Approximate Causal Abstractions

Sander Beckers, Frederick Eberhardt, Joseph Y. Halpern
Proceedings of The 35th Uncertainty in Artificial Intelligence Conference, PMLR 115:606-615, 2020.

Abstract

Scientific models describe natural phenomena at different levels of abstraction. Abstract descriptions can provide the basis for interventions on the system and explanation of observed phenomena at a level of granularity that is coarser than the most fundamental account of the system. Beckers and Halpern (2019), building on prior work of Rubinstein et al. (2017), developed an account of abstraction for causal models that is exact. Here we extend this account to the more realistic case where an abstract causal model only offers an approximation of the underlying system. We show how the resulting account handles the discrepancy that can arise between low- and high-level causal models of the same system, and in the process provide an account of how one causal model approximates another, a topic of independent interest. Finally, we extend the account of approximate abstractions to probabilistic causal models, indicating how and where uncertainty can enter into an approximate abstraction.

Cite this Paper

BibTeX
@InProceedings{pmlr-v115-beckers20a, title = {Approximate Causal Abstractions}, author = {Beckers, Sander and Eberhardt, Frederick and Halpern, Joseph Y.}, booktitle = {Proceedings of The 35th Uncertainty in Artificial Intelligence Conference}, pages = {606--615}, year = {2020}, editor = {Adams, Ryan P. and Gogate, Vibhav}, volume = {115}, series = {Proceedings of Machine Learning Research}, month = {22--25 Jul}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v115/beckers20a/beckers20a.pdf}, url = {https://proceedings.mlr.press/v115/beckers20a.html}, abstract = {Scientific models describe natural phenomena at different levels of abstraction. Abstract descriptions can provide the basis for interventions on the system and explanation of observed phenomena at a level of granularity that is coarser than the most fundamental account of the system. Beckers and Halpern (2019), building on prior work of Rubinstein et al. (2017), developed an account of abstraction for causal models that is exact. Here we extend this account to the more realistic case where an abstract causal model only offers an approximation of the underlying system. We show how the resulting account handles the discrepancy that can arise between low- and high-level causal models of the same system, and in the process provide an account of how one causal model approximates another, a topic of independent interest. Finally, we extend the account of approximate abstractions to probabilistic causal models, indicating how and where uncertainty can enter into an approximate abstraction.} }
Endnote
%0 Conference Paper %T Approximate Causal Abstractions %A Sander Beckers %A Frederick Eberhardt %A Joseph Y. Halpern %B Proceedings of The 35th Uncertainty in Artificial Intelligence Conference %C Proceedings of Machine Learning Research %D 2020 %E Ryan P. Adams %E Vibhav Gogate %F pmlr-v115-beckers20a %I PMLR %P 606--615 %U https://proceedings.mlr.press/v115/beckers20a.html %V 115 %X Scientific models describe natural phenomena at different levels of abstraction. Abstract descriptions can provide the basis for interventions on the system and explanation of observed phenomena at a level of granularity that is coarser than the most fundamental account of the system. Beckers and Halpern (2019), building on prior work of Rubinstein et al. (2017), developed an account of abstraction for causal models that is exact. Here we extend this account to the more realistic case where an abstract causal model only offers an approximation of the underlying system. We show how the resulting account handles the discrepancy that can arise between low- and high-level causal models of the same system, and in the process provide an account of how one causal model approximates another, a topic of independent interest. Finally, we extend the account of approximate abstractions to probabilistic causal models, indicating how and where uncertainty can enter into an approximate abstraction.
APA
Beckers, S., Eberhardt, F. & Halpern, J.Y.. (2020). Approximate Causal Abstractions. Proceedings of The 35th Uncertainty in Artificial Intelligence Conference, in Proceedings of Machine Learning Research 115:606-615 Available from https://proceedings.mlr.press/v115/beckers20a.html.

Related Material