On Graph Reconstruction via Empirical Risk Minimization: Fast Learning Rates and Scalability

Authors: Guillaume Papa, Aurélien Bellet, Stephan Clémençon

NeurIPS 2016 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental Finally, we illustrate our theoretical results by numerical experiments on synthetic and real graphs.
Researcher Affiliation Academia Guillaume Papa, Stéphan Clémençon LTCI, CNRS, Télécom Paris Tech, Université Paris-Saclay 75013, Paris, France first.last@telecom-paristech.fr Aurélien Bellet INRIA 59650 Villeneuve d Ascq, France aurelien.bellet@inria.fr
Pseudocode No The paper does not contain any structured pseudocode or clearly labeled algorithm blocks.
Open Source Code No The paper does not provide any statement or link indicating the availability of open-source code for the methodology described.
Open Datasets No The paper describes generating its own synthetic graph data ('We create a synthetic graph with n nodes as follows.') and mentions 'experiments on a real network' in the Supplementary Material section, but it does not provide concrete access information (e.g., a specific link, DOI, or a formal citation with authors and year for a publicly available dataset) for either the synthetic or real data.
Dataset Splits No The paper describes generating a 'training graph' and a 'test graph' separately but does not specify explicit percentages or sample counts for training/validation/test splits from a single dataset, nor does it mention cross-validation. The phrase 'dataset splitting strategy given by (6)' refers to an alternative empirical risk estimation method, not necessarily their experimental setup's data partitioning.
Hardware Specification No The paper does not provide any specific details about the hardware used for running its experiments (e.g., GPU/CPU models, memory, or cloud resources).
Software Dependencies No The paper does not specify any software dependencies with version numbers (e.g., libraries, frameworks, or solvers).
Experiment Setup Yes We create a synthetic graph with n nodes as follows. Each node i has features Xtrue i Rq sampled uniformly over [0, 1]...Using this procedure, we generate a training graph with n = 1,000,000 and q = 100. We set the threshold τ such that there is an edge between about 20% of the node pairs, and set p = 0.05...Table 1 shows the test error (averaged over 10 runs)...