Don’t let an outbreak come back and a DI model predicts or be a good idea

January 30 news, since the outbreak of the new coronavirus pneumonia outbreak, scientific and medical workers have come up with a number of ways to deal with, including artificial intelligence (AI) technology. Whenever a mysterious outbreak occurs suddenly, it may be difficult for government and public health officials to quickly gather relevant information and coordinate responses.

Don't let an outbreak come back and a DI model predicts or be a good idea

Infographic

But AI systems that automatically mine news stories and online content from around the world can help experts identify anomalies that could lead to a potential epidemic or worse. In other words, AI may actually help us survive the next plague.

The capabilities of these AI systems have been demonstrated in recent outbreaks of coronavirus pneumonia. The Canadian company, BlueDot, which claims to have discovered the coronavirus early on, is one of many companies that use data to assess public health risks. BlueDot, which claims to be mainly engaged in “automatic infectious disease surveillance,” informed customers of the new coronavirus at the end of December 2019.

It wasn’t until a few days later that the U.S. Centers for Disease Control and Prevention (CDC) and the World Health Organization (WHO) issued official notices. Now, towards the end of January, the coronavirus outbreak has claimed more than 100 lives, and several other countries, including the United States, have reported cases of infection.

Kamran Khan, an infectious disease physician and founder and chief executive of BlueDot, explained in an interview that the company’s early warning system uses AI, which includes natural language processing and machine learning, by analyzing about 100,000 articles a day in 65 languages. to track outbreaks of more than 100 infectious diseases. This data helps companies know when to notify customers of potential outbreaks and spreads of infectious diseases.

Other data, such as traveler’s itinerary information and flight paths, could help the company provide additional clues about how the disease may spread. Earlier this month, for example, blueDot researchers predicted that the coronavirus would appear in other Asian cities after it appeared in China.

The idea behind the BlueDot model, whose final results are then analyzed by human researchers, is to provide information to health care providers as soon as possible in the hope that they will be able to diagnose or isolate infected people and potential sources of infection at an early stage. “Official information is sometimes not timely, ” says Mr Kamran. The difference between an infection in travellers and an outbreak depends on whether front-line health care can identify a particular disease. This may be the key to preventing a real outbreak. “

Kamran added that the company’s AI system could also use a range of other data, such as information about climate, temperature and even local livestock in an area, to predict whether people infected with the disease could lead to outbreaks in the area. He noted that as early as 2016, BlueDot was able to predict an outbreak in the United States six months before the Zika virus actually appeared in Florida.

Similarly, Metabiota, an outbreak monitoring company, claims that countries such as Thailand, South Korea and Japan have the highest risk of outbreaks of coronavirus pneumonia, which occur seven days earlier than officially reported, in part by looking at flight data. Metabiota, like BlueDot, uses natural language processing to evaluate online reports of potential diseases, and it is developing the same technology for social media data.

Online platforms and forums can also give signs of an outbreak, explains Mark Gallivan, metabiota’s director of data science. Metabiota also claims that it can estimate the risk of social and political disruption caused by the spread of the disease based on information such as the symptoms of the disease, mortality and accesstos to treatment. At the time of publication, for example, Metabiota rated the risk of public anxiety caused by the new coronavirus in the United States and China as “advanced,” but it rated the risk of monkeypox virus in the Democratic Republic of the Congo as “moderate.”

It’s hard to know how accurate the rating system or platform itself can be, but Mr. Galvan said the company is working with the U.S. intelligence community and the Defense Department to address coronavirus-related issues. Metabiota also advertises its platform to reinsurers, which may want to manage financial risks associated with the potential spread of the disease.

But AI’s ability is far from just to allow epidemiologists and government officials to be notified when disease son. Researchers have developed AI-based models that can predict outbreaks of the Zika virus in real time, which could inform doctors how to respond to potential crises. AI can also be used to guide public health officials on how to allocate resources during a crisis. In fact, AI will be the new line of defense against disease.

More broadly, AI is already helping to research new drugs, treat rare diseases, and detect breast cancer. AI has even been used to identify whether insects that transmit Chagas disease, which has infected about 8 million people in Mexico and Central and South America. There is also growing interest in using unhealthy data, such as social media posts, to help health policy makers and pharmaceutical companies understand the breadth of the health crisis. For example, AI can mine social media posts to track the sale of illegal opioids and keep public health officials informed of the spread of these controlled substances.

Of course, these AI systems, including those of Metabiota and BlueDot, can only be used to evaluate situations similar to a given data. In general, AI has a bias problem, either from the engineer who built the system or from the data it trains. And the use of AI in health care is by no matter how much it is.

The idea of AI fighting deadly diseases provides us with new use cases that, even if they don’t give us all hope, don’t make us feel very upset. If developed and used properly, perhaps this technology could indeed help save more lives. (From: Vox Author: Rebecca Heilweil Compilation: Netease Intelligence Participation: Small)