Expert detection within crowdsourcing forums is crucial for enhancing content accuracy and the decision-making process, as well as identifying knowledgeable individuals on a certain topic. While current expert detection methods identify knowledgeable users effectively, these methods depend on various user behavioral statistics, which may not be available for some crowdsourcing forums, and they only recognize expert users at the forum level on a general topic, such as data science, losing specific expertise on subtopics, such as clustering. In this paper, we propose a novel method that identifies experts for each subtopic using only the user's interactions and ratings, without relying on additional user behavioral statistics. We define responding/commenting to the same post/questions as the user interaction and create graphs and hypergraphs via these interactions. Then, we define a sheaf data structure on these networks to keep each user's tag-based knowledge and connections. To model information diffusion within the network considering subtopics, we develop a novel network diffusion model via sheaf Laplacian that captures subtopic-based knowledge diffusion over the network. Furthermore, we define a new centrality method, called sheaf Laplacian centrality, to measure a given user's expertise in each subtopic. Through extensive experiments conducted on four Stack Exchange networks, we show that our models outperform baseline models in detecting subtopic-based experts.

Telegram emerged as a crucial platform for both parties during the conflict between Russia and Ukraine. Per its minimal policies for content moderation, Pro-Kremlin narratives and potential misinformation were spread on Telegram, while anti-Kremlin narratives with related content were also propagated, such as war footage, troop movements, maps of bomb shelters, and air raid warnings. This paper presents a dataset of posts from both pro-Kremlin and anti-Kremlin Telegram channels, collected over a period spanning a year before and a year after the Russian invasion. The dataset comprises 404 pro-Kremlin channels with 4,109,645 posts and 114 anti-Kremlin channels with 1,117,768 posts. We provide details on the data collection process, processing methods, and dataset characterization. Lastly, we discuss the potential research opportunities this dataset may enable researchers across various disciplines.

Detecting influential users, called the influence maximization problem on social networks, is an important graph mining problem with many diverse applications such as information propagation, market advertising, and rumor controlling. There are many studies in the literature on the influential user detection problem in social networks. Although the current methods are successfully used in many different applications, they typically operate under the assumption that users are honest with one another, overlooking the role of deception in these environments. On the other hand, deception appears to be surprisingly common among humans within social networks. In this paper, we study the effect of deception in influence maximization on social networks. We begin by modeling deception within social networks and then explore opinion dynamics in these networks, incorporating deception through a recent opinion dynamics framework based on the sheaf Laplacian. We propose two methods for detecting influential nodes: sheaf Laplacian centrality and sheaf DFF centrality, both of which are designed to assess the influence of deception in the context of influence maximization. Our experimental results on synthetic and realworld networks suggest that liars are more influential than honest users in social networks. Our study underscores the potential danger and implications of deceptive practices in social network contexts.

Graph analytics has become instrumental in uncovering insights across various domains, specifically in social networks. It serves as a crucial tool for analyzing the relationship between users in different online platforms. In this research, we apply methods of social network analysis to examine the communication patterns among participants in an online forum recognized for far-right extremism. Our study demonstrates the actors’ relationships and activities through different aspects of applications over networks. In extensive analysis, we identify the influential actors and map their relationships throughout the course of 76 monthly networks. Moreover, we illustrate the evolution of networks over that period, and their connections with significant events. The findings of this analysis aim to understand the nature of interactions and networks, and to allow practitioners to take necessary precautions to mitigate far-right activities on various online platforms.

Predicting events, ranging from political unrest to disease outbreaks and criminal activities, stands as a pivotal task in proactively addressing emerging challenges. Despite the richness of textual data as a source for event detection, it is challenging to extract contextual information from documents due to their complex structure and the dynamic evolution of events. In response to this challenge, dynamic Graph Neural Networks (GNNs) have emerged as a promising tool for capturing the intricate patterns embedded within textual data graphs. Nevertheless, many models in this domain primarily rely on local node-level representations, overlooking the essential global graph-level context. However, both node-level and graph-level representations are critical for effective event prediction. Node-level representations provide insight into the local structure, while graph-level representations offer an understanding of the global structure and the evaluation of temporal graphs. To address these challenges, in this paper, we propose a Dynamic Graph Contrastive Learning (DyGCL) method for event prediction. Our model DyGCL first employs a local view encoder to effectively capture the local dynamic structure of input graphs as the evolving node representations. Then, it performs a global view encoder to perceive the hierarchical dynamic graph representation of the input graphs. Finally, the graph representations from both encoders, optimized via contrastive learning, are combined with an attention mechanism and utilized to predict future events. Our extensive experiments demonstrate that our proposed method outperforms the state-of-the-art methods for event prediction on six real-world datasets.

Modeling the interactions between drugs, targets, and diseases has significant implications for drug discovery, precision medicine and personalized treatments. Current computational approaches consider pairwise interaction, including drug-target or drug-disease interaction individually. On the other hand, within human metabolic systems, the interaction of drugs with protein targets in cells influences target activities. Moving beyond binary relationships and exploring tighter relationships together as triple is essential to understanding drugs' mechanism of action (MoAs). Moreover, considering the heterogeneity of drugs, targets, and diseases, along with their distinct characteristics, it is critical to model these complex interactions appropriately. To address these challenges, we develop a novel Heterogeneous Graph Triplet Attention Network (HeTan)by modeling the interconnectedness of all entities in a heterogeneous graph. HeTAN introduces a novel triplet message passing and triplet-wise attention mechanism within this heterogeneous graph structure. In contrast to focusing only on pairwise attention as the importance of an entity for the other, we define triplet attention to model the importance of pairs for the other in the drug-target-disease triplet prediction problem. We perform extensive experiments on real-world datasets and our results show that HeTAN outperforms several baselines, demonstrating its superior performance in uncovering novel drug-target-disease relationships.

Extracting meaningful features from sequences and devising effective similarity measures are vital for sequence data mining tasks, particularly sequence classification. While neural network models are commonly used to automatically learn sequence features, they are limited to capturing adjacent structural connection information and ignoring global, higher-order information between the sequences. To address these challenges, we propose a novel Hypergraph Attention Network model, namely Seq-HyGAN for sequence classification problems. To capture the complex structural similarity between sequence data, we create a novel hypergraph model by defining higher-order relations between subsequences extracted from sequences. Subsequently, we introduce a Sequence Hypergraph Attention Network that learns sequence features by considering the significance of subsequences and sequences to one another. Through extensive experiments, we demonstrate the effectiveness of our proposed Seq-HyGAN model in accurately classifying sequence data, outperforming several state-of-the-art methods by a significant margin.

Drug-Drug Interactions (DDIs) may hamper the functionalities of drugs, and in the worst scenario, they may lead to adverse drug reactions (ADRs). Predicting all DDIs is a challenging and critical problem. Most existing computational models integrate drug-centric information from different sources and leverage them as features in machine learning classifiers to predict DDIs. However, these models have a high chance of failure, especially for new drugs when all the information is not available. This paper proposes a novel Hypergraph Neural Network (HyGNN) model based on only the Simplified Molecular Input Line Entry System (SMILES) string of drugs, available for any drug, for the DDI prediction problem. To capture the drug chemical structure similarities, we create a hypergraph from drugs’ chemical substructures extracted from the SMILES strings. Then, we develop HyGNN consisting of a novel attention-based hypergraph edge encoder to get the representation of drugs as hyperedges and a decoder to predict the interactions between drug pairs. Furthermore, we conduct extensive experiments to evaluate our model and compare it with several state-of-the-art methods. Experimental results demonstrate that our proposed HyGNN model effectively predicts DDIs and impressively outperforms the baselines with a maximum F1 score, ROC-AUC, and PR-AUC of 94.61%, 98.69%, and 98.68%, respectively. Finally, we show that our models also work well for new drugs.

A drug-drug interaction (DDI) occurs when a drug is combined with other drug(s). DDIs have the potential to obstruct, increase, or diminish the intended impact of a drug or, in the worst-case scenario, induce an undesirable side effect. While it is critical to discover DDIs during clinical trials, it is impractical and expensive to detect all possible DDIs for a drug. Although several computational approaches for this problem have been developed, many of these methods need external biomedical knowledge that makes them difficult to generalize to drugs in early development phase. In this paper, we propose a novel method for predicting DDIs based on the vital chemical substructure of drugs extracted from their SMILES strings. We construct a graph that connects drugs based on their common functional chemical substructures. Furthermore, we apply different well-known graph neural network (GNN) methods to generate drug embeddings. Drug embeddings of individual drugs are concatenated to generate features of drug pairs. Finally, drug pair features are fed to different machine learning (ML) classifiers for DDI prediction. We evaluate our model on DrugBank dataset. Our result shows promising results and our model outperforms a baseline model based on different DDI representation creation methods.

With the increasing demand for digital privacy, content-hiding (or vault) apps are becoming popular among mobile phone users. Content-hiding apps affiliate to decoy apps. They are used for hiding photos, text, or videos and appear to have an interface very similar to commonly-used utility/productivity/gaming applications (for example, a calculator user interface). While these kinds of applications are convenient for people and let them hide private data, it raises concerns among app security researchers about their presence in legit and illicit app markets. It can also set a barrier for digital investigators, practitioners, victim service agencies, and the intelligence community since these apps are known to encrypt/delete data and make it unrecoverable. Such data could be anything ranging from contraband to classified data. Our research focuses on developing a fully automated Android Vault app Identification and Extraction system, primarily from the Google Play store. Through the feature extractions from description and images of applications followed by various machine learning and deep learning models, the system successfully identifies the content-hiding applications. The system can also automatically extract the user data from vault applications running on Android phones. To facilitate the advancement of research, we also keep an inventory of vault apps found in the Google Play store and offer to trace such apps even if they get removed from the Google Play store for security/other reasons. Our methodology and findings can be further extended to detect and classify content-hiding and anti-forensic apps in any Android app market and not limited to the Google Play store.

The rate of non-medical use of opioid drugs has increased markedly since the early 2000s. Due to this non-medical use, abusers suffer from different adverse effects that include physical and psychological problems. Many studies have been done to detect Drug Abuse (DA) events from social media data using machine learning and deep learning concepts. Moreover, Graph Neural Networks (GNNs) have recently become popular in text classification tasks due to their high accuracy and capability to handle complex structures. In this work, we collect drugs-related Twitter data (tweets) and build text graphs (corpus-level and document-level) to capture word-word, document-word, and document-document relations. Then we apply different GNN models on those text graphs and thus turn the text classification task into a node classification (for corpus-level graph) and graph classification (for document-level graph) task to detect DA events. Finally, we compare our graph-based DA detection models with different types of baselines models, including rule-based, traditional machine learning, and deep learning models. Our result shows graph-based models outperform the traditional machine learning and deep learning-based models.

The community search problem is defined as finding densely connected subgraphs in a large graph containing a given set of query nodes. One of the limitations of many current community search models is that detected communities may include irrelevant nodes, called the “free riders”. In this paper, we study the community search problem in the truss-based closest community model aimed to discover community for a given query set with avoiding the free rider effect. The Closest Truss Community (CTC) is a densely connected k-truss subgraph that contains the query nodes and has the smallest diameter. We use a greedy approach and propose a Truss EQuivalence based index graph (TEQ) that supports the search of closest trust community. To further improve the efficiency of the search, we create a maximum spanning tree of TEQ, as an index tree. We also propose an early pruning algorithm to improve the efficiency of free rider elimination. Extensive experiments on large real-world networks validate the efficiency and effectiveness of our algorithms over the state-of-the-art methods.

Recently, network embedding that encodes structural information of graphs into a vector space has become popular for network analysis. Although recent methods show promising performance for various applications, the huge size of graphs may hinder a direct application of the existing network embedding method to them. This paper presents NECL, a novel efficient Network Embedding method as answers to the following two questions: 1) Is there an ideal network Compression designed specifically for embedding? 2) Does the network compression significantly boost the network representation Learning? For the first problem, we propose a neighborhood similarity based graph compression method that compresses the input graph to a smaller graph without losing any/much information about its global structure and the local proximity of its vertices. For the second problem, we employ the compressed graph for network embedding instead of the original large graph to bring down the embedding cost. NECL is a general meta-strategy to improve the efficiency of all of the state-of-the-art graph embedding algorithms based on random walks, including DeepWalk and Node2vec, without losing their effectiveness. Extensive experiments validate the efficiency of NECL method that yields an average improvement of 23 -57% embedding time, including walking and learning time, without decreasing classification accuracy as evaluated on single and multi-label classification tasks on large real-world graphs.

Searching for relevant literature is a fundamental part of academic research. The search for relevant literature is becoming a more difficult and time-consuming task as millions of articles are published each year. As a solution, recommendation systems for academic papers attempt to help researchers find relevant papers quickly. This paper focuses on graph-based recommendation systems for academic papers using citation networks. This type of paper recommendation system leverages a graph of papers linked by citations to create a list of relevant papers. In this study, we explore recommendation systems for academic papers using citation networks incorporating citation relations. We define citation relation based on the number of times the origin paper cites the reference paper, and use this citation relation to measure the strength of the relation between the papers. We created a weighted network using citation relation as citation weight on edges. We evaluate our proposed method on a real-world publication data set, and conduct an extensive comparison with three state-of-the-art baseline methods. Our results show that citation network-based recommendation systems using citation weights perform better than the current methods.

Graph clustering is a fundamental problem in social network analysis, the goal of which is to group vertices of a graph into a series of densely knitted clusters with each cluster well separated from all the others. Classical graph clustering methods take advantage of the graph topology to model and quantify vertex proximity. With the proliferation of rich graph contents, such as user profiles in social networks, and gene annotations in protein interaction networks, it is essential to consider both the structure and content information of graphs for high-quality graph clustering. In this paper, we propose a graph embedding approach to clustering content-enriched graphs. The key idea is to embed each vertex of a graph into a continuous vector space where the localized structural and attributive information of vertices can be encoded in a unified, latent representation. Specifically, we quantify vertex-wise attribute proximity into edge weights, and employ truncated, attribute-aware random walks to learn the latent representations for vertices. We evaluate our attribute-aware graph embedding method in real-world attributed graphs, and the results demonstrate its effectiveness in comparison with state-of-the-art algorithms.

We consider the community search problem defined upon a large graph G : given a query vertex q in G , to find as output all the densely connected subgraphs of G , each of which contains the query v . As an online, query-dependent variant of the well-known community detection problem, community search enables personalized community discovery that has found widely varying applications in real-world, large-scale graphs. In this paper, we study the community search problem in the truss-based model aimed at discovering all dense and cohesive k -truss communities to which the query vertex q belongs. We introduce a novel equivalence relation, k-truss equivalence , to model the intrinsic density and cohesiveness of edges in k -truss communities. Consequently, all the edges of G can be partitioned to a series of k -truss equivalence classes that constitute a space-efficient, truss-preserving index structure, EquiTruss. Community search can be henceforth addressed directly upon EquiTruss without repeated, time-demanding accesses to the original graph, G , which proves to be theoretically optimal. In addition, EquiTruss can be efficiently updated in a dynamic fashion when G evolves with edge insertion and deletion. Experimental studies in real-world, large-scale graphs validate the efficiency and effectiveness of EquiTruss, which has achieved at least an order of magnitude speedup in community search over the state-of-the-art method, TCP-Index.