Recovering Causal Structures from Low-Order Conditional Independencies

Authors: Marcel Wienöbst, Maciej Liskiewicz10302-10309

AAAI 2020 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We experimentally compare this algorithm to previous approaches in Section 6. Finally, we discuss our results in Section 7.
Researcher Affiliation Academia Marcel Wien obst, Maciej Li skiewicz Institute of Theoretical Computer Science, University of L ubeck, Germany {wienoebst, liskiewi}@tcs.uni-luebeck.de
Pseudocode Yes Algorithm 1: The LOCI algorithm computes the representation G for a DAG-representable set of CIs up to order k.
Open Source Code No The paper does not provide an explicit statement or link for open-sourcing the code for the described methodology.
Open Datasets No The paper states how data is generated: 'An undirected graph with n nodes is drawn randomly. More precisely, each edge is present with probability d/(n 1), meaning every node has expected degree d. Afterwards, a topological ordering of the nodes is randomly chosen in order to obtain a DAG D from the generated graph. From this DAG we can read off all zeroand first-order independencies...' This describes a data generation process rather than the use of a publicly available dataset with concrete access information.
Dataset Splits No The paper uses generated data for experimental analysis ('All values are the means of 100 independent trials') but does not specify explicit train/validation/test splits, as the focus is on recovering causal structures in an oracle model rather than typical predictive model training.
Hardware Specification No The paper does not provide any specific details about the hardware used to run the experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers.
Experiment Setup No The paper describes the parameters for generating random DAGs for evaluation (e.g., 'random DAGs with n nodes and expected node degree d'), but it does not specify hyperparameters or system-level training settings for a model in the traditional sense, as it focuses on an algorithm for causal structure learning.