Události

Čt 06.10.2022

Simon Scheidegger (University of Lausanne) "Deep learning for climate change mitigation"

Simon Scheidegger, Ph.D.

University of Lausanne, Switzerland

Abstract: There is  a  growing  demand  to  quantify  parametric  uncertainty  as  well  as  economic  and  climate uncertainty on the climate policies to tackle global warming. To investigate parametric uncertainty and nonlinear interactions among the uncertain model parameters, we develop a high-dimensional stochastic climate-economy model that propagates parametric uncertainty as pseudo-states. We approximate all equilibrium functions using deep equilibrium nets. To limit the number of model evaluations to obtain convergent  statistics,  we  further  interpolate  the  outcomes  of  the  cheap-to-evaluate  surrogate  model employing  Gaussian  processes  in  combination  with  Bayesian  active  learning,  from  which  we analytically  estimate  the  Sobol'  indices  and  univariate  effects.  The  uncertainty  quantification  results show that the equilibrium climate sensitivity dominates the level of the social cost of carbon. In contrast, the stochastic and persistent long-run growth risk characterizes the volatilities of economic moments.