A unifying causal framework for analyzing dataset shift-stable learning algorithms

Subbaswamy Adarsh (), Chen Bryant () and Saria Suchi ()
Additional contact information
Subbaswamy Adarsh: Department of Computer Science, Johns Hopkins University, Baltimore, MD 21218, United States
Chen Bryant: Brex Inc, San Francisco, California, United States
Saria Suchi: Department of Computer Science, Johns Hopkins University and Bayesian Health, Baltimore, MD 21218, United States

Journal of Causal Inference, 2022, vol. 10, issue 1, 64-89

Abstract: Recent interest in the external validity of prediction models (i.e., the problem of different train and test distributions, known as dataset shift) has produced many methods for finding predictive distributions that are invariant to dataset shifts and can be used for prediction in new, unseen environments. However, these methods consider different types of shifts and have been developed under disparate frameworks, making it difficult to theoretically analyze how solutions differ with respect to stability and accuracy. Taking a causal graphical view, we use a flexible graphical representation to express various types of dataset shifts. Given a known graph of the data generating process, we show that all invariant distributions correspond to a causal hierarchy of graphical operators, which disable the edges in the graph that are responsible for the shifts. The hierarchy provides a common theoretical underpinning for understanding when and how stability to shifts can be achieved, and in what ways stable distributions can differ. We use it to establish conditions for minimax optimal performance across environments, and derive new algorithms that find optimal stable distributions. By using this new perspective, we empirically demonstrate that that there is a tradeoff between minimax and average performance.

Keywords: dataset shift; transportability; invariance; stability (search for similar items in EconPapers)
Date: 2022
References: View references in EconPapers View complete reference list from CitEc
Citations:

Downloads: (external link)
https://doi.org/10.1515/jci-2021-0042 (text/html)

Related works:
This item may be available elsewhere in EconPapers: Search for items with the same title.

Export reference: BibTeX RIS (EndNote, ProCite, RefMan) HTML/Text

Persistent link: https://EconPapers.repec.org/RePEc:bpj:causin:v:10:y:2022:i:1:p:64-89:n:2

DOI: 10.1515/jci-2021-0042

Access Statistics for this article

Journal of Causal Inference is currently edited by Elias Bareinboim, Jin Tian and Iván Díaz

More articles in Journal of Causal Inference from De Gruyter
Bibliographic data for series maintained by Peter Golla ().