In this project Orit Halpern and Johannes Bruder use the same mathematical approach to model a population of individuals and their response to an epidemic’ (Friston et al. 2020a). These models made neural nets proxies for human behavior in order to model disease spread during the pandemic.
The COVID 19 pandemic has seemingly naturalized the relationship between computation and human survival. Digital systems, at least in the Global North, sustain our supply chains, labor, vaccine development, public health, and virtually every manner of social life. Nowhere has this link become more powerful then at the intersection of statistics, artificial intelligence and disease modelling.
Early in the pandemic hopes were high that tracing apps, mobility data and statistical models might give the predictions and models for human behavior and social interactions to stop the progress of COVID-19. All over the world, scientists gathered to pool their knowledge in the modelling of complex processes and provide guidelines for measures to contain the spread of the virus. For instance, a group of like-minded scientists formed the Independent Scientific Advisory Group for Emergenics (IndieSAGE), an organization providing independent scientific advice to the UK government and public on how to minimize deaths and support Britain’s recovery from the COVID-19 crisis.1 Somewhat surprisingly, the group did not simply rely on epidemiological models but used one of their member’s hard-won knowledge about interactions of neurons in the human brain to infer on how social interaction could influence the spread of the virus in the UK (Friston et al. 2020a, Friston et al. 2020b). ‘Here, we use a ubiquitous form of model’, the authors write, ‘namely a mean field approximation to loosely coupled ensembles or populations. In the neurosciences, this kind of model is applied to populations of neurons that respond to experimental stimulation… Here we use the same mathematical approach to model a population of individuals and their response to an epidemic’ (Friston et al. 2020a). These models made neural nets proxies for human behavior in order to model disease spread during the pandemic.