Learning Regularization for Graph Inverse Problems

Authors: Moshe Eliasof, Md Shahriar Rahim Siddiqui, Carola-Bibiane Schönlieb, Eldad Haber

AAAI 2025 | Venue PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental We evaluate the five models discussed in Section 4 on the diverse set of GRIPs discussed in Section 2. We utilize several datasets to demonstrate GRIPs, ranging from synthetic graphs (CLUSTER) to geometric datasets (Shape Net), natural images (CIFAR10), road traffic (METR-LA), and spread of disease (Chickenpox-Hungary, shortened to CPOX). We provide full details on the datasets and our motivation for using them to demonstrate different problems, in Appendix D. We also provide a comprehensive description of the training, evaluation, and hyperparameter selection procedures for each problem in Appendix E, hyperparameter sensitivity results in Appendix F, and a discussion of the complexity and runtimes of the methods, in Appendix G.
Researcher Affiliation Academia Moshe Eliasof1, Md Shahriar Rahim Siddiqui2, Carola-Bibiane Sch onlieb1, Eldad Haber3 1Department of Applied Mathematics, University of Cambridge, Cambridge, United Kingdom 2Department of Physics and Astronomy, University of British Columbia, Vancouver, Canada 3Department of Earth, Ocean and Atmospheric Sciences, University of British Columbia, Vancouver, Canada EMAIL, EMAIL, EMAIL, EMAIL
Pseudocode No The paper describes methodologies using mathematical equations and descriptions of steps, but does not present any explicitly labeled 'Pseudocode' or 'Algorithm' block, nor any structured code-like procedures.
Open Source Code Yes Our code is implemented in Pytorch (Paszke et al. 2017), and is available at https://github.com/nafi007/Graph Inverse Problems.
Open Datasets Yes We utilize several datasets to demonstrate GRIPs, ranging from synthetic graphs (CLUSTER) to geometric datasets (Shape Net), natural images (CIFAR10), road traffic (METR-LA), and spread of disease (Chickenpox-Hungary, shortened to CPOX).
Dataset Splits Yes We provide full details on the datasets and our motivation for using them to demonstrate different problems, in Appendix D. We also provide a comprehensive description of the training, evaluation, and hyperparameter selection procedures for each problem in Appendix E, hyperparameter sensitivity results in Appendix F, and a discussion of the complexity and runtimes of the methods, in Appendix G.
Hardware Specification No The paper does not provide specific details about the hardware used for running experiments, such as GPU models, CPU types, or memory specifications.
Software Dependencies No Our code is implemented in Pytorch (Paszke et al. 2017), and is available at https://github.com/nafi007/Graph Inverse Problems.
Experiment Setup Yes We provide full details on the datasets and our motivation for using them to demonstrate different problems, in Appendix D. We also provide a comprehensive description of the training, evaluation, and hyperparameter selection procedures for each problem in Appendix E, hyperparameter sensitivity results in Appendix F, and a discussion of the complexity and runtimes of the methods, in Appendix G.