This article comes from WeChat public account: amplified light (ID: guokr233) , planning: light amplification team of: large sheep, drawing from the title: Baidu official micro (Apollo eco-partner intellectual Walker clean and disinfect vehicles)
When it comes to the Great Leap Forward in China’s science and technology, people will always praise the three good students of the “Internet”. But under the prestige, whether the Internet is as magical as people see and hope, and the sudden outbreak at the end of 2019 is a touchstone. Although it is good for Internet companies to donate money, it is no longer enough to satisfy people’s “technical imagination”: Facing the epidemic, what can Internet companies do?
Extreme skill, test it with epidemic
To this day, the best way to control the epidemic is still simple physical isolation, which is essentially cutting the “chain of relations” . Similarly, the business model of most Internet companies is working on the “relationship chain”: serving independent users scattered in front of individual terminal devices. Therefore, the control of the epidemic can be described as a hit with the Internet.
Internet companies contribute money to provide services for the epidemic, but this is fundamentally different from the work of medical institutions. Can the “new technology” they come up with really help the epidemic prevention and control?
Use the most data to make the most like platform
In the three apps of Baidu Map, Gaode Map and Tencent Map, the “Outbreak Report” entrance has been added to the prominent position on the homepage. The three companies cooperate with the Health and Health Commission and other agencies in data. In addition to real-time updates of detailed data on the epidemic situation across the country and across the country, they are also uniformly equipped with functions such as intra-city inquiry and hot-out diagnosis.
Comparison of data entry 丨 Enlarged lamp team drawing
The product functions of the three companies are similar. The data comes from the national and local health and health commissions. (Occasionally, due to the speed of real-time updates) >. Not only BAT, but similar epidemic broadcast products are not uncommon in the core products of many large technology companies, such as maps, browsers, news apps, etc., intended to provide users with reliable information services.
The repeated information provided by the functions that consume a lot of resources can of course meet the user’s needs for epidemic information. However, if the user appears suspicious symptoms, it is difficult to say whether he can use these products to get treatment. With the launch of epidemic reporting products, related functions are no longer a bonus for App or brand customer acquisition capabilities.
There are also a handful of differentiated products, such as Baidu’s Migration System. It has some reference value for office workers to plan the return time and avoid the peak of passenger flow during the Spring Festival, but that’s all.
Baidu migration system | Official website screenshot
In short, big data is just an efficient information provider. Data platforms can neither eliminate viruses nor protect against virus infections.Its role is nothing more than reminding users that “the epidemic is on their side” and do personal protection in a tense atmosphere.
Cloud computing makes detection faster, but who is doing it?
In terms of cloud computing, the first response is still BAT. On January 29, Ali, Baidu, and Tencent successively opened cloud computing power to research institutions to support virus genetic sequencing and mutation prediction, new drug screening and research and development.
On February 1st, Zhejiang Provincial Center for Disease Control and Prevention launched a whole-genome detection and analysis platform jointly developed with the medical AI team of Alibaba Group and Jieyi Biotechnology Company. The diagnosis of suspected pneumonia was shortened from hours to half an hour. [1]
Although Jieyi Biological’s new coronavirus nucleic acid detection kit has not yet been approved by the State Food and Drug Administration for marketing, it has also been allowed to be tried out in medical institutions and reviewed according to procedures after the epidemic has been resolved.
In the final analysis, the first attempt of cloud computing in the field of anti-epidemic can only help, and it really depends on biotechnology companies.
One to two days after the publication of the new coronary pneumonia virus gene sequence, (January 12) Kits have been launched for virus detection. On February 2, a new coronavirus detection kit produced by BGI was approved. In just one week, BGI produced more than 650,000 copies of the new coronavirus detection kit, with a daily production capacity of 100,000 copies / day. Determined body temperature, uncertain privacy
Relevant departments and organizations such as the Ministry of Industry and Information Technology, the Ministry of Science and Technology have solicited artificial intelligence technology solutions related to “infrared temperature measurement” from the society in order to cope with the peak return year after. Pan-Internet companies such as Baidu, Deaf, Shangtang, and Sankai Kaitian have launched corresponding solutions.
Baidu and Shangtang ’s “infrared temperature detection + identification” solution is more conventional: As long as you set a fixed point at the flow checkpoint,Install an infrared camera and use infrared thermal imaging technology to detect its forehead temperature. The technology supports the simultaneous detection of 3 to 5 people in a certain area at most, and does not require the active cooperation of passersby. [4] , mainly optimize the recognition scene wearing a mask, and improve the accuracy of AI infrared temperature measurement.
