Robust Model Selection of Gaussian Graphical Models
Authors: Abrar Zahin, Rajasekhar Anguluri, Lalitha Sankar, Oliver Kosut, Gautam Dasarathy
TMLR 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details
| Reproducibility Variable | Result | LLM Response |
|---|---|---|
| Research Type | Experimental | We further provide finite sample guarantees in the high-dimensional regime for our algorithm and validate our results through numerical simulations. We also show the efficacy of our algorithm through experiments on synthetic and realistic network structures. We perform experiments on both a synthetic graph and an IEEE 33 bus system, which is a graphical representation of established IEEE-33 bus benchmark distribution system (Baran & Wu, 1989), to assess the validity of our theoretical results and to demonstrate the performance of No MAD. |
| Researcher Affiliation | Academia | Abrar Zahin EMAIL School of Electrical, Computer & Energy Engineering Arizona State University, Rajasekhar Anguluriú EMAIL Department of Computer Science & Electrical Engineering University of Maryland, Baltimore County, Lalitha Sankar EMAIL School of Electrical, Computer & Energy Engineering Arizona State University, Oliver Kosut EMAIL School of Electrical, Computer & Energy Engineering Arizona State University, Gautam Dasarathy EMAIL School of Electrical, Computer & Energy Engineering Arizona State University |
| Pseudocode | Yes | A high level overview of the algorithm is given in Algorithm 1. Its operation may be divided into two main steps: (a) learning Pop and Aop; and (b) learning Eop. These steps are summarized in the following. A formal algorithmic listing and a full description can be found in the Appendix. Complete pseudocode for this step appears in Subroutine 3 in Appendix. |
| Open Source Code | No | The paper does not contain any explicit statements about providing access to source code, nor does it provide any links to code repositories. |
| Open Datasets | Yes | We perform experiments on both a synthetic graph and an IEEE 33 bus system, which is a graphical representation of established IEEE-33 bus benchmark distribution system (Baran & Wu, 1989), to assess the validity of our theoretical results and to demonstrate the performance of No MAD. |
| Dataset Splits | No | The paper does not explicitly provide details about training/test/validation dataset splits, only mentioning 'a fixed sample size across 15 trials' for the experiments. |
| Hardware Specification | No | The paper does not provide any specific hardware details such as GPU models, CPU types, or memory amounts used for running the experiments. |
| Software Dependencies | No | The paper mentions using GLASSO, but does not provide specific version numbers for any software, libraries, or programming languages used in the implementation. |
| Experiment Setup | Yes | Our protocol for selecting the best GLASSO is as follows: for a fixed maximum allowable diagonals, we selected the regularization parameter for which the output graph given the noise covariance matrix to GLASSO is in equivalence class. Then, for that fixed regularization parameter we report the performance of GLASSO for varying numbers of (increased) maximum allowable diagonals. To investigate this influence, we conducted the experiment with a fixed sample size across 15 trials. |