# Local Bayesian Dirichlet mixing of imperfect models

arXiv: https://arxiv.org/abs/2311.01596 · HTML: https://mlsanity.com/paper/2311.01596/

## Key facts

- Cited by 2 papers in the map (data through 14 Oct 2026).
- Builds on 10 of its 54 references that are in the map.
- First posted on arXiv 2 Nov 2023.

## Builds on (most cited first, 10 of 10)

- [Using stacking to average Bayesian predictive distributions](https://mlsanity.com/paper/1704.02030/) (2017)
- [Bayesian model averaging: A systematic review and conceptual classification](https://mlsanity.com/paper/1509.08864/) (2015)
- [Bayesian hierarchical stacking: Some models are (somewhere) useful](https://mlsanity.com/paper/2101.08954/) (2021)
- [Bayesian approach to model-based extrapolation of nuclear observables](https://mlsanity.com/paper/1806.00552/) (2018)
- [Quantified limits of the nuclear landscape](https://mlsanity.com/paper/2001.05924/) (2020)
- [Neutron drip line in the Ca region from Bayesian model averaging](https://mlsanity.com/paper/1901.07632/) (2019)
- [Machine Learning in Nuclear Physics](https://mlsanity.com/paper/2112.02309/) (2021)
- [Model Mixing Using Bayesian Additive Regression Trees](https://mlsanity.com/paper/2301.02296/) (2023)
- [Statistical aspects of nuclear mass models](https://mlsanity.com/paper/2002.04151/) (2020)
- [Beyond the proton drip line: Bayesian analysis of proton-emitting nuclei](https://mlsanity.com/paper/1910.12624/) (2019)

## Built on by (most cited first, 2 of 2)

- [Model orthogonalization and Bayesian forecast mixing via Principal Component Analysis](https://mlsanity.com/paper/2405.10839/) (2024)
- [Bridging Ab Initio Symmetries and Global Nuclear Masses with Interpretable Neural Networks](https://mlsanity.com/paper/2606.28287/) (2026)

## Abstract (arXiv)

To improve the predictability of complex computational models in the experimentally-unknown domains, we propose a Bayesian statistical machine learning framework utilizing the Dirichlet distribution that combines results of several imperfect models. This framework can be viewed as an extension of Bayesian stacking. To illustrate the method, we study the ability of Bayesian model averaging and mixing techniques to mine nuclear masses. We show that the global and local mixtures of models reach excellent performance on both prediction accuracy and uncertainty quantification and are preferable to classical Bayesian model averaging. Additionally, our statistical analysis indicates that improving model predictions through mixing rather than mixing of corrected models leads to more robust extrapolations.