AI infrared temperature measurement system “Ming 骥” Backstage | 猎 云 网
Sanlit Kaitian’s public emergency medical tracing system combines AI technology and big data. After identifying high-temperature personnel, it compares with the public security face database based on face information. This determines the identity of the object. Liberation Radiology? May not
Since the outbreak of the new crown pneumonia epidemic, the dispute over the “diagnostic criteria” for nucleic acid kits and CT examinations has attracted attention. Internet companies cannot produce kits or CT machines, but they can write articles when interpreting CT images.
On February 10th, Huawei Cloud announced the development of an AI-assisted diagnostic service for new coronary pneumonia, which generates reading results within a few seconds. 7] ; 5 days later, Alibaba Cloud also launched similar products, claiming an accuracy rate of 96% sup> .
Huawei Cloud AI Quantitative Assisted Diagnosis | Huawei official website
How fast the AI can be is not the key to the interpretation system. The radiologist is more concerned about the accuracy. Coronary pneumonia is an acute illness, but AI learning often requires slow work and fine work.
AI reading is not a new technology. Shanghai Chest Hospital introduced a set of “AI reading” auxiliary system in 2017, which reduced the reading speed of 10 minutes / sheet to a few seconds. At that time, the Shanghai Chest Hospital had more than 16,000 lung cases per month, and the monthly CT readings exceeded 6.4 million. , Alibaba Cloud’s products have 5000 learning cases [8] The tens of thousands of samples from chest hospitals cannot make the AI reading perfect, let alone the new disease of new coronary pneumonia.
Artificial Intelligence Reading System | Shanghai Chest Hospital
A bite-feeding AI reading system, how should we use it?
From the outbreak of the epidemic to February 7, 21 doctors of the Radiology Department of Wuhan Jinyintan Hospital worked in shifts for more than 30 days and watched about 1.5 million frames. (600 ~ 800 frames per film) . [10] If the AI can replace the reading, can the 21 doctors at Jinyintan Hospital not need to use it? So hard? No.
In fact, in the workflow of the imaging department, what AI can do is to perform the first screening work, improve the work efficiency, and find out the image details that may have problems. The follow-up interpretation still needs the image doctor do.
Robot that beats hands
In order to reduce unnecessary contact between health care workers and patients, and to avoid accidental infections as much as possible, automated intelligent hardware is a good idea. More and more hospitals are using intelligent robots developed by Internet companies.
In this epidemic, the companies that earlier applied intelligent robots to medical institutions are Cheetah Mobile and Dada Technology. The former is used in place of manual guidance, preliminary diagnosis and treatment of patients, and even as a communication intermediary to help doctors perform remote diagnosis and treatment to avoid direct contact between medical staff and patients. [11] ; the latter is used to replace the medical staff to perform remote care, temperature measurement, disinfection, cleaning and drug delivery, and to alert the emergency situation
Cheetah Mobile Intelligent Service Robot | Hexun.com
Dartech’s intelligent service robot launched in Beijing Ditan Hospital | Dalmat’s official website
From the perspective of functions, the two medical robots are expected to liberate many grass-roots medical workers who need to have a lot of contact with patients, replacing some basics such as guidance, temperature measurement, drug delivery, and disinfection that do not require much professional skills. Work, but it can only be “hit and play” like this, at best it is “service robot”, which is not higher than the automatic food delivery robot you use when eating at Haidilao ( In fact, Purdue Technology and Qinglang Technology, manufacturers of food delivery robots for Haidilao, both opened the robot to the quarantine zone [13] ) .
Self-driving cars, please don’t mess up
In the past few years, Baidu has been proud of the Apollo self-driving car, but in the Wuhan epidemic, Huawei, which was killed halfway, took the lead.
On February 4th, two Huawei 5G unmanned vehicles were sent to Wuhan Pneumonia Epidemic Prevention and Control Headquarters. Huawei’s unmanned vehicles include three modules: autonomous driving, remote driving, and dispatch monitoring background. They will be used for hospital isolation. Ward homework [14 ] . Until February 10, the long-awaited Baidu officially announced the free opening of low-speed miniature unmanned vehicle kits.
According to the information released by Baidu Apollo, Baidu only promised to provide technical adaptation, supply chain and cloud service support to the companies that meet the conditions during the epidemic, and did not provide free vehicles [15] ; Baidu’s partner companies provide low-speed micro-unmanned vehicles that are suitable for cleaning and disinfection, logistics delivery, and other scenarios. Nursing robots are quite similar and highly alternative.
