Behind the battle against the epidemic, humans use AI technology to stand firm together

On May 18, humanity’s fight against the novel coronavirus reached a milestone. Moderna, the first company to begin human clinical trials of a vaccine, announced early results: the vaccine produced an immune response that protected volunteers from viral infection. Next, the biotech company will continue more experiments, hoping to have a vaccine for emergency use by the fall. [1] Previously, Moderna developed the vaccine in just 42 days, thanks to artificial intelligence technology: selecting the best messenger RNA, using AI to complete real-time quality control, and improving clinical work efficiency all rely on AI. [2]

There is no doubt that vaccines will be the most effective line of defense for mankind. But before a widely available vaccine appears, we still need more other means of epidemic prevention, such as detection and diagnosis, prevention and control screening, epidemic analysis, transmission location, etc. All of these will build multiple lines of defense to protect people's lives and safety. In these measures, AI also plays a vital role.

Intel, Huiyi Huiying, and 4Paradigm use artificial intelligence technology to consolidate the above-mentioned lines of defense and also promote the continuous transformation of medical digitalization.

Huiyi Huiying was founded in 2015. It is a startup company that helps doctors use deep learning technology to read medical images. By applying deep learning technology to medical images, it provides hospitals with medical diagnostic software products, including AI diagnostic cloud platforms, radiomics cloud platforms, etc., to assist doctors in the detection and treatment of lung nodules, fractures, etc., and provide assistance to doctors in scientific research.

In the early stage of the epidemic, due to the lack of knowledge about the virus and effective detection methods, many hospitals and doctors screened patients based on chest CT images to determine whether they were infected. However, with the increasing number of patients, hundreds of CT films also overwhelmed doctors. In the race against the virus, this will obviously greatly reduce the efficiency of medical treatment and delay the best time for diagnosis and treatment.

(Photo source: Huiyi Huiying)

This reflects the AI ​​technology advantage of Huiyi Huiying. Through cooperation with many hospitals, Huiyi Huiying quickly collected CT images of nearly 4,000 COVID-19 patients, and combined with more than 13,000 pneumonia images of non-COVID-19 patients, it quickly developed an artificial intelligence system - by analyzing indicators such as ground-glass opacity (GGO) in CT images, it provides doctors and experts with the probability of suspected infection for further testing by doctors and experts.

According to Chai Xiangfei, founder and CEO of Huiyi Huiying, " Based on Intel's computing power and optimization platform, and combined with artificial intelligence neural network models , Huiyi Huiying's solution can help doctors make diagnoses faster and more accurately, preventing infected people from spreading the virus to others. At the same time, it provides doctors with quantitative information on lesions and brings more accurate treatment plans to patients."

Huiyi Huiying COVID-19 AI-assisted screening system

(Photo source: Huiyi Huiying)

Practice has proved that in the early stage of the fight against the epidemic, when RT-PCR (reverse transcription-polymerase chain reaction) was not yet popular, the diagnostic solution provided by Huiyi Huiying played a positive auxiliary role and became an effective supplement to standard laboratory testing. Once this technology came out, it was quickly promoted to more than 20 hospitals on the front line of the fight against the epidemic in China, greatly alleviating the work pressure of frontline doctors. With the spread of the epidemic, this technology has now been deployed and used in more than 40 countries overseas, allowing more medical staff and patients to benefit from it.

In recent days, the news of the outbreak in Shulan City, Jilin Province has touched the nerves of people across the country and made everyone further realize the importance of epidemic prevention and control. We often say "prevention is better than cure". On the one hand, we need to actively treat COVID-19 patients, and on the other hand, we need to manage the suspected population from the source and from the perspective of mobility. At this time, in addition to the big data of population mobility, we also need an intelligent analysis platform.

4Paradigm is such an artificial intelligence platform company that provides intelligent AI services. 4Paradigm has always maintained a close joint technology innovation relationship with Intel. At the first moment of the outbreak, 4Paradigm formed a joint team with the LAMDA Institute of Nanjing University and clinical experts from a hospital in northern Jiangsu to carry out technical research for key links in the epidemic prevention and control process of relevant departments.

It is reported that the solutions provided by 4Paradigm for this epidemic prevention and control mainly include three parts: high-dimensional machine learning prevention and control screening model, provincial, municipal, district and county digital twin system and new coronavirus transmission path analysis system:

By leveraging AI-based self-learning capabilities and the professional experience of clinical research experts, we build a closed loop of data and systems and conduct continuous iterations, ensuring that efficient and accurate judgment support is provided to relevant departments within a limited time based on dynamic environmental changes (including virus mutations and population dynamic complexity), and that high-risk susceptible populations are accurately screened to achieve precise prevention and control.

Based on high-dimensional machine learning technology and multi-dimensional data, a digital twin system with finer granularity and closer to the actual situation is built. It can fully consider the impact of various unexpected factors, such as traffic control, resumption of work time and drug delivery, on the epidemic, and analyze the development of the epidemic in real time through simulation function rehearsal, thereby providing a basis for the formulation and adjustment of prevention and control policies;

We use machine learning technology to build a data-driven digital twin system for the spread of the new coronavirus. Based on this, we build a potential infection network. Combined with patient information, we find possible transmission paths and potential super spreaders in the network, quickly analyze the transmission paths, and help epidemic prevention departments cut off the source of the spread of the epidemic as soon as possible.

According to Tu Weiwei, chief scientist at 4Paradigm, this innovative solution further enriches the existing prevention and control screening model, and the screening accuracy of potential spreaders has increased from the original 5.8% to more than 90% . Practice has proved that this innovative solution launched by the joint team performs better than the traditional epidemic transmission model. It allows AI to play an irreplaceable role in epidemic prevention and control, and has made important contributions to epidemic prevention and control.

In the face of the new coronavirus, technology, including AI, is a powerful weapon in our hands. Humans have used it to do things that were unimaginable in the past. In this special battle, Intel has joined hands with many industry partners to provide strong support for people: from assisting in vaccine development, manufacturing masks and ventilators, to helping people work remotely, Intel technology is present.