Oscar Lao: Natural computing applied to population genomics

Natural computing refers to a family of algorithms that try to mimic how natural processes are optimized. Natural algorithms include artificial neural networks, social behaviour and evolutionary strategies among others. These algorithms have been shown to be particularly powerful approaches for estimating the optimal -or sub-optimal- solution in NP-hard problems such as finding the shortest route visiting all the edges of a graph, identifying informative features in a dataset or identifying patterns in image recognition.

In my group we are applying natural computing algorithms to address basic problems in population genomic such as the demographic processes that shape the current observed genetic diversity at populations.

In this presentation I will introduce the Approximate Bayesian Computation algorithm coupled to Deep Learning that we have developed for demographic model ascertainment and parameter estimation.