Characterization and Learning of Causal Graphs with Small Conditioning Sets

Authors: Murat Kocaoglu

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

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we conduct synthetic, and semi-synthetic experiments to demonstrate that the k-PC algorithm enables more robust causal discovery in the small sample regime compared to the baseline algorithms.
Researcher Affiliation Academia Murat Kocaoglu School of Electrical and Computer Engineering Purdue University mkocaoglu@purdue.edu
Pseudocode Yes Algorithm 1 k-PC Algorithm [...] Algorithm 2 FCI_Orient
Open Source Code Yes The Python code is provided at https://github.com/CausalML-Lab/kPC.
Open Datasets Yes We test our algorithm on the semi-synthetic Asia dataset from bnlearn repository and compare it with PC algorithm. [...] We randomly sample DAGs using the pyAgrum package.
Dataset Splits No The paper mentions generating multiple datasets per model instance but does not specify a formal validation set or explicit training/validation/test splits, percentages, or absolute sample counts for reproduction.
Hardware Specification No The paper does not provide specific hardware details such as exact GPU/CPU models, processor types, or memory amounts used for running its experiments.
Software Dependencies No The paper mentions 'causal-learn package', 'pyAgrum package', and 'pcalg package in R' but does not provide specific version numbers for these software components.
Experiment Setup Yes For each DAG, conditional probability tables are independently and uniformly randomly filled from the corresponding probability simplex. [...] For each DAG, a linear SCM is sampled as follows: Each coefficient is chosen randomly in the range [-3, 3]. Exogenous noise terms are jointly independent unit Gaussian.