Graph-like Data Research Articles

Graph neural network based phishing account detection in Ethereum

Abstract In recent years, the widespread adoption of Ethereum-based transactions, such as cryptocurrencies and blockchain technologies, have revolutionized the way financial transactions are conducted. These decentralized and transparent systems offer numerous advantages, including enhanced security, immutability, and reduced transaction costs. However, alongside their benefits, Ethereum-based transactions have also attracted the attention of malicious actors seeking to exploit unsuspecting users through phishing scams. Phishing scams have thus become frequent in this scenario. Therefore, it is required to implement an effective and reliable phishing scam detection method. In this paper, we present the implementation of a highly efficient detection method by carrying out a graph-like data network formation, over which we then apply models that are based on graph neural networks like Magnet Link Prediction and Graph AutoEncoder Pathfinder Discovery Network Algorithm (GAE_PDNA). This helps in extracting useful information from the nodes of the graph. After relevant embeddings have been obtained, the classification of the phishing account is performed using AdaBoost classifier that helps in complex decision-making and detects the accounts related to the phishing scams. Our best model attains a precision of 0.99 and an F1 score of 0.99. Highlights

Beyond the individual: An improved telecom fraud detection approach based on latent synergy graph learning

The development of telecom technology not only facilitates social interactions but also inevitably provides the breeding ground for telecom fraud crimes. However, telecom fraud detection is a challenging task as fraudsters tend to commit co-fraud and disguise themselves within the mass of benign ones. Previous approaches work by unearthing differences in calling sequential patterns between independent fraudsters, but they may ignore synergic fraud patterns and oversimplify fraudulent behaviors. Fortunately, graph-like data formed by traceable telecom interaction provides opportunities for graph neural network (GNN)-based telecom fraud detection methods. Therefore, we develop a latent synergy graph (LSG) learning-based telecom fraud detector, named LSG-FD, to model both sequential and interactive fraudulent behaviors. Specifically, LSG-FD introduces (1) a multi-view LSG extractor to reconstruct synergy relationship-oriented graphs from the meta-interaction graph based on second-order proximity assumption; (2) an LSTM-based calling behavior encoder to capture the sequential patterns from the perspective of local individuals; (3) a dual-channel based graph learning module to alleviate the disassortativity issue (caused by the camouflages of fraudsters) by incorporating the dual-channel frequency filters and the learnable controller to adaptively aggregate high- and low-frequency information from their neighbors; (4) an imbalance-resistant model trainer to remedy the graph imbalance issue by developing a label-aware sampler. Experiment results on the telecom fraud dataset and another two widely used fraud datasets have verified the effectiveness of our model.

Improving Network Delay Predictions Using GNNs

Autonomous network management is crucial for Fifth Generation (5G) and Beyond 5G (B5G) networks, where a constantly changing environment is expected and network configuration must adapt accordingly. Modeling tools are required to predict the impact on performance (packet and delay loss) when new traffic demands arrives and when changes in routing paths are applied in the network. Mathematical analysis and network simulators are techniques for modeling networks but both have limitations, as the former provides low accuracy and the latter requires high execution times. To overcome these limitations, machine learning (ML) algorithms, and more specifically, graph neural networks (GNNs), are proposed for network modeling due to their ability to capture complex relationships from graph-like data while predicting network properties with high accuracy and low computational requirements. However, one of the main issues when using GNNs is their lack of generalization capability to larger networks, i.e., when trained in small networks (in number of nodes, paths length, links capacity), the accuracy of predictions on larger networks is poor. This paper addresses the GNN generalization problem by the use of fundamental networking concepts. Our solution is built from a baseline GNN model called RouteNet (developed by Barcelona Neural Networking Center-Universitat Politècnica de Catalunya (BNN-UPC)) that predicts the average delay in network paths, and through a number of simple additions significantly improves the prediction accuracy in larger networks. The improvement ratio compared to the baseline model is 101, from a 187.28% to a 1.828%, measured by the Mean Average Percentage Error (MAPE). In addition, we propose a closed-loop control context where the resulting GNN model could be potentially used in different use cases.

Vehicular-Network-Intrusion Detection Based on a Mosaic-Coded Convolutional Neural Network

With the development of Internet of Vehicles (IoV) technology, the car is no longer a closed individual. It exchanges information with an external network, communicating through the vehicle-mounted network (VMN), which, inevitably, gives rise to security problems. Attackers can intrude on the VMN, using a wireless network or vehicle-mounted interface devices. To prevent such attacks, various intrusion-detection methods have been proposed, including convolutional neural network (CNN) ones. However, the existing CNN method was not able to best use the CNN’s capability, of extracting two-dimensional graph-like data, and, at the same time, to reflect the time connections among the sequential data. Therefore, this paper proposed a novel CNN model, based on two-dimensional Mosaic pattern coding, for anomaly detection. It can not only make full use of the ability of a CNN to extract grid data but also maintain the sequential time relationship of it. Simulations showed that this method could, effectively, distinguish attacks from the normal information on the vehicular network, improve the reliability of the system’s discrimination, and, at the same time, meet the real-time requirement of detection.

