Fighting The Pandemic With Data — AI + ML Approach. Insights from GST Webinar
Last week we held a webinar panel discussion with leaders from RoundSqr, a digital/data consulting company to discuss their work in utilising data to fight the pandemic. Their work began with three phases — understanding the disease, making the transition and finally gearing up for the new world.
They used Xrays to at first distinguish regular pneumonia from Covid pneumonia, then moved on to utilising Xray data as a composite to predict the progression of the disease. They learned that utilising their Xray model for early diagnosis was not effective but as a predictive model, it was able to map out the progression of the disease in a patient.
Some additional insights from their work regarding the future of the business world:
- Workplaces of the future. Things cannot be as reactive as they are now. We must apply ideas or methods used in one industry to other industries. For example, the retail industry studies the customer patterns when they are in a store so they can stock their items in places where the customer flocks more often.
- We can use the same frequent foot analysis in post-covid scenarios. You can then proactively prioritize and clean/disinfect these areas more often than you would if there were set areas to clean (new-age facilities management).
- You need to change things once you know the percentage of people that are returning to the office: which spaces can be optimized, what will be the temperature of the aircon, which elevators will be working and the wait time between elevators. Everything will have to be predicted and made more efficient.
- Companies can’t run facilities at the same cost as they were before with 100% occupancy.
- We will now transition into more facial recognition capabilities to avoid contact with biometrics facilities.
- You would also want to track and trace people who visited your premises, even for full-time employees.
- Work contractors who visit will be easily tracked in their movements through CCTVs, low-cost beacons. People will be traced and tracked more. Remote monitoring will be more normal
- We will reach a point when the contact tracing apps will be able to warn us if we came into contact with someone with a high risk for infection.
- You can also use CCTV more proactively: it can help determine if there are large gatherings in an area, and you can disperse the crowd immediately.
Problems:
- Can we predict the progression or regression of the disease so healthcare workers can determine which patients are more at risk? Also to help healthcare workers in putting a treatment plan in place? How can these available data help healthcare workers become more proactive?
- How can we make sense of all the data lying around about coronaviruses? Can I segment them into different topics? Will I able to answer questions if I search for them? Which one will be more helpful? How can you cluster the documents according to topics?
- To get out of the lockdowns, one of the key metrics is to understand the R0 or the RT metric. How many other people will each infected person infect?
- As an employer, what can you do more diligently so your employees feel safe to go back to work? How can we become more risk-aware as employers?
- Can we do an objective triage? Can you use an app, a health kiosk, a digital low range bluetooth medical device to collect information with a no-touch policy? Can you feed that information automatically?
- Does it make sense to add some mental health triage questions?
- How can you optimize the working space that you’ve got?
Please check out the slides here https://s3.amazonaws.com/globalsouth.net/Presentation_For_GST_Webinar+v2.pdf
