# Discussion of Big Bayes Stories and BayesBag

arXiv: https://arxiv.org/abs/1405.4977 · HTML: https://mlsanity.com/paper/1405.4977/

## Key facts

- Cited by 6 papers in the map (data through 14 Oct 2026).
- First posted on arXiv 20 May 2014.

## Built on by (most cited first, 6 of 6)

- [Model-based metrics: Sample-efficient estimates of predictive model subpopulation performance](https://mlsanity.com/paper/2104.12231/) (2021)
- [Uncertainty Quantification and Propagation in Surrogate-based Bayesian Inference](https://mlsanity.com/paper/2312.05153/) (2023)
- [A Probabilistic Framework for Mutation Testing in Deep Neural Networks](https://mlsanity.com/paper/2208.06018/) (2022)
- [Stabilizing black-box model selection with the inflated argmax](https://mlsanity.com/paper/2410.18268/) (2024)
- [Reproducible Parameter Inference Using Bagged Posteriors](https://mlsanity.com/paper/2311.02019/) (2023)
- [Variational bagging: a robust approach for Bayesian uncertainty quantification](https://mlsanity.com/paper/2511.20594/) (2025)

## Abstract (arXiv)

I congratulate all the authors for their insightful papers with wide-ranging contributions. The articles demonstrate the power and elegance of the Bayesian inference paradigm. In particular, it allows to incorporate prior knowledge as well as hierarchical model building in a convincing way. Regarding the latter, the contribution by Raftery, Alkema and German is a very fascinating piece, as it addresses a set of problems of great public interest and presents predictions for the world populations and other interesting quantities with uncertainty regions. Their approach is based on a hierarchical model, taking various characteristics into account (e.g., fertility projections). It would have been very difficult to come up with a "better" solution which would be as clear in terms of interpretation (in contrast to a "black-box machine") and which would provide (model-based) uncertainties for the predictions into the future.