Continuous-time methods in Macroeconomics with Applications to Machine Learning Summer School at The University of Oxford
Continuous-time methods draw from the vast body of mathematical research on partial differential equations and provide some advantages over more frequently used discrete time methods. I participated in the summer school on continuous-time methods with applications to machine learning taught by Jesús Fernández-Villaverde from the University of Pennsylvania. The advantages of continuous time methods are especially important when we are dealing with models with uncertainty, as the application of Ito's Lemma allows us to substitute the integrals of discrete time for derivatives in continuous time and lead to considerable gains in terms of computational speed. Although continuous-time methods have a long and illustrious tradition in finance, only recently did they start being used widely in business cycle research. Jesús Fernández-Villaverde highlighted that they can be employed together with machine learning tools and applied very successfully to complex models such as those that I am using in my research on the distributional consequences of bailouts.