Direct Estimation of Differences in Causal Graphs

Authors: Yuhao Wang, Chandler Squires, Anastasiya Belyaeva, Caroline Uhler

NeurIPS 2018 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the properties of our method through a simulation study and apply it to the analysis of gene expression data from ovarian cancer and during T-cell activation.
Researcher Affiliation Academia Yuhao Wang Lab for Information & Decision Systems and Institute for Data, Systems and Society Massachusetts Institute of Technology Cambridge, MA 02139 yuhaow@mit.edu
Pseudocode Yes Algorithm 1 Difference Causal Inference (DCI) algorithm
Open Source Code Yes The code utilized for the following experiments can be found at https://github.com/csquires/dci.
Open Datasets Yes We tested our method on an ovarian cancer data set [37] that contains two groups of patients with different survival rates and was previously analyzed using the DPM algorithm in the undirected setting [43]. We applied DCI to single-cell gene expression data of naive and activated T cells in order to study the pathways involved during the immune response to a pathogen. We analyzed data from 377 activated and 298 naive T cells obtained by [34] using the recent drop-seq technology.
Dataset Splits Yes We initiated DCI using KLIEP, thresholding the edge weights at 0.005, and ran DCI for different tuning parameters and with cross-validation to obtain the final DCI output shown in Figure 2 (c) using stability selection as described in [21].
Hardware Specification No The paper does not provide specific details about the hardware used for running the experiments, such as GPU/CPU models, memory, or cloud instance types.
Software Dependencies No The paper mentions using KLIEP, stability selection, PC, and GES algorithms but does not provide specific version numbers for these software components or any underlying libraries/frameworks.
Experiment Setup Yes In Figure 1 we analyzed how the performance of the DCI algorithm changes over different choices of significance levels α. The simulations were performed on graphs with p = 10 nodes, neighborhood size of s = 3 and sample size n {103, 104}.