Though research on the Semantic Web has progressed at a steady pace, its promise has yet to be realized. One major difficulty is that, by its very nature, the Semantic Web is a large, uncensored system to which anyone may contribute. This raises the question of how much credence to give each source. We cannot expect each user to know the trustworthiness of each source, nor would we want to assign top-down or global credibility values due to the subjective nature of trust. We tackle this problem by employing a web of trust, in which each user maintains trusts in a small number of other users. We then compose these trusts into trust values for all other users. The result of our computation is not an agglomerate "trustworthiness" of each user. Instead, each user receives a personalized set of trusts, which may vary widely from person to person. We define properties for combination functions which merge such trusts, and define a class of functions for which merging may be done locally while maintaining these properties. We give examples of specific functions and apply them to data from Epinions and our BibServ bibliography server. Experiments confirm that the methods are robust to noise, and do not put unreasonable expectations on users. We hope that these methods will help move the Semantic Web closer to fulfilling its promise.

Trust management for the Semantic Web

Graph representation learning serves as the core of important prediction tasks, ranging from product recommendation to fraud detection. Reallife graphs usually have complex information in the local neighborhood, where each node is described by a rich set of features and connects to dozens or even hundreds of neighbors. Despite the success of neighborhood aggregation in graph neural networks, task-irrelevant information is mixed into nodes’ neighborhood, making learned models suffer from sub-optimal generalization performance. In this paper, we present NeuralSparse, a supervised graph sparsification technique that improves generalization power by learning to remove potentially task-irrelevant edges from input graphs. Our method takes both structural and non-structural information as input, utilizes deep neural networks to parameterize sparsification processes, and optimizes the parameters by feedback signals from downstream tasks. Under the NeuralSparse framework, supervised graph sparsification could seamlessly connect with existing graph neural networks for more robust performance. Experimental results on both benchmark and private datasets show that NeuralSparse can yield up to 7.2% improvement in testing accuracy when working with existing graph neural networks on node classification tasks.

/pdf/robust-graph-representation-learning-via-neural-2hxgzk0xtr.pdf

Robust Graph Representation Learning via Neural Sparsification

Applying effective methods to identify important nodes in a complex network is highly invaluable. Recently, in a complex network, finding a powerful leader of the community to spread information quickly throughout the network is the concern of many researchers. In this paper, to identify influential nodes in a large and complex network, community-based mediator (CbM), which considers the entropy of a random walk from a node to each community is proposed as a metrics. CbM describes how the node is essential to connect two or more than two communities of the network. Correlations between CbM and other classical methods used to identify influential nodes are discussed. The performance of CbM is evaluated by susceptible-infected-recovered (SIR) model. In SIR model, the node is the most powerful node in the network, if the percentage of infected node is more while the node is used as the source of infection. Simulation results show that the proposed method performs better than the existing methods to spread information quickly and it can also introduce new influential nodes that other methods failed to identify.

Identifying Influential Nodes Based on Community Structure to Speed up the Dissemination of Information in Complex Network

Stochastic Diffusion Search, first incepted in 1989, belongs to the extended family of swarm intelligence algorithms. In contrast to many nature-inspired algorithms, stochastic diffusion search has a strong mathematical framework describing its behaviour and convergence. In addition to concisely exploring the algorithm in the context of natural swarm intelligence systems, this paper reviews various developments of the algorithm, which have been shown to perform well in a variety of application domains including continuous optimisation, implementation on hardware and medical imaging. This algorithm has also being utilised to argue the potential computational creativity of swarm intelligence systems through the two phases of exploration and exploitation.

Stochastic Diffusion Search Review

Report On The Investigation Into Russian Interference In The 2016 Presidential Election. Volumes I & II. (Redacted version of 4/18/2019)

The study of a large real world network in terms of graph sample representation constitutes a very powerful and useful tool in several domains of network analysis. This is the motivation that has led the work of this paper towards the development of a new graph sampling algorithm. Previous research in this area proposed simple processes such as the classic Random Walk algorithm, Random node and Random edge sampling and has evolved during the last decade to more advanced graph exploration approaches such as Forest Fire and Frontier sampling. In this paper, we propose a new graph sampling method based on edge selection. In addition, we crawled Facebook collecting a large dataset consisting of 10 million users and 80 million users' relations, which we have also used to evaluate our sampling algorithm. The experimental evaluation on several datasets proves that our approach preserves several properties of the initial graphs, leading to representative samples and outperforms all the other approaches.