Apollo Eco-Partner Smart Walker Cleaning and Disinfecting Vehicle | Baidu
Unmanned vehicles and robots play a similar role in the epidemic: we hope it will replace humans and do more work in epidemic areas to reduce the chance of infection by workers or susceptible people.
But where can these devices be used in practice? How many people work? Equipment maintenance and disinfection costs? Comprehensive evaluation of price-performance is also required. At the same time, we also need to assess whether there are risks that are not conducive to normal epidemic prevention.
As some communication cases claim, its products are “unmanned vehicles that can carry four people and two stretchers.” The question here is that ambulance isolation cockpit is a simpler and more feasible solution than unmanned driving; even if reducing one driver can indeed reduce the risk of infection, can unmanned driving break through red lights and whistle like an ambulance to open the road? Because of the “non-standardized” driving needs of special vehicles, at this stage, unmanned ambulances are impractical.
Of course, the tacit understanding of an industry is that certain driverless products are not really used in the frontline of the epidemic, and are only active in information articles. It is good to be able to make some contributions to the epidemic situation, but it is best not to be enthusiastic about medical treatment and epidemic prevention for publicity gimmicks.
Admit it, the “anti-epidemic” front line is not the main battlefield of Internet companies
The Internet industry does have its own advantages. They have both the ability to leverage 100 billion leverage in the capital market, and the spirit to research technology from scratch. But unlike the Silicon Valley giant Alphabet, which is trying to fill the technology tree, Chinese Internet companies have been able to obtain the so-called technology in the past 20 years, which is intended to serve the business model.
Whether it is a variety of AI temperature measurement or 5G “empowered” unmanned vehicles, what Internet technology can do in the epidemic is to use its inherent advantages to make articles on “chain relationships”-but And that’s it. They knocked on the “cut off the route of transmission of new coronary pneumonia”, and as more companies resume work and the virus latency increases, these so-called “high-end technologies” are becoming increasingly boring and useless.
On the other hand, in eradicating infectious diseases, we can’t expect Internet companies to come up with any hard-core technology. Their existing technology is even difficult to directly participate in the treatment of the disease. Even cloud computing can play a role in stopping New Coronavirus sequencing and predicting mutations. The role of Internet companies is still very limited in finding and developing specific drugs or broad-spectrum vaccines against the new crown virus.
Specialized in the surgery industry, Internet companies cannot be forced to possess the strength of a biomedical company. Most companies use the bottom of the technology in the outbreak to protect the people from illness. But goodwill ≠ correct, Goodwill ≠ effective. In many cases, these technologies are mostly useless on the front line. What’s more, many Internet companies have over-packed conventional technology, taking advantage of the outbreak of outrage.
Internet companies don’t have to look forward to themselves because they stick to the “technology company” label, they don’t need to be superior in all technical fields. In contrast, instead of brushing presence on the front battlefield, Internet companies that open up new battlefields in the rear and maintain people’s normal lives with strength are even more respectable.
Compared with SARS in 2003, the Internet today has a complete supply chain and logistics, a real-time information dissemination platform, a full set of online services and online office, and an online education system that Jingdong could not reach at the time. Asia’s No. 1 Intelligent Park has become one of the channels for Wuhan residents to purchase daily necessities; students have been badly criticized by students, and they have also become air classrooms in special times; Wuhan residents have inconvenience to travel and have support teams of Didi. Protecting emergency needs … These “techniques that can operate even if they cannot return to work or study again” are exactly the advantages that SARS did not have.
Can Internet companies improve the situation? Yes, it’s just slower.
Amara ’s Law states this: “We often overestimate the short-term impact of a technology, but underestimate its long-term impact.” Regardless of whether the outbreak was 17 years ago or not Achieved Ali and JD. One fact is that e-commerce did indeed flourish after 2003, and the digital industry also prospered at an unprecedented rate.
Now, the unexpected arrival of new coronary pneumonia has brought a new proposition to this nothing new Internet: Will the epidemic trigger a real Internet technology revolution instead of a superficial model prosperity?
References:
[1] [13] Leifeng.net: First Line | Robots in the quarantine area during the epidemic prevention and control war
[14] AI Global Headquarters: 5G unmanned vehicles, enter the” Wuhan epidemic area “! Awesome!
This article comes from WeChat public account: amplification lamp (ID: guokr233) , planning: the amplification lamp team, author: big sheep, drawing from the title: Baidu official micro