A Proxy Variable View of Shared Confounding

Authors: Yixin Wang, David Blei

ICML 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental 5. A Simulation Study In this section, we see the identification results in action. We find that the identification conditions discussed in Sections 2 and 3 are crucial for producing correct causal estimates. The theoretical results and the conditions required by Theorems 1, 2, 5 and 6 are practically important. Specifically, we consider a linear data generating process in Section 4 with a one-dimensional U and three treatments A1, A2, A3. We explore two configurations of the unobserved confounder U. In one configuration, U is normally distributed, and the resulting observational data satisfies the completeness condition in Assumption 1.2. Figure 3a shows the mean squared error (RMSE) of the deconfounder average treatment effect (ATE) estimate stays low even if the confounding strength is high while the RMSE of naive regression quickly blows up.
Researcher Affiliation Academia 1University of California, Berkeley 2Columbia University. Correspondence to: David M. Blei <david.blei@columbia.edu>.
Pseudocode No The paper describes the deconfounder algorithm in a step-by-step manner within paragraph text, but it does not present it as a formally labeled 'Pseudocode' or 'Algorithm' block.
Open Source Code No The paper mentions and builds upon the deconfounder algorithm from Wang & Blei (2019a) but does not provide a statement or link for open-source code specific to the contributions or experiments in this paper.
Open Datasets No The paper describes a 'linear data generating process' for its simulation study, implying the use of synthetic data, and does not provide concrete access information or citations for any publicly available or open datasets.
Dataset Splits No The paper describes a simulation study but does not provide specific dataset split information (e.g., percentages, sample counts, or methodology for training, validation, and testing splits).
Hardware Specification No The paper does not provide any specific hardware details (e.g., exact GPU/CPU models, memory amounts, or detailed computer specifications) used for running its simulation experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers) needed to replicate the experiments.
Experiment Setup Yes Specifically, we consider a linear data generating process in Section 4 with a one-dimensional U and three treatments A1, A2, A3. We explore two configurations of the unobserved confounder U. In one configuration, U is normally distributed... In a second configuration, U is uniformly distributed...