Notice: The reproducibility variables underlying each score are classified using an automated LLM-based pipeline, validated against a manually labeled dataset. LLM-based classification introduces uncertainty and potential bias; scores should be interpreted as estimates. Full accuracy metrics and methodology are described in Coakley et alK. L. Coakley, T. Snelleman, H. Hoos, and O. E. Gundersen, "The embrace of open science: An analysis of a decade of AI research and 56 800 conference papers," Under Review, 2026..
DSCS: Fast CPDAG-Based Verification of Collapsible Submodels in High-Dimensional Bayesian Networks
Authors: Wentao Wu, Shiyuan He, Jianhua Guo
NeurIPS 2025 | Venue PDF | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | Extensive numerical experiments demonstrate the practicality, scalability, and efficiency of our proposed approach. ... Through numerical experiments, we examine the performance of the proposed DSCS Algorithm in Algorithm 1. Our algorithm is compared to the approach of directly detecting sequential c-removability of M, as proposed by [28]. |
| Researcher Affiliation | Academia | 1Northeast Normal University, 2Beijing Technology and Business University |
| Pseudocode | Yes | Algorithm 1: DSCS Algorithm |
| Open Source Code | Yes | Additionally, we provide the code in the supplemental material to ensure the easy reproduction of all reported results. |
| Open Datasets | Yes | we selected three real-world Bayesian networks from the R-package bnlearn WIN95PTS (76 nodes, 112 edges), LINK (724 nodes, 1125 edges), and MUNIN (1041 nodes, 1397 edges). A detailed description of the three Bayesian networks can be found at https://www.bnlearn.com/bnrepository/. |
| Dataset Splits | Yes | In each simulation replicate, we randomly generate a DAG with n {500, 1000, 1500, 2000, 4000, 6000, 8000, 10000} and p {0.1, 0.05, 0.01, 0.005}, and the vertex subset M was randomly selected from the graph with cardinality |M|. ... For each combination of n, p, and |M|, the simulation was run 30 times. |
| Hardware Specification | Yes | The experiments were implemented with R and run on a computer with 2.20GHz CPU and 256 GB memory. |
| Software Dependencies | No | The experiments were implemented with R and run on a computer... Erdös-Rényi graphs were randomly generated using the R-package pcalg [54]... |
| Experiment Setup | Yes | In each simulation replicate, we randomly generate a DAG with n {500, 1000, 1500, 2000, 4000, 6000, 8000, 10000} and p {0.1, 0.05, 0.01, 0.005}, and the vertex subset M was randomly selected from the graph with cardinality |M|. ... For each combination of n, p, and |M|, the simulation was run 30 times. |