Spatial Pooling Graph Convolutional Network for Hyperspectral Image Classification

Graph convolution networks (GCNs) have been applied in a variety of fields due to their powerful ability in processing graph-like data. However, the massive number of hyperspectral pixels makes it challenging to define general graph structures on hyperspectral images (HSIs). On the other hand, convolutional neural networks (CNNs) take in regular image regions with fixed square size, and have demonstrated impressive accuracy while being efficient in computation. Inspired by the classification framework of CNNs, we develop a GCN-based model that generates effective local spectral–spatial features for HSI classification. Specifically, graph convolutions are performed separately on every local region, which significantly limits the graph’s size. While graph convolution extracts features of every pixel, it does not reduce the number of them. To fuse suitable representations for the classification task, we develop a graph pooling operation to preserve classification-specific features and reduce redundant pixels. Based on local regions of HSIs, pooling in the graph domain is equivalent to spatial pooling in the spatial domain. The proposed method is thus named the spatial pooling graph convolutional network (SPGCN). Experimental results on several typical datasets demonstrated that the proposed SPGCN provides competitive results compared with other state-of-the-art CNN-based methods.

Power-Law Distributed Graph Generation With MapReduce

A graph generator is a tool which allows to create graph-like data whose structural properties are very similar to those found in real world networks. This paper presents two methods to generate graphs with power-law edge distribution based on the MapReduce processing model that can be easily implemented to run on top of Apache Hadoop. The proposed methods allow the generation of directed and undirected power-law distributed graphs without repeated edges. Our experimental evaluation shows that our methods are efficient and scalable in terms of both graph size and cluster capacity.

Suitability of Graph Database Technology for the Analysis of Spatio-Temporal Data

Every day large quantities of spatio-temporal data are captured, whether by Web-based companies for social data mining or by other industries for a variety of applications ranging from disaster relief to marine data analysis. Making sense of all this data dramatically increases the need for intelligent backend systems to provide realtime query response times while scaling well (in terms of storage and performance) with increasing quantities of structured or semi-structured, multi-dimensional data. Currently, relational database solutions with spatial extensions such as PostGIS, seem to come to their limits. However, the use of graph database technology has been rising in popularity and has been found to handle graph-like spatio-temporal data much more effectively. Motivated by the need to effectively store multi-dimensional, interconnected data, this paper investigates whether or not graph database technology is better suited when compared to the extended relational approach. Three database technologies will be investigated using real world datasets namely: PostgreSQL, JanusGraph, and TigerGraph. The datasets used are the Yelp challenge dataset and an ambulance response simulation dataset, thus combining real world spatial data with realistic simulations offering more control over the dataset. Our extensive evaluation is based on how each database performs under practical data analysis scenarios similar to those found on enterprise level.

Using Graph Databases

Data in High Energy Physics (HEP) usually consist of complext complex data structures stored in relational databases and files with internal schema. Such architecture exhibits many shortcomings, which could be fixed by migrating into Graph Database storage. The paper describes basic principles of the Graph Database together with an overview of existing standards and implementations. The usefulness and usability are demonstrated using the concrete example of the Event Index of the ATLAS experiment at LHC in two approaches as the full storage (all data are in the Graph Database) and meta-storage (a layer of schema-less graph-like data implemented on top of more traditional storage). The usability, the interfaces with the surrounding framework and the performance of those solutions are discussed. The possible more general usefulness for generic experiments’ storage is also discussed.

Mapping RDF Databases to Property Graph Databases

RDF triplestores and property graph databases are two approaches for data management which are based on modeling, storing, and querying graph-like data. In spite of such common principles, they present special features that complicate the task of database interoperability. While there exist some methods to transform RDF graphs into property graphs, and vice versa, they lack compatibility and a solid formal foundation. This paper presents three direct mappings (schema-dependent and schema-independent) for transforming an RDF database into a property graph database, including data and schema. We show that two of the proposed mappings satisfy the properties of semantics preservation and information preservation. The existence of both mappings allows us to conclude that the property graph data model subsumes the information capacity of the RDF data model.

