Causal Discovery in Semi-Stationary Time Series

Authors: Shanyun Gao, Raghavendra Addanki, Tong Yu, Ryan Rossi, Murat Kocaoglu

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

Reproducibility Variable Result LLM Response
Research Type Experimental We validate the algorithm with extensive experiments on continuous and discrete simulated data. We also apply our algorithm to a real-world climate dataset.
Researcher Affiliation Collaboration Shanyun Gao Purdue University gao565@purdue.edu Raghavendra Addanki Adobe Research raddanki@adobe.com Tong Yu Adobe Research tyu@adobe.com Ryan A. Rossi Adobe Research ryrossi@adobe.com Murat Kocaoglu Purdue University mkocaoglu@purdue.edu
Pseudocode Yes Algorithm 1 PCMCIΩ, Algorithm A1 PCqmax, Algorithm A2 MCI, Algorithm B1 PCMCIΩ
Open Source Code Yes The Python code is provided at https://github.com/Causal ML-Lab/PCMCI-Omega.
Open Datasets No The paper mentions using "continuous and discrete simulated data" and a "real-world climate dataset" from 1948-2022, but it does not provide concrete access information (link, DOI, specific citation with authors/year) for the climate dataset, nor does it make the simulated data publicly available.
Dataset Splits No The paper describes how synthetic data is generated but does not specify explicit training, validation, or test splits. It states that performance statistics are averaged over 100 trials.
Hardware Specification Yes All experiments, including those detailed in the main paper, are conducted on a single node with one core, utilizing 512 GB of memory in the Gilbreth cluster at Purdue University.
Software Dependencies No The paper mentions that "The Python code is provided at https://github.com/Causal ML-Lab/PCMCI-Omega." implying Python is used, but it does not provide specific version numbers for Python or any other libraries/packages.
Experiment Setup Yes Set τub = 15, ωub = 15 for all variables.