This paper introduces a simulation-based deep learning approach to Bayesian inference to the economic literature. This technique allows for the estimation of a posterior even when the likelihood is intractable and works even with data that has a latent time structure which is ubiquitous in macroeconomic structural modeling. This approach allows value function iteration and other methods that don't return a likelihood function to be estimated in a Bayesian fashion for likely the first time.
This paper uses a transformer neural network model to perform imputation of missing data. The method returns a distribution which allows for easy marginalization which can allow for statistically efficient analysis when combined with a model for inference.
This paper demonstrates how to use Variational Inference as well as a flexible variational family that allows fast estimation of DSGE models, which roughly reduces computation time by a factor of 100.
This paper demonstrates how to use variational inference as well as a flexible variational family that allows fast estimation of DSGE models, which roughly reduces computation time by a factor of 100 compared to Markov chain Monte Carlo.