Student Brown Bag Seminar

Speaker:          Adrienne Kinney, Program in Applied Mathematics, University of Arizona

Title:               Assessing neural network models of mosquito abundance for vector surveillance

Abstract:         Vector-borne disease outbreaks are closely tied to vector abundance, which makes knowledge of population dynamics useful in preventing future outbreaks. Here we use the Aedes-AI framework we previously developed to produce probabilistic forecasts of mosquito abundance for neighborhoods in Puerto Rico.

The Aedes-AI models are a collection of neural network models of Aedes aegypti abundance [1]. The models are trained on synthetic data generated from a mechanistic model, in contrast to other models of mosquito abundance that rely on noisy, real world trap data for training. We previously demonstrated that the neural networks can learn the spatiotemporal features of mosquito populations.

In this work, we use the Aedes-AI models to generate predictions using local weather and present a methodology of scaling the predictions and forecasting mosquito abundance based on past trap data. We assess the ability of the forecasts to capture trends in future trap data and compare them with outbreaks of mosquito-borne diseases in the region. We conclude with a discussion on how the Aedes-AI models are appealing from a public health perspective and can be used to supplement vector surveillance efforts.

[1] Kinney, A.C., Current, S., and Lega, J., “Aedes-AI: Neural network models of mosquito abundance, PLoS Comp. Bio. (2021)