Multiplex Network Inference With Sparse Tensor Decomposition for Functional Connectivity

Functional connectivity (FC) is a graph-like data structure commonly used by neuroscientists to study the dynamic behaviour of brain activity. However, these analyses rapidly become complex and time-consuming, since the number of connectivity components to be studied is quadratic with the number of electrodes. In this work, we address the problem of clustering FC into relevant ensembles of simultaneously activated components, yielding a multiplex network that reveals characteristic patterns of the epileptic seizures of a given patient. While $k-$ means is certainly the most popular method for data clustering, it is known to perform poorly on large dimensional data sets, and to be highly sensitive to noise. To overcome the so-called curse of dimensionality, we propose a new tensor decomposition to reduce the size of the data set formed by FC time-series recorded for several seizures, prior to apply $k$ -means. We propose an adapted procedure to infer a multiplex network from several FC time series, and we emphasise one particular variant that imposes sparsity constraint. Then, we conduct a real case study, applying the proposed sparse tensor decomposition to iEEG data to infer a multiplex network corresponding to the different stages of an epileptic seizure.

Efficient Distributed SPARQL Queries on Apache Spark

RDF is a widely-accepted framework for describing metadata in the web due to its simplicity and universal graph-like data model. Owing to the abundance of RDF data, existing query techniques are rendered unsuitable. To this direction, we adopt the processing power of Apache Spark to load and query a large dataset much more quickly than classical approaches. In this paper, we have designed experiments to evaluate the performance of several queries ranging from single attribute selection to selection, filtering and sorting multiple attributes in the dataset. We further experimented with the performance of queries using distributed SPARQL query on Apache Spark GraphX and studied different stages involved in this pipeline. The execution of distributed SPARQL query on Apache Spark GraphX helped us study its performance and gave insights into which stages of the pipeline can be improved. The query pipeline comprised of Graph loading, Basic Graph Pattern and Result calculating. Our goal is to minimize the time during graph loading stage in order to improve overall performance and cut the costs of data loading.

Seam: provably safe local edits on graphs

Algorithms that create and mutate graph data structures are challenging to implement correctly. However, verifying even basic properties of low-level implementations, such as referential integrity and memory safety, remains non-trivial. Furthermore, any extension to such a data structure multiplies the complexity of its implementation, while compounding the challenges in reasoning about correctness. We take a language design approach to this problem. We propose Seam, a language for expressing local edits to graph-like data structures, based on a relational data model, and such that data integrity can be verified automatically. We present a verification method that leverages an SMT solver, and prove it sound and precise (complete modulo termination of the SMT solver). We evaluate the verification capabilities of Seam empirically, and demonstrate its applicability to a variety of examples, most notably a new class of verification tasks derived from geometric remeshing operations used in scientific simulation and computer graphics. We describe our prototype implementation of a Seam compiler that generates low-level code, which can then be integrated into larger applications. We evaluate our compiler on a sample application, and demonstrate competitive execution time, compared to hand-written implementations.

Query Optimization Techniques in Graph Databases

Graph databases (GDB) have recently been arisen to overcome the limits of traditional databases for storing and managing data with graph-like structure. Today, they represent a requirement for many applications that manage graph-like data, like social networks. Most of the techniques, applied to optimize queries in graph databases, have been used in traditional databases, distribution systems... or they are inspired from graph theory. However, their reuse in graph databases should take care of the main characteristics of graph databases, such as dynamic structure, highly interconnected data, and ability to efficiently access data relationships. In this paper, we survey the query optimization techniques in graph databases. In particular, we focus on the features they have introduced to improve querying graph-like data.

S2RDF

RDF has become very popular for semantic data publishing due to its flexible and universal graph-like data model. Thus, the ever-increasing size of RDF data collections raises the need for scalable distributed approaches. We endorse the usage of existing infrastructures for Big Data processing like Hadoop for this purpose. Yet, SPARQL query performance is a major challenge as Hadoop is not intentionally designed for RDF processing. Existing approaches often favor certain query pattern shapes while performance drops significantly for other shapes. In this paper, we introduce a novel relational partitioning schema for RDF data called ExtVP that uses a semi-join based preprocessing, akin to the concept of Join Indices in relational databases, to efficiently minimize query input size regardless of its pattern shape and diameter. Our prototype system S2RDF is built on top of Spark and uses SQL to execute SPARQL queries over ExtVP. We demonstrate its superior performance in comparison to state of the art SPARQL-on-Hadoop approaches.

