Data-informed parameter synthesis for stochastic population models
Tatjana Petrov, University of Konstanz
Tatjana Petrov will speak at the Summer Seminar Series on "Data-informed parameter synthesis for stochastic population models"
Stochastic population models are widely used to model biological phenomena such as molecular interactions or collective animal behaviour. Quantitative analysis of stochastic population models easily becomes challenging, due to the combinatorial propagation of dependencies across the population. The complexity becomes especially prominent when model's parameters are not known and available measurements are limited. In the talk, I will illustrate this challenge in a concrete scenario: we assume a simple communication scheme among identical individuals, inspired by how social honeybees emit the alarm pheromone to protect the colony in case of danger. Together, n individuals induce a population Markov chain with n parameters. We assume to be able to experimentally observe the states only after the steady-state is reached. We obtain the parameters of the individual's behaviour, by utilising the data measurements for population, in two steps. First, we use the tools for parameter synthesis for Markov chains with respect to temporal logic properties, to get a symbolic representation of steady-state distribution. Then, we employ a counterexample-guided abstraction refinement (CEGAR) reasoning with SMT solvers, in order to find the viable parameter space up to desired coverage. In the end of the talk, I will discuss the algorithm performance on a number of synthetic data sets, and outlook to the future challenges. This is joint work with Matej Hajnal, Morgane Nouvian and David Safranek.