Deep Learning based Network Similarity for Model Selection

Tracking #: 624-1604

Responsible editor: 

Michael Maes

Submission Type: 

Research Paper


Capturing data in the form of network’s is becoming an increasingly popular approach for modeling, analyzing and visualizing complex phenomena, to understand the important properties of the underlying complex processes. Access to many large-scale network datasets is restricted due to the privacy and security concerns. Also for several applications (such as functional connectivity networks), generating large scale real data is expensive. For these reasons, there is a growing need for advanced mathematical and statistical models (also called generative models) that can account for the structure of these large scale networks, without having to materialize them in the real world. The objective is to provide a comprehensible description of the network properties and to be able to infer previously unobserved properties. Various models have been developed by researchers, which generate synthetic networks that adhere to the structural properties of real networks. However, the selection of the appropriate generative model for a given real-world network remains an important challenge. In this paper, we investigate this problem and provide a novel technique (named as TripletFit) for model selection (or network classification) and estimation of structural similarities of the complex networks. The goal of network model selection is to select a generative model that is able to generate a structurally similar synthetic network for a given real-world (target) network. We consider six outstanding generative models as the candidate models. The existing model selection methods mostly suffer from sensitivity to network perturbations, dependency on the size of the networks, and low accuracy. To overcome these limitations, we considered a broad array of network features, with the aim of representing different structural aspects of the network and employed deep learning techniques such as deep triplet network architecture and simple feed-forward network for model selection and estimation of structural similarities of the complex networks. Our proposed method, outperforms existing methods with respect to accuracy, noise-tolerance, and size independence on a number of gold standard data set used in previous studies.


Supplementary Files (optional): 


  • Reviewed

Data repository URLs: 

<p> aliakbary/datasets.html</p><p></p><p> alexandre.arenas/data/welcome.htm</p>

Date of Submission: 

Saturday, March 28, 2020

Date of Decision: 

Wednesday, July 8, 2020

Nanopublication URLs:



Solicited Reviews:

1 Comment

Meta-Review by Editor

Reviewer 1 is positive about your paper but points to concepts that need to be clarified. Please note that the reviewer, like parts of the journal’s target audience, is from the social sciences. While these readers have a very solid background in formal methods, they are often unaware of disciplinary details. Please try to make your paper more accessible also to these readers, following the reviewer’s suggestions. Reviewer 2 is less positive but provides a detailed list of aspects that need to be clarified or defended. In addition, the reviewer points out that the analyses leading to Table 1 and Table 2 are not demonstrating the main point of the paper. Please address this with great care, in order to better drive home the message of your paper.

Michael Maes (