Shortest Path Finding Using a Star Algorithmand Minimum weight Node First Principle

The shortest paths, sets of paths with the shortest distance between a single initial (source) point and all other destination points, as well as between all pairs of points, are to be found. For each of these approaches, individual algorithms with specific features have been worked out over the past decades. Graph-like data appears in many applications, such as social networks, internet hyperlinks, roadmaps, etc. and in most cases, graphs are dynamic, evolving through time

Graph anonymization via metric embeddings: Using classical anonymization for graphs

With the unstoppable growth of applications requiring data to be represented as graphs, the interest for keeping this type of data private also grows. While many efforts have been made in order to anonymize tabular data, anonymizing graphs is a recent topic of research. Previous work on graph anonymization assumes that techniques proposed for tabular data are not suitable for anonymizing graph-like data because these ignore the topological properties of the graph. Because of this, they resort to new graph anonymization techniques that require devising new complex algorithms. In this paper, we show that, contrarily to this well-established assumption, it is possible to use classical anonymization techniques for graph anonymization. For this, we propose to embed the graph into a multidimensional vector space that approximately preserves the distances between any two vertices in the graph. This way, the graph can be represented as a list of vectors on which we can use tabular anonymization techniques. We show that, with our proposal, we can successfully anonymize graphs directly using the most common tabular techniques.

Computer-aided assembly planning for the diemaking industry

The paper describes a CAD-based approach to some management tasks related to the manufacture of stamping tools for car body parts. The proposed method generates assembly plans for draw dies and trim/pierce dies from their design information. Die assembly is a highly constrained process including a variety of part handling, measurement and surface finishing operations. As such, it escapes some critical assumptions underlying most generative planning methodologies in literature. Such complexity is faced through a comprehensive description model of the assembly process, which represents the space of all feasible operation sequences for any allowable die configuration within a predefined domain. Once a specific configuration has been retrieved from solid and surface CAD descriptions of die parts, the assembly model is instantiated into a graph-like data structure, which include only applicable operations and their precedence constraints. An instance describes operations at a sufficient detail level to support time estimation, process documentation and production scheduling. A prototype software tool derived from the assembly planning method has been tested in a real industrial context, in order to evaluate its potential impact on the efficiency of die manufacture.

A graph grammars based framework for querying graph-like data

The widespread use of graph-based models for representing data collections (e.g. object-oriented data, XML data, etc.) has stimulated the database research community to investigate the problem of defining declarative languages for querying graph-like databases. In this paper, a new framework for querying graph-like data based on graph grammars is proposed. The new paradigm allows us to verify structural properties of graphs and to extract sub-graphs. More specifically, a new form of query (namely graph query) is proposed, consisting in a particular graph grammar which defines a class of graphs to be matched on the graph representing the database. Thus, differently from path queries, the answer of a graph query is not just a set of nodes, but a subgraph, extracted from the input graph, which satisfies the structural properties defined by the graph grammar. Expressiveness and complexity of different forms of graph queries are discussed, and some practical applications are shown.

H-MDS: a new approach for interactive visualization with multidimensional scaling in the hyperbolic space

We introduce a novel projection-based visualization method for high-dimensional data sets by combining concepts from MDS and the geometry of the hyperbolic spaces. This approach hyperbolic multi-dimensional scaling (H-MDS) is a synthesis of two important concepts for explorative data analysis and visualization: (i) multi-dimensional scaling uses proximity or pair distance data to generate a low-dimensional, spatial presentation of the data; (ii) previous work on the “hyperbolic tree browser” demonstrated the extraordinary advantages for an interactive display of graph-like data in the two-dimensional hyperbolic space ( H 2) . In the new approach, H-MDS maps proximity data directly into the H 2 . This removes the restriction to “quasi-hierarchical”, graph-based data—a major limitation of (ii). Since a suitable distance function can convert all kinds of data to proximity (or distance-based) data, this type of data can be considered the most general. We review important properties of the hyperbolic space and, in particular, the circular Poincaré model of the H 2 . It enables effective human–computer interaction: by mouse dragging the “focus”, the user can navigate in the data without loosing the context. In H 2 the “fish-eye” behavior originates not simply by a non-linear view transformation but rather by extraordinary, non-Euclidean properties of the H 2 . Especially, the exponential growth of length and area of the underlying space makes the H 2 a prime target for mapping hierarchical and (now also) high-dimensional data. Several high-dimensional mapping examples including synthetic and real-world data are presented. Since high-dimensional data produce “ring”-shaped displays, we present methods to enhance the display by modulating the dissimilarity contrast. This is demonstrated for an application for unstructured text: i.e., by using multiple film critiques from news:rec.art.movies.reviews and www.imdb.com, each movie is placed within the H 2 —creating a “space of movies” for interactive exploration.

Feature Based Macroplanning Including Fixturing

Macroplanning is concerned with deciding on setups and sequences of operations by taking into account interactions between features and issues of fixturing. The algorithms presented in this paper use a set of features arranged in a graph-like data structure to determine the fixture configurations and setups. Operation sequences for the features and tool-paths are determined by separate modules working in conjunction with the macroplanner. Examples are presented to demonstrate the capability of the algorithms implemented in an integrated CAD/ CAM system called CyberCut.