Andy and Dave, co-hosts of AI with AI, CNA’s popular podcast on artificial intelligence, have compiled a second annotated list of AI developments and resources related to COVID-19. Entries are cross-referenced with the podcast episodes in which Dave and Andy discussed them. Part 1 of this resource list was published on May 6.

qXR: A validated AI-based chest x-ray system

qXR is an AI-based chest x-ray system that uses a combination of deep-learning models to detect common types of lung abnormalities. Mumbai-based Qure.ai started developing this system before the outbreak. In September 2019, Rizwan Malik, lead radiologist at the Royal Bolton Hospital in England identified qXR as the most promising system for a clinical trial. After the pandemic hit, the proposal was quickly adjusted, and a preliminary validation study using more than 11,000 patient images found the tool was able to distinguish between COVID-19 and non-COVID-19 patients with 95% accuracy.

Discussed on "AI with AI" Podcast 3.27

https://www.technologyreview.com/2020/05/07/1000961/launching-mittr-covid-tracing-tracker/

Ethical guidelines for SARS-CoV-2 digital tracking and tracing systems

The Digital Ethics Lab at the University of Oxford released a framework to evaluate whether the use of digital systems to track potentially infected individuals is legal and ethical. The framework consists of high-level principles and enabling factors. High-level principles include necessity, proportionality scientific soundness and time-boundedness. The enabling factors are concrete considerations, derived from high-level governance principles and digital ethics. They enable one to answer the question "Is the digital tracking and tracing system being developed correctly?" (This is what engineers would call "verification" of the product.) The enabling factors translate high-level principles (the what) into practical, ground-level considerations for designers and deployers of tracking and tracing systems (the how).

Discussed on "AI with AI" Podcast 3.27

https://papers.ssrn.com/sol3/papers.cfm?abstract_id=3582550

Carnegie Mellon Interactive COVID-19 Maps

Carnegie Mellon University unveiled five interactive maps on its COVIDcast website displaying real-time information on symptoms, doctor visits, medical tests and browser searches related to COVID-19 in the United States, which is used to estimate disease activity at the county level. COVIDcast was created by Carnegie Mellon's Delphi Research Group and its COVID-19 Response Team, with support from Google, Facebook, Quidel Corporation and a national health system.

Discussed on "AI with AI" Podcast 3.27

https://covidcast.cmu.edu/

Facebook & Carnegie Mellon University COVID-19 Symptom Map:

https://covid-survey.dataforgood.fb.com/

The National Security Commission on Artificial Intelligence white papers

"Privacy and Ethics Recommendations for Computing Applications Developed to Mitigate COVID-19"

This paper offers recommendations for putting civil liberties at the center of contact tracing methods, ensuring that federally funded AI tools used in pandemic response account for potential bias and avoid introducing additional unfairness into healthcare delivery and outcomes.

https://drive.google.com/file/d/1m0AT21dS2XJ6JIGMgo7SuLSLveWIO8WK/view

"Mitigating Economic Impacts of the COVID-19 Pandemic and Preserving U.S. Strategic Competitiveness in AI"

This paper offers several steps policymakers can take to use and protect U.S. capabilities in AI and associated technologies during the current crisis and beyond.

"AI presents opportunities to help safely reopen and grow the US economy while minimizing the risk of future outbreaks, but the AI ecosystem itself is one of the most critical sectors in need of assistance and protection from foreign competitors as it weathers the brutal economic crisis."

Discussed on "AI with AI" Podcast 3.28 and 3.21

https://drive.google.com/file/d/1vSRRfqV6S4xGoMxueC-CMG0tYYTGa1Ly/view

Mini-model ensemble approach to predicting possible trajectories of COVID-19’s death toll

FiveThirtyEight,with the help of the Reich Lab at the University of Massachusetts, Amherst, has released a new tool that compares six of the most reputable models side by side and explains the differences in their predictions and assumptions. The models’ forecasts are updated every week along with actual coronavirus data, showing how the models fared in the past. “Ensemble” forecasts are useful because they help us understand the most likely outcomes as well as best- and worst-case scenarios.

Discussed on "AI with AI" Podcast 3.29

https://projects.fivethirtyeight.com/covid-forecasts/

MIT Technology Review releases COVID Tracing Tracker

The COVID Tracing Tracker is a database to capture details of every significant automated contact tracing effort around the world. The site has documented 25 significant automated contact tracing efforts globally, including details on what they are, how they work, and what policies and processes have been put in place around them.

https://www.technologyreview.com/2020/05/07/1000961/launching-mittr-covid-tracing-tracker/

Discussed on "AI with AI" Podcast 3.29

John Hopkins: 2019 Novel Coronavirus Research Compendium

Kate Grabowski, an infectious disease epidemiologist at Johns Hopkins University's Bloomberg School of Public Health, is leading an effort to create a curated set of pandemic papers. The 2019 Novel Coronavirus Research Compendium focuses on quality over quantity. It complements OpenAI's CORD-19 effort. The Johns Hopkins effort focuses on studies in humans, and the intended readers are primarily health care workers and policymakers, as well as researchers.

