Combating COVID-19 with the Aid of Artificial Intelligence Technology
-
As of 16:00 on February 4th, the total number of confirmed cases of the novel coronavirus infection nationwide reached 20,502, with 23,214 suspected cases, 426 cumulative deaths, and 693 cumulative recoveries.
The rising number of confirmed cases has gripped the hearts of everyone, and all sectors of society have joined this silent battle against the epidemic. Beyond the tireless efforts of medical workers, artificial intelligence technology has also played a significant role in this fight—rapid temperature screening, big data monitoring, automated diagnosis, robotic reception... Undoubtedly, AI technology is gradually becoming humanity's new guardian.
1. Rapid Temperature Detection
With the overlap of the Spring Festival travel rush and the critical period of epidemic prevention, controlling the spread during this time is particularly crucial. Traditional handheld forehead or ear thermometers at transportation hubs like high-speed rail stations and airports are insufficient for handling dense crowds.
AI-powered image recognition combined with infrared thermal imaging technology enables rapid screening and early warning of forehead temperatures in crowded areas, overcoming challenges posed by facial recognition due to masks and hats. For instance, Baidu has implemented an AI-based multi-person temperature screening solution at Beijing Qinghe Railway Station, utilizing facial key-point detection and infrared temperature analysis algorithms for efficient crowd screening.
2. Big Data for Epidemic Alerts and Prevention
Cloud computing and big data technologies enable precise data collection and analysis, aiding governments in making informed decisions. For example, Canadian AI startup BlueDot used natural language processing to analyze foreign news reports and official announcements, issuing a warning on December 31, 2019, and accurately predicting the virus's spread from Wuhan to Bangkok, Seoul, Taipei, and Tokyo.
Big data also helps predict infection rates, optimize resource allocation, and track potential carriers. By analyzing travel histories, authorities can identify individuals who visited high-risk areas like the Huanan Seafood Market in Wuhan, enhancing targeted prevention efforts. Additionally, integrating data from maps, aviation, mobile communications, and e-commerce supports comprehensive modeling for epidemic response.
3. AI-Powered Diagnosis to Ease Doctor Workloads
Misconceptions about the epidemic have led many with minor symptoms like headaches or coughs to flood hospitals, increasing cross-infection risks and straining healthcare resources. While online consultations exist, limited doctor availability restricts their effectiveness.
To alleviate frontline pressure, AI-driven consultation robots, such as those developed by KuaiShangTong, assist doctors by handling routine inquiries, freeing them for critical tasks. These robots, now deployed in hospitals like Beijing Tongrentang and partnered with medical institutions, provide free support to healthcare providers and organizations.
4. Robots Minimize Contact Transmission
Robots reduce human-to-human contact, curbing virus spread. In treating COVID-19 cases, robots equipped with cameras, microphones, and stethoscopes have been used for remote patient care, as seen in the treatment of the first U.S. case.
Beyond treatment, robots perform non-contact tasks like delivering meals to isolation zones, preventing cross-infection. Innovations like contactless self-service terminals, developed by the First Affiliated Hospital of USTC, allow users to complete registrations and payments via mid-air gestures, eliminating surface contact.
5. AI Algorithms Identify Virus Hosts
Identifying the natural and intermediate hosts of the virus is crucial for cutting transmission routes. On January 25th, a team led by Professor Zhu Huaiqiu of Peking University published a study using a dual-path convolutional neural network (BiPathCNN) to predict the virus's hosts. Their analysis suggested bats as the natural reservoir and minks as a potential intermediate host.
6. AI Accelerates Antiviral Drug Development
Traditional drug discovery involves screening vast compound libraries for effective, low-toxicity candidates—a time-consuming and costly process. AI streamlines this by learning from known 3D protein structures and drug interactions, then autonomously screening compounds for potential efficacy. Researchers only need to validate a handful of AI-selected candidates, significantly reducing development time and costs.
Amid the outbreak, tech giants like BAT have donated substantial AI computing resources to expedite drug screening for scientists.