On the Recoverability of Causal Relations from Temporally Aggregated I.I.D. Data

Authors: Shunxing Fan, Mingming Gong, Kun Zhang

ICML 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We conducted five simulation experiments to support our findings, focusing on theorems for functional and conditional consistency, examining the effect of the k value, and proposing a trivial solution for the aggregation issue.
Researcher Affiliation Academia 1Department of Machine Learning, Mohamed bin Zayed University of Artificial Intelligence, Abu Dhabi, United Arab Emirates 2School of Mathematics and Statistics, The University of Melbourne, Melbourne, Australia 3Department of Philosophy, Carnegie Mellon University, Pittsburgh, PA, United States. Correspondence to: Kun Zhang <kunz1@cmu.edu>.
Pseudocode No The paper does not contain any pseudocode or algorithm blocks.
Open Source Code No The paper does not provide any statement or link indicating the release of open-source code for the described methodology.
Open Datasets No The paper describes generating synthetic datasets for simulations (e.g., 'generating the dataset X, Y with a sample size of 10,000'), rather than using pre-existing publicly available datasets that would require access information.
Dataset Splits No The paper describes generating data and running simulations (e.g., 'generating the dataset X, Y with a sample size of 10,000', '100 repetitions') but does not specify explicit training, validation, and test dataset splits in percentages or absolute counts for reproducibility of data partitioning.
Hardware Specification No The paper does not provide any specific details about the hardware (e.g., GPU models, CPU types, memory) used to run the experiments.
Software Dependencies No The paper mentions various methods and tools like 'PC', 'FCI', 'GES', 'Direct Li NGAM', 'ANM', 'Kernel Conditional Independence test (KCI)', but it does not specify version numbers for any of these software components, libraries, or programming languages.
Experiment Setup Yes Regarding the method parameters, we used the Fisher-Z test for PC and FCI, and the BIC score for GES in the linear scenario. In the nonlinear scenario, we set the conditional independence test for PC and FCI as the Kernel Conditional Independence Test (KCI) with the default kernel and chose the local score CV general score function for GES. All other settings are kept at their default values.