Facilitating the Spread of Knowledge and Innovation in Professional Software Development

Write for InfoQ


Choose your language

InfoQ Homepage News COVID-19 and AI: Virtual Conference at Stanford Discusses the Future

COVID-19 and AI: Virtual Conference at Stanford Discusses the Future

This item in japanese

A recent virtual conference was rapidly organized to discuss the use of data analysis to understand and fight COVID-19. Following the growing number of COVID-19 cases and in accordance with the school's policy, the original in-person conference was cancelled and replaced with a virtual conference titled "COVID-19 and AI: a virtual conference". The conference was split into four primary sessions, which were named "Landscape and Framing", "Social Impacts and Biosecurity", "Tracking the Epidemic" and "Treatments and Vaccines". Notable speakers included congressman Dr. Ami Bera, the U.S. Representative of the Seventh Congressional District, and Anthony Goldbloom, founder and CEO of Kaggle, a competitive platform for data scientists and machine learning researchers.

On the machine learning and data science end, Nigam Shah, associate professor of Medicine (Biomedical Informatics) and of Biomedical Data Science discussed the efforts of 60+ data scientists to see what they could do about the pandemic. Shah discussed the three areas of focus for data scientists: operational planning, like how many beds a hospital might need; clinical care decisions like which patients should get tested; and research questions like which drugs have the highest efficacy at combating the virus. There are two kinds of models, according to Shah, SEIR simulations that capture dynamics of an epidemic and simple calculators that capture day-to-day trends. SEIR simulators require extremely accurate inputs, whereas simple calculators are easier estimates with low-dimensional inputs.

In Session III, Jason Wang, associate professor of Pediatrics at Lucile Packard Children’s Hospital, discussed Taiwan's big-data dependent response towards the COVID-19 epidemic.  Using surveys collected immediately after arriving in Taiwan, customs separated people into high risk and low risk categories. Linking the National Health Insurance Research database to the Customs and Immigration database, doctors could be informed in advance of potential high risk patients at point of care. Early action mitigated the potential risk associated with the Diamond Princess cruise ship, which stopped in Taiwan before Japan: the Taiwan government sent text alerts to citizens who had been in areas where tourists from the Diamond Princess had visited to alert them of possible risk.

Later in Session III, John Brownstein, chief innovation officer at Boston Children's Hospital and professor at Harvard University, briefly discussed HealthMap, a data-mining tool for WeChat, local news and media outlets, to map the rise of COVID-19 cases in China. Using a broad collaboration across the University of Washington, Oxford and other universities, they have been line-mapping every case they can. Brownstein and others have used publicly available dataset to assess the efficacy of human mobility and control measures on the COVID-19 epidemic in China. Other tools like Thermia, to measure temperature, were deployed in China, and Buoy, an NLP-powered symptom checker, was being deployed in Massachusetts.

Sponsored by the Wu Tsai Neurosciences Institute and the Stanford Department of Psychology, the Spring 2020 conference on Human-Centered AI originally intended on discussing the intersection of psychology, neuroscience and artificial intelligence.

Rate this Article