Estimating Identifiable Causal Effects through Double Machine Learning

Authors: Yonghan Jung, Jin Tian, Elias Bareinboim12113-12122

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

Reproducibility Variable Result LLM Response
Research Type Experimental Simulation results corroborate with the theory. ... We evaluate the proposed estimators on the models in Examples 1 and 2.
Researcher Affiliation Academia Yonghan Jung1, Jin Tian2, Elias Bareinboim 3 1 Department of Computer Science, Purdue University 2 Department of Computer Science, Iowa State University 3 Department of Computer Science, Columbia University jung222@purdue.edu, jtian@iastate.edu, eb@cs.columbia.edu
Pseudocode Yes Algorithm 1: DML-ID (x, y, G, P) ... Algorithm 2: COMPONENTUIF(Aj, Mj r)
Open Source Code No The paper does not provide any explicit statement or link to open-source code for the described methodology.
Open Datasets Yes The causal graphs are constructed from the classic Alarm network (Beinlich et al. 1989), originally collected from a system used to monitor patients conditions.
Dataset Splits No The paper mentions 'Split D randomly into two halves: D0 and D1' for cross-fitting, but does not provide specific train/validation/test percentages, sample counts, or explicit cross-validation setups for general reproducibility of data partitioning.
Hardware Specification No The paper does not provide specific hardware details such as GPU/CPU models, processor types, or memory used for running the experiments.
Software Dependencies No The paper mentions 'Nuisance functions are estimated using gradient boosting models called XGBoost (Chen and Guestrin 2016)' but does not provide a specific version number for XGBoost or any other software.
Experiment Setup No The paper states 'Details of the models and the data-generating process are described in Appendix B.' indicating that specific experimental setup details (like hyperparameters or training schedules) are not provided in the main text.