Differentially Private Nonlinear Causal Discovery from Numerical Data

Authors: Hao Zhang, Yewei Xia, Yixin Ren, Jihong Guan, Shuigeng Zhou

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

Reproducibility Variable Result LLM Response
Research Type Experimental Extensive simulations and real-world experiments for both conditional independence test and causal discovery are conducted, which show that our method is effective in handling nonlinear numerical cases and easy to implement.
Researcher Affiliation Academia 1Shanghai Key Lab of Intelligent Information Processing, and School of Computer Science, Fudan University, China 2Department of Computer Science & Technology, Tongji University, China {haoz15, ywxia21, yxren21, sgzhou}@fudan.edu.cn; jhguan@tongji.edu.cn
Pseudocode Yes Algorithm 1: Differentially Private Nonlinear Causal Discovery (PCD)
Open Source Code Yes The source code of our method and data are available at https://github.com/Causality-Inference/PCD.
Open Datasets Yes The source code of our method and data are available at https://github.com/Causality-Inference/PCD.
Dataset Splits No The paper describes generating samples for experiments (e.g., "{200, 500} samples" and "{1000, 2000, 4000, 8000} samples") but does not provide specific training, validation, and test dataset splits with percentages or absolute counts for model evaluation.
Hardware Specification No The paper does not provide specific hardware details (exact GPU/CPU models, processor types with speeds, memory amounts, or detailed computer specifications) used for running its experiments.
Software Dependencies No The paper mentions using specific methods (e.g., KRR, HSIC), but it does not provide specific software dependencies with version numbers (e.g., Python 3.8, PyTorch 1.9) required to replicate the experiments.
Experiment Setup Yes We fix the regularization parameter λ = 1 for KRR, and set the kernel size to median distance between points and k = 100 permutations for HSIC. These are normal parameter settings for KRR and HSIC.