Rank degree: an efficient algorithm for graph sampling

Graph sampling is a widely used procedure in social network analysis, has attracted great interest in the scientific community and is considered as a very powerful and useful tool in several domains of network analysis. Apart from initial research in this area, which has proposed simple processes such as the classic Random Walk algorithm, Random Node and Random Edge sampling, during the last decade, more advanced graph sampling approaches have been emerged. In this paper, we extensively study the properties of a newly proposed method, the Rank Degree method, which leads to representative graph subgraphs. The Rank Degree is a novel graph exploration method which significantly differs from other existing methods in the literature. The novelty of the Rank Degree lies on the fact that its core methodology corresponds to a deterministic graph exploration; one specific variation corresponds to a number of parallel deterministic traverses that explore the graph. We perform extensive experiments on twelve real-world datasets of a different type, using a variety of measures and comparing our method with Forest Fire, Metropolis Hastings Random Walk and Metropolis Hastings. We provide strong evidence that our approach leads to highly efficient graph sampling; the generated samples preserve several graph properties, to a large extent.

Deterministic graph exploration for efficient graph sampling

This paper represents the SECONDO framework to assist organizations with decisions related to cybersecurity investments and cyber-insurance. The platform supports cybersecurity and cyber-insurance decisions by implementing and integrating a number of software components. SECONDO operates in three distinct phases: (i) cyber-physical risk assessment and continuous monitoring; (ii) investment-driven optimized cyber-physical risk control; and (iii) blockchain-enabled cyber-insurance contract preparation and maintenance. Insurers can leverage SECONDO functionalities to actively participate in the management of cyber-physical risks of a shipping company to reduce their insured risk.

SECONDO: A platform for cybersecurity investments and cyber insurance decisions

In this paper we investigate the correlation between the social network communities as defined by a community detection algorithm and the Facebook pages annotated as Likes by its users Our goal is twofold First, we aim to examine the relation between the underlined social dynamic, as expressed indirectly by a community structure, with the users' characteristics represented by Likes Second, to valuate the outcome of the community detection algorithm To the best of our knowledge this is the first study of the correlation between community structure and users' Likes in Facebook Using a standard crawling method, such as the Breadth First Search, we collect: a) several snapshots of a sub graph of Facebook, b) the users' Likes in Web and Facebook pages and c) the pages' categories as classified by the owner of the page We study several graph samples along with their community structure The experimental results demonstrate that in the case of users' Likes, the correlation ranges from small to medium between communities and the whole population, while it is even smaller between communities Moreover, there is a high correlation in terms of Likes' categories between the different communities and between communities and the whole population This fact proves that Likes constitute a criterion of distinction among the communities and verifies the intuition that lead us towards this research

Discovering Correlation between Communities and Likes in Facebook

The problem of identifying the influential spreaders - the important nodes - in a real world network is of high importance due to its theoretical interest as well as its practical applications, such as the acceleration of information diffusion, the control of the spread of a disease and the improvement of the resilience of networks to external attacks. In this paper, we propose a graph exploration sampling method that accurately identifies the influential spreaders in a complex network, without any prior knowledge of the original graph, apart from the collected samples/subgraphs. The method explores the graph, following a deterministic selection rule and outputs a graph sample - the set of edges that have been crossed. The proposed method is based on a version of Rank Degree graph sampling algorithm. We conduct extensive experiments in eight real world networks by simulating the susceptible-infected-recovered (SIR) and susceptible-infected-susceptible (SIS) epidemic models which serve as ground truth identifiers of nodes spreading efficiency. Experimentally, we show that by exploring only the 20% of the network and using the degree centrality as well as the k-core measure, we are able to identify the influential spreaders with at least the same accuracy as in the full information case, namely, the case where we have access to the original graph and in that graph, we compute the centrality measures. Finally and more importantly, we present strong evidence that the degree centrality - the degree of nodes in the collected samples - is almost as accurate as the k-core values obtained from the original graph.

/pdf/a-graph-exploration-method-for-identifying-influential-2mbeofsx2y.pdf

Nikos Salamanos

Papers

Rank degree: an efficient algorithm for graph sampling

Deterministic graph exploration for efficient graph sampling

SECONDO: A platform for cybersecurity investments and cyber insurance decisions

Discovering Correlation between Communities and Likes in Facebook

A graph exploration method for identifying influential spreaders in complex networks