Scalable Structure Learning of Continuous-Time Bayesian Networks from Incomplete Data

Authors: Dominik Linzner, Michael Schmidt, Heinz Koeppl

NeurIPS 2019 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We demonstrate the effectiveness of our method on synthetic and two real-world data sets. For all experiments, we consider a fixed set of hyper-parameters.
Researcher Affiliation Academia Dominik Linzner1 Michael Schmidt1 Heinz Koeppl1,2 1Department of Electrical Engineering and Information Technology 2Department of Biology Technische Universität Darmstadt {dominik.linzner, michael.schmidt, heinz.koeppl}@bcs.tu-darmstadt.de
Pseudocode Yes Algorithm 1 Stationary points of Euler Lagrange equation
Open Source Code Yes An implementation of our method is available via Git1. 1https://git.rwth-aachen.de/bcs/ssl-ctbn
Open Datasets Yes We applied our method to the British Household Panel Survey (ESRC Research Centre on Micro-social Change, 2003).
Dataset Splits No The paper describes how data was used (e.g., 'a varying number of trajectories', 'picked 600 at random') but does not specify explicit train/validation/test dataset splits with percentages or sample counts for reproducibility.
Hardware Specification No The paper does not provide specific hardware details (e.g., GPU/CPU models, memory amounts) used for running its experiments.
Software Dependencies No The paper mentions using 'standard Matlab implementation of the interior-point method' but does not specify a version number for Matlab or any other software dependencies.
Experiment Setup Yes For all experiments, we consider a fixed set of hyper-parameters. We set the Dirichlet concentration parameter ci = 0.9 for all i {1, . . . , N}. Further, we assume a prior for the generators, which is uninformative on the structure αi(x, x | u) = 5 and βi(x | u) = 10, for all x, x Xi, u Ui.