Discussion on Using Stacking to Average Bayesian Predictive Distributions by Yao et al

TL;DR

This paper discusses the stacking method for averaging Bayesian predictive distributions, exploring its theoretical foundations, graphical modeling perspective, and potential applications in distributed computing for aggregating (sub)posterior distributions.

Contribution

It provides a comprehensive discussion on the stacking approach, including its theoretical aspects and potential for broader applications in Bayesian inference and distributed computing.

Findings

Stacking effectively averages Bayesian predictive distributions.

Graphical modeling offers insights into stacking methods.

Potential for stacking in distributed (sub)posterior aggregation.

Abstract

I begin by summarizing key ideas of the paper under discussion. Then I will talk about a graphical modeling perspective, posterior contraction rates and alternative methods of aggregation. Moreover, I will also discuss possible applications of the stacking method to other problems, in particular, aggregating (sub)posterior distributions in distributed computing.