Over 82 million people have been infected worldwide, and the number of new COVID-19 cases has continued to climb in recent months. As we anxiously await vaccines, artificial intelligence is already battling the virus on a number of fronts — from predicting protein structure and diagnosing patients to automatically disinfecting public areas.
As part of our year-end series, Synced highlights 10 AI-powered efforts that contributed to the fight against COVID-19 in 2020.
To help the global research community better understand the coronavirus, UK-based AI company and research lab DeepMind in March leveraged their AlphaFold system to releasestructure predictions for six proteins associated with SARS-CoV-2, the virus that causes COVID-19. In August, DeepMind released additional SARS-CoV-2 structure predictions for five understudied SARS-CoV-2 targets.
AlphaFold was introduced in December 2018. The deep learning system is designed to accurately predict protein structure even when no structures of similar proteins are available, and can generate 3D models of proteins with SOTA accuracy. On November 30, the latest version of AlphaFold was recognized for solving the biennial Critical Assessment of Protein Structure Prediction (CASP) grand challenge with unparalleled levels of accuracy. DeepMind says AlphaFold’s success “demonstrates the impact AI can have on scientific discovery and its potential to dramatically accelerate progress in some of the most fundamental fields that explain and shape our world.”
Diagnosing COVID-19 Infection in Seconds
CT (computed tomography) lung scans and nucleic acid tests are the two main diagnostic tools doctors use in confirming COVID-19 infections, and CT imaging is crucial for lung infection diagnosis verification and severity assessment.
In January, the Shanghai Public Health Clinical Center (SPHCC) partnered with Chinese AI startup YITU Technology’s healthcare division — which provides AI-powered medical imaging solutions for lung cancer diagnosis — to build an AI CT image reader.
By February 5, the AI diagnostic system had been deployed in four Hubei province hospitals struggling with ongoing doctor and supply shortages. Functionalities catering to the specific needs of clinical departments such as radiology, respiratory, emergency, intensive care, etc., were built into the system.
In response to the COVID-19 pandemic, the US White House joined with research groups in March to announce the release of the COVID-19 Open Research Dataset (CORD-19) of scholarly literature about COVID-19, SARS-CoV-2, and the coronavirus group. The release came with an urgent call to action to the world’s AI experts to “develop new text and data mining techniques that can help the science community answer high-priority scientific questions related to COVID-19.”
The online ML community Kaggle is hosting a CORD-19 dataset challenge that defines 10 tasks based on key scientific questions developed in coordination with the WHO and the National Academies of Sciences, Engineering, and Medicine’s Standing Committee on Emerging Infectious Diseases and 21st Century Health Threats.
Developed by the Pande Laboratory at Stanford University in 2000 as a distributed computing project for simulating protein dynamics — including the process of protein folding and the movements of protein implicated in a variety of diseases — the Folding@Home project aims to build a network of protein dynamics simulations run on volunteers’ personal computers to provide insights that could help researchers develop new therapeutics.
The current focus of Folding@Home is modelling the structure of the 2019-nCoV spike protein to identify sites that can be targeted by therapeutic antibodies. Coronaviruses invade cells via spike protein on their surfaces, which binds to a lung cell’s receptor protein. Understanding the structure of viral spike protein and how it binds to the ACE-2 human host cell receptor can help scientists stop viral entry into human cells. Anyone who would like to donate their unused computing power can join Folding@Home’s fight against the coronavirus.
In March, Canadian startup DarwinAI released COVID-Net, an open-sourced neural network for COVID-19 detection using chest radiography (X-Rays). Company CEO Sheldon Fernandez says COVID-Net has been leveraged by researchers in Italy, Canada, Spain, Malaysia, India and the US.
Fernandez explains that rather than treating AI as a tool, his company reimagines AI as a collaborator that learns from a developer’s needs and subsequently proposes multiple design approaches with different trade-offs in order to enable a rapid and iterative approach to model building.
In response to the COVID-19 pandemic, Andrej Karpathy — director of artificial intelligence and Autopilot Vision at Tesla and developer of the arXiv sanity preserver web interface — introduced “Covid-Sanity,” a web interface designed to navigate the flood of bioRxiv and medRxiv COVID-19 papers and make the research within more searchable and sortable.
Covid-Sanity organizes COVID-19-related papers with a “most similar” search that uses an exemplar SVM trained on TF-IDF feature vectors from the abstracts of the papers. This is similar to the Google search engine, which responds by finding the relevance of a query in all texts, ranks by similarity scores and returns the top-k results. Based on paper abstracts, the web interface returns all papers similar to the best-matched paper result to a query.
Volunteer Drone Teams for COVID-19 Disinfection
Disinfection of public areas is a challenging but crucial process in the fight to stop the spread of the COVID-19. In China, ad hoc teams of DJI drone hobbyists sprung up nationwide to provide this service for free. By February, total DJI agricultural drone disinfection coverage had exceeded 600 million square meters across more than 1,000 villages — including schools, isolation wards, food waste treatment plants, waste incineration plants, livestock and poultry epidemic prevention centres and more.
Shenzhen-based DJI is a leading drone and associated technologies company. In February they launched the “DJI Army Against the Virus” project, providing subsidies to support working pilots, with provisions for pilot protective kits and assistance to villages who perform drone disinfection. Spare parts and drone repair services were also provided during the missions.
Autonomous Delivery Vehicles Navigate the Pandemic
Many AI-powered autonomous vehicles navigated Chinese streets in response to the COVID-19 outbreak. Developed and modified for the purpose by Chinese O2O local life service company Meituan, Modai (“Magic Bag”) vehicles delivered much-needed groceries to communities in Beijing’s Shunyi District.
Self-driving delivery vehicles like Modai were an effective solution to the COVID-triggered surge of online grocery orders and the need to reduce interpersonal contact to slow disease spread. The urgent needs and empty streets drew many companies into autonomous delivery — JD Logistics developed self-driving delivery vehicles in Wuhan and for the first time delivered medical supplies to the Wuhan Ninth Hospital, and the Suning Logistics 5G Wolong self-driving car delivered its first orders in Suzhou.
Hand Washing AI
Japan’s Fujitsu Ltd developed an artificial intelligence monitor to ensure healthcare, hotel, and food industry workers wash their hands properly, according to a Reuters report. The system is based on crime surveillance technology that detects suspicious body movements, and can recognize and classify complex hand movements. It checks whether people complete a Japanese health ministry six-step hand washing procedure similar to guidelines issued by the WHO (clean palms, wash thumbs, between fingers and around wrists and scrub fingernails). The monitor can even tag instances of people not using soap.
AI-Assisted Elder Care Solution
Fei-Fei Li, Stanford computer science professor and co-director of Stanford’s Human-Centered AI Institute (HAI), shared her thoughts on AI technologies that could help seniors during the coronavirus pandemic in April’s COVID-19 and AI: A Virtual Conference. Li identified AI-powered smart home sensor technology as a way to help families and clinicians remotely monitor housebound seniors for infection symptoms or symptom progression or regression and potentially also help manage their chronic health issues.
Research institution Strategy Analytics predicts the smart home market will resume in 2021 and consumer spending will increase to US$62 billion. The post-pandemic global smart home device market is expected to maintain a compound annual growth rate of 15 percent.
As efforts to control the spread of COVID-19 continue, contact tracing has emerged as a public health tool where ML can play an important role in optimizing systems. Various countries have developed digital contact tracing processes with mobile applications, utilizing technologies like Bluetooth, the Global Positioning System (GPS), social graphs, network-based API, mobile tracking data, system physical addresses, etc. These apps collect massive data from individuals, which ML and AI tools analyze to identify and trace vulnerable people.
A study published by the US National Library of Medicine shows that by June, over 36 countries had successfully employed digital contact tracing systems using a mixture of ML and other techniques.
Reporter: Yuan Yuan | Editor: Michael Sarazen
This report offers a look at how China has leveraged artificial intelligence technologies in the battle against COVID-19. It is also available on Amazon Kindle. Along with this report, we also introduced a database covering additional 1428 artificial intelligence solutions from 12 pandemic scenarios.
Click here to find more reports from us.
We know you don’t want to miss any news or research breakthroughs. Subscribe to our popular newsletter Synced Global AI Weekly to get weekly AI updates.