Automated Conjecturing & Bootstrap Percolation

A network is being infected by a very mathematical virus -- any susceptible node becomes infected if it is connected to at least two already infected nodes.  This process is known as 2-bootstrap percolation, and it has its origins in statistical physics.  In the Summer of 2019, with a large group ranging from high school students to full professors, we used Automated Conjecturing software to discover which networks are MOST susceptible.  In this talk, we give a history of this model and a few of its applications, discuss the Automated  Conjecturing software we used, and see some of the results obtained during and since this working group.  Further, we'll spend a little time discussing the somewhat unusual (and budget-free) large lab environment in which this work was initiated.
This talk is intended for a general audience, and assumes no particular mathematical background; the software used is open-source and publicly available.