Interpreting Time-Varying Dynamic Bayesian Networks for Earth Climate Modelling

Enrique Valero-Leal, Pedro Larrañaga, Concha Bielza
Proceedings of The 11th International Conference on Probabilistic Graphical Models, PMLR 186:373-384, 2022.

Abstract

Bayesian networks tend to be considered as transparent and interpretable, but for big and dense networks they become harder to understand. This is the case of non-stationary, and more generally time-varying dynamic Bayesian networks, as the relations change over time and cannot be represented with a single template model. We introduce methods to explain how the model evolves qualitatively over time, and quantify these changes. In addition, we offer a functional implementation for time-varying dynamic Bayesian networks that includes our explainability proposals and some extensions that are targeted to simplify the networks in the specific field of climate sciences.

Cite this Paper

BibTeX
@InProceedings{pmlr-v186-valero-leal22a, title = {Interpreting Time-Varying Dynamic Bayesian Networks for Earth Climate Modelling}, author = {Valero-Leal, Enrique and Larra{\~n}aga, Pedro and Bielza, Concha}, booktitle = {Proceedings of The 11th International Conference on Probabilistic Graphical Models}, pages = {373--384}, year = {2022}, editor = {Salmerón, Antonio and Rumı́, Rafael}, volume = {186}, series = {Proceedings of Machine Learning Research}, month = {05--07 Oct}, publisher = {PMLR}, pdf = {https://proceedings.mlr.press/v186/valero-leal22a/valero-leal22a.pdf}, url = {https://proceedings.mlr.press/v186/valero-leal22a.html}, abstract = {Bayesian networks tend to be considered as transparent and interpretable, but for big and dense networks they become harder to understand. This is the case of non-stationary, and more generally time-varying dynamic Bayesian networks, as the relations change over time and cannot be represented with a single template model. We introduce methods to explain how the model evolves qualitatively over time, and quantify these changes. In addition, we offer a functional implementation for time-varying dynamic Bayesian networks that includes our explainability proposals and some extensions that are targeted to simplify the networks in the specific field of climate sciences.} }
Endnote
%0 Conference Paper %T Interpreting Time-Varying Dynamic Bayesian Networks for Earth Climate Modelling %A Enrique Valero-Leal %A Pedro Larrañaga %A Concha Bielza %B Proceedings of The 11th International Conference on Probabilistic Graphical Models %C Proceedings of Machine Learning Research %D 2022 %E Antonio Salmerón %E Rafael Rumı́ %F pmlr-v186-valero-leal22a %I PMLR %P 373--384 %U https://proceedings.mlr.press/v186/valero-leal22a.html %V 186 %X Bayesian networks tend to be considered as transparent and interpretable, but for big and dense networks they become harder to understand. This is the case of non-stationary, and more generally time-varying dynamic Bayesian networks, as the relations change over time and cannot be represented with a single template model. We introduce methods to explain how the model evolves qualitatively over time, and quantify these changes. In addition, we offer a functional implementation for time-varying dynamic Bayesian networks that includes our explainability proposals and some extensions that are targeted to simplify the networks in the specific field of climate sciences.
APA
Valero-Leal, E., Larrañaga, P. & Bielza, C.. (2022). Interpreting Time-Varying Dynamic Bayesian Networks for Earth Climate Modelling. Proceedings of The 11th International Conference on Probabilistic Graphical Models, in Proceedings of Machine Learning Research 186:373-384 Available from https://proceedings.mlr.press/v186/valero-leal22a.html.

Related Material