States, the brain stroke AI products of viz.ai, the brain image analysis AI products of aidoc in Israel, that is, pet related AI products of the Chinese American team shuttle medical Domestic enterprises dare not slack in the direction of examination and approval In addition to striving for innovative device certification and nmpa class II device examination and approval qualification, many enterprises have also applied for products to FDA and obtained certain results In addition to a number of AI companies approved by Category II devices in 2018, at the end of 2018, the "ai-ecg platform" of the ECG AI automatic analysis and diagnosis system of Lepu medical was approved by the U.S FDA, becoming the first ECG AI product approved by the FDA in China China is far ahead in healthcare financing Although in the first quarter of 2018 alone, American AI start-ups swept away $1.9 billion in venture capital, according to the statistics of the first three quarters, only a few of them flowed into healthcare According to the data of eggshell Research Institute, from January to August 2018, there are only 14 medical AI start-up projects in the United States, with a total financing amount of about 470 million US dollars, while there are 55 in China, with a total financing amount of 792 million US dollars From the perspective of the overall financing trend, although there are few projects in the United States, the single financing amount is relatively large, the projects are relatively mature, and the distribution is relatively uniform In China, most AI products are concentrated in the field of artificial intelligence medical imaging The financing amount of Yitu Medical Co., Ltd and yunzhisheng is nearly 200 million US dollars The total financing amount of other 53 projects is less than the total financing amount of 14 projects financed by the United States But from the perspective of start-up companies, investment in China and the United States has its own advantages and disadvantages Although the amount of financing for various projects in the United States is large, the lack of early projects means that there is not enough fresh blood to participate in the competition, which is not the appearance of a cutting-edge technology; although there are many new projects in China, there is a trend of image aggregation, competitive understanding, and the technology of new drug research and development, hospital management and so on is lack of participation The layout of medical AI companies in China and the United States (data from Liu Shiyuan's sharing at the first medical imaging Conference) in the final analysis, the difficulty of commercialization is the common problem faced by China and the United States, and it is also the core of restricting capital inflow As Dr Fan Wei, head of Tencent AI laboratory, said, "we are focused on R & D, not on commercialization." Although artificial intelligence has been popular for several years, it is only a few years In this regard, we can see that in the chapter of "government's investment and development in artificial intelligence" in the fourth part, the author puts forward that "the project in charge of the project leader of artificial intelligence will be regarded as a priority, and has the qualification of preferentially obtaining budget R & D funds." This will solve the problem of project funding next to data Perhaps, we can see the US AI start-up projects in more fields in the new year As the unicorns of hope, what are the comparisons between the two sides? Recently, CB insights released the top 100 list of global AI Among them, six Chinese companies were selected, including Shangtang, Yitu, Kuangshi, momenta, horizon and the fourth paradigm If startups represent the vitality of an area, unicorn is a symbol of hope in that area Of the 11 unicorns on the list, 5 are from China, and the other 5 are from the United States They are tied In these enterprises, Shangtang, the fourth paradigm and Yitu all have made corresponding layout in the medical field, and Yitu takes it as the core business, and establishes Yitu medical separately to develop the application of artificial intelligence in medical treatment, which has made great achievements in lung nodule image, natural semantic processing, etc Recently, a paper entitled "evaluation and accurate diagnosis of pediatric diseases using artificial intelligence" was published online in nature medicine, an international well-known medical research journal The experiment is based on NLP technology So, although Trump's action is radical, it is still reasonable How can the consistent American pride make him bear the grievance of equal standing? What is the layout of faage and bat? Even in terms of innovation, China has enough capital to compete But when comparing strengths, we must consider the world giants like faage in Silicon Valley, in which Google has poured huge energy In November 2018, deepmind announced that its health department, deepmind health, and the team responsible for promoting "streams" (mobile app to help doctors identify and diagnose patients' conditions faster) will be adjusted and merged into Google's newly established "Google Health" department The atomic company deepmind health will no longer exist as an independent brand, but other departments of deepmind will remain independent In contrast, bat's layout in the field of artificial intelligence also has no trace In the direction of medical AI, Tencent regards it as an important layout to get through to B path In addition to the core Tencent medical AI laboratory, relevant excellent map laboratory and AI Lab also provide suggestions for Tencent's medical AI layout At the end of 2018, Tencent won the key special project of "research and development of digital diagnosis and treatment equipment" to develop artificial intelligence assisted clinical decision support system (aiacdss) At that time, Tencent will obtain a lot of training data and AI development experience from scientific research cooperation, and Tencent cloud will play a more in-depth role in AI medical treatment Alibaba deployed et industrial brain, et urban brain, et agricultural brain and et medical brain with its powerful cloud service In the medical field, Alibaba cloud is committed to working with participants in the medical field to focus, be cautious and be smart in applying data intelligence to help doctors and nurses provide better medical services for patients and save more lives Baidu's AI strength has been ranked at the top of the world, with strong computing power Under the premise that most AI image algorithms are migration algorithms, Baidu, which focuses on computing technology awareness and driverless technology, wants to enter the field of medical AI very easily The key is whether it is necessary to make direct investment in this field At present, baidu Lingyi has made efforts in the medical field Today, the advantages of the United States are not as obvious as before, and faage itself has constantly created problems for the federal government due to monopoly, data and other reasons Tencent, which just won the "tieba" of the United States, shows more the ambition of bat's merger and acquisition of the world No matter from which angle, it is imperative to keep the dominant position of the United States in the frontier technology Looking back at Huawei and ZTE, the unease of the United States may be that although the gap between China and the United States in artificial intelligence technology is narrowing, the United States still has the means to control China's progress Arterial network reporter learned that in the field of medical artificial intelligence, almost all enterprises have used NVIDIA and Intel chips (or cloud computing) to provide computing power support for training algorithms Thoughtfully researched and developed the AI medical related chips, but the training algorithm still needs the support of NVIDIA Huawei is not the first company to be restricted by law in the United States In the past decade, there have been dozens of large-scale judicial and regulatory agencies in the United States