Discussed on "AI with AI" Podcast 3.30

https://ncrc.jhsph.edu/

NLP-powered COVID-19 Primer

The COVID-19 Primer summarizes research trends as well as the news coverage and social media discussion on scientific research about the COVID-19 pandemic and the SARS-CoV-2 virus. It employs natural language processing (NLP) to read and analyze research papers, identify trending discussions, and generate summaries. The COVID-19 Primer is generated by Primer AI, a machine intelligence company in San Francisco.

Discussed on "AI with AI" Podcast 3.30

https://covid19primer.com/dashboard/

C3.ai COVID-19 Data Lake

C3.ai — an enterprise AI software provider — has created a "data lake" of information about the COVID-19 pandemic for researchers around the world. The data lake is a unique, unified source of comprehensive COVID-19 data and a knowledge graph that is structurally different from other COVID-19 data collections, enabling researchers to accelerate efforts to mitigate the spread of the pandemic.

Discussed on "AI with AI" Podcast 3.30

C3.ai homepage (access port): https://c3.ai/products/c3-ai-covid-19-data-lake/

AI-enabled rapid diagnosis of patients with COVID-19

Mount Sinai researchers in New York built an AI system using data and images to rapidly diagnose patients with COVID-19, according to a study published in Nature on May 19, 2020. The researchers believe they are the first in the U.S. to develop such a system, which uses an algorithm trained on more than 900 CT scans from Chinese hospitals. The algorithm detects the disease based on CT scans of a patient's lungs, along with data about age, bloodwork, symptoms, and possible contact with virus.

Discussed on "AI with AI" Podcast 3.31

https://www.nature.com/articles/s41591-020-0931-3.pdf

MIT-IBM Watson AI Lab to fund 10 AI research projects to help fight COVID-19

The MIT-IBM Watson AI Lab announced on May 19 that it will fund ten MIT research projects incorporating AI and other advanced technology to fight the coronavirus:

  • Early detection of sepsis in COVID-19 patients.
  • Designing proteins to block SARS-CoV-2.
  • Saving lives while restarting the U.S. economy.
  • Which materials make the best face masks?
  • Treating COVID-19 with repurposed drugs.
  • A privacy-first approach to automated contact tracing.
  • Overcoming manufacturing and supply hurdles to provide global access to a coronavirus vaccine.
  • Finding better ways to treat COVID-19 patients on ventilators.
  • Leveraging electronic medical records to find a treatment for COVID-19.
  • Returning to normal via targeted lockdowns, personalized treatments and mass testing.

Discussed on "AI with AI" Podcast 3.31

http://news.mit.edu/2020/mit-marshaling-artificial-intelligence-fight-against-covid-19-0519

Virtual Conferences

AI for COVID-19 Response: Federal Health Virtual Forum

Leaders from the federal government, Booz Allen and select industry and academic partners provide a candid look at how AI is already being applied and opportunities to leverage it further. Topics include:

  • How agencies and industry experts are leveraging AI for COVID-19 response.
  • Best practices for data management workflows.
  • How to identify and mitigate common challenges.
  • The importance of collaboration to overcome challenges related to COVID-19.

On-demand video including slides — 1.5 hrs. Requires name, email, company affiliation to get link.

https://boozallen.com/s/event/federal-health-ai-webinar.html?origref=&vURL=/Covid19AIEvent

AI for COVID-19 in LA: A Virtual Symposium

A virtual symposium hosted on May 8 by University of Southern California’s Viterbi School of Engineering, Center for AI in Society and ML Center.

Video proceedings (5.5 hr):https://sites.google.com/usc.edu/ai4covid-19/video

Presentation slides:https://sites.google.com/usc.edu/ai4covid-19/presentations

Andrew Ilachinski specializes in the mathematical and computer modeling of complex adaptive systems. He pioneered the application of agent-based modeling (ABM), and evolutionary programming techniques to military operations research problems, and developed two of the earliest ABMs of land warfare.

David Broyles directs the Special Activities & Intelligence Program. Broyles specializes in cyber operations and special operations, as well as experimentation and innovation in the Department of Defense.