John Baez | Category Theory in Epidemiology

African Mathematics Seminar | November 2, 2022

Virtually hosted by the University of Nairobi


Sponsor: International Science Programme

[fredeisele]

I think the idea you are making by composing using the pushout to R with R' (and S with S') would be better served if you used letters from different languages.
That is (for example) Re for Recovered and Im for Imepona, Su for Susceptible and In for Inashukiwa.
Then your pushout would effectively fuse Re with Im and Su with In.
This conveys the idea that the models being composed come from distinct domains (or cultures).
I used some Swahili words but you could chose words particular to your audience.

[dwinsemius]

You asked for questions around @9:00 and if I had been in the audience I would have asked why the dR/dt value did nod depend on both I and R which was has how the dS/dt value was presented (as dS/dt=-phi_i(S, I). I guess it's because you have two absorbing states and S is just a "donor state"? You then answerred the question how to go from the diagram to the system of Eqs but didn't really answer the question: Why no blue arrow from R to the I-R flow? (When I learned an analogous description the authors talked about "through variables" (like your double arrows) which were things like masses or quantities and "across variables" (like your blue arrows)which were potentials that could be used to create "forces".

[SalixAlba]

With the idea of composition of diagrams, presumably we can do decompose a diagram into a composition of simpler diagram. Do we end up with some sort of irreducible set?