Learning Nonparametric Latent Causal Graphs with Unknown Interventions

Authors: Yibo Jiang, Bryon Aragam

NeurIPS 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we verify our theoretical results in a simulation study. ... We test the theoretical results on simulated datasets under two settings: pure child and single latent source. ... The results show that even without the graphical assumptions, our method can be effective in recovering the DAG for nonlinear models.
Researcher Affiliation Academia Yibo Jiang University of Chicago yiboj@uchicago.edu Bryon Aragam University of Chicago bryon@chicagobooth.edu
Pseudocode Yes Pseudocode for the overall approach can be found in Algorithm 2 in Appendix G.
Open Source Code No The paper does not provide an explicit statement or a link to open-source code for the methodology described.
Open Datasets No The paper states, "We test the theoretical results on simulated datasets" but does not provide access information (link, citation, or repository) for a publicly available dataset.
Dataset Splits No The paper states "We test the theoretical results on simulated datasets" and mentions running "100 runs" but does not specify training, validation, or test splits. The term "validation" is used in the schema but not in the context of dataset splits in the paper.
Hardware Specification No The paper does not explicitly describe the hardware used to run its experiments.
Software Dependencies No The paper mentions "Chatterjee’s coefficient" but does not list any software components or libraries with specific version numbers.
Experiment Setup Yes For each variable Vi in the causal graph, the structural equation is simply Vi P Vj pa(Vi) f(Vj) + ϵ, where ϵ is Gaussian noise, and f is a nonlinear function. We set f to be a quadratic function. To test independence, we adopt Chatterjee’s coefficient [10].