Effective Causal Discovery under Identifiable Heteroscedastic Noise Model

Authors: Naiyu Yin, Tian Gao, Yue Yu, Qiang Ji

AAAI 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We show significant empirical gains of the proposed approaches over state-of-the-art methods on both synthetic data and real data.
Researcher Affiliation Collaboration Naiyu Yin1, Tian Gao2, Yue Yu3, Qiang Ji1 1Rensselaer Polytechnic Institute, Troy, NY. 2IBM Research, Yorktown Heights, NY. 3Lehigh University, Bethlehem, PA.
Pseudocode No We outline the full procedure (Algorithm 1), Phase-I procedure (Algorithm 2), Phase-II procedure (Algorithm 3) in supplementary Section E.
Open Source Code No Please refer to the ar Xiv version of this paper for supplementary materials. https://arxiv.org/abs/2312.12844
Open Datasets No We apply our method and baselines to two real datasets: Sachs and cause-effect pairs.
Dataset Splits No No specific information about training/validation/test dataset splits is provided.
Hardware Specification No The paper does not provide specific hardware details (e.g., exact GPU/CPU models or memory) used for running its experiments.
Software Dependencies No The paper does not provide specific ancillary software details (e.g., library or solver names with version numbers).
Experiment Setup No The paper describes the model architecture (2-layer MLPs, activation functions) and optimization method (ALM, k iterations of two phases) but does not provide specific numerical hyperparameter values (e.g., learning rate, batch size, number of epochs).