Node Labels Research Articles

New strategy for anti-loop formulations

This paper presents a strategy based on binary labelling of nodes for the creation of anti-loop formulations from existing strategies. This strategy prevents by default the formation of odd cycles, therefore it can have important role in iterative procedures based on generating subtour elimination constraints. It can also be used to modify the classic strategies used in problems associated to graphs. In this paper we focus on this last application. The behavior of this strategy is analyzed with two problems associated with graphs, the Asymmetric Traveling Salesman Problem (ATSP) and the Steiner Problem, where two configurations that modify the Miller-Tucking-Zemlig proposal to avoid cycles are compared. The experimental analysis shows that this strategy keep a good convergence, highlighting its use for the Steiner problem.

Further investigation on the super classical mean labeling of graphs obtained from paths

Consider the graph G, with the injection Ω from node set to the first p + q natural numbers. Let us assume that the ceiling function of the classical average of the node labels of the end nodes of each link is the induced link assignment Ω *. If the union of range of Ω of node set and the range of Ω * of link set is all the first p + q natural numbers, then Ω is called a classical mean labeling. A super classical mean graph is a graph with super classical mean labeling. In this research effort, we attempted to address the super classical meanness of graphs generated by paths and those formed by the union of two graphs.

When Causal Inference Meets Graph Machine Learning

Graphs (i.e., networks) are ubiquitous in daily life, as they can effectively model a plethora of real-world systems with connected units, such as social networks and biological networks. Recent years have witnessed rapid development in graph-based machine learning (GML) in various high-impact domains. Currently, the mainstream GML methods are based on statistical learning, e.g., utilizing the statistical correlations between node features, graph structure, and labels for node classification. However, statistical learning has been widely criticized for only capturing the superficial relations between variables in the data system, and consequently, rendering the lack of trustworthiness in real-world applications. Therefore, it is crucial to understand the causality in the data system and the learning process. Causal inference is the discipline that investigates the causality inside a system, for example, to identify and estimate the causal effect of a certain treatment (e.g., wearing a face mask) on an important outcome (e.g., COVID-19 infection). Involving the concepts and philosophy of causal inference in ML methods is often considered significant for human-level intelligence and can serve as the foundation of artificial intelligence (AI). However, most traditional causal inference studies rely on strong assumptions, and focus on independent and identically distributed (i.i.d.) data, while causal inference on graphs is faced with many barriers. Therefore, we aim to bridge the gap between causal inference and GML.

Learning Efficient and Robust Multi-Agent Communication via Graph Information Bottleneck

Efficient communication learning among agents has been shown crucial for cooperative multi-agent reinforcement learning (MARL), as it can promote the action coordination of agents and ultimately improve performance. Graph neural network (GNN) provide a general paradigm for communication learning, which consider agents and communication channels as nodes and edges in a graph, with the action selection corresponding to node labeling. Under such paradigm, an agent aggregates information from neighbor agents, which can reduce uncertainty in local decision-making and induce implicit action coordination. However, this communication paradigm is vulnerable to adversarial attacks and noise, and how to learn robust and efficient communication under perturbations has largely not been studied. To this end, this paper introduces a novel Multi-Agent communication mechanism via Graph Information bottleneck (MAGI), which can optimally balance the robustness and expressiveness of the message representation learned by agents. This communication mechanism is aim at learning the minimal sufficient message representation for an agent by maximizing the mutual information (MI) between the message representation and the selected action, and simultaneously constraining the MI between the message representation and the agent feature. Empirical results demonstrate that MAGI is more robust and efficient than state-of-the-art GNN-based MARL methods.

Structural Information Enhanced Graph Representation for Link Prediction

Link prediction is a fundamental task of graph machine learning, and Graph Neural Network (GNN) based methods have become the mainstream approach due to their good performance. However, the typical practice learns node representations through neighborhood aggregation, lacking awareness of the structural relationships between target nodes. Recently, some methods have attempted to address this issue by node labeling tricks. However, they still rely on the node-centric neighborhood message passing of GNNs, which we believe involves two limitations in terms of information perception and transmission for link prediction. First, it cannot perceive long-range structural information due to the restricted receptive fields. Second, there may be information loss of node-centric model on link-centric task. In addition, we empirically find that the neighbor node features could introduce noise for link prediction. To address these issues, we propose a structural information enhanced link prediction framework, which involves removing the neighbor node features while fitting neighborhood graph structures more focused through GNN. Furthermore, we introduce Binary Structural Transformer (BST) to encode the structural relationships between target nodes, complementing the deficiency of GNN. Our approach achieves remarkable results on multiple popular benchmarks, including ranking first on ogbl-ppa, ogbl-citation2 and Pubmed.

Revisiting Graph-Based Fraud Detection in Sight of Heterophily and Spectrum

Graph-based fraud detection (GFD) can be regarded as a challenging semi-supervised node binary classification task. In recent years, Graph Neural Networks (GNN) have been widely applied to GFD, characterizing the anomalous possibility of a node by aggregating neighbor information. However, fraud graphs are inherently heterophilic, thus most of GNNs perform poorly due to their assumption of homophily. In addition, due to the existence of heterophily and class imbalance problem, the existing models do not fully utilize the precious node label information. To address the above issues, this paper proposes a semi-supervised GNN-based fraud detector SEC-GFD. This detector includes a hybrid filtering module and a local environmental constraint module, the two modules are utilized to solve heterophily and label utilization problem respectively. The first module starts from the perspective of the spectral domain, and solves the heterophily problem to a certain extent. Specifically, it divides the spectrum into various mixed-frequency bands based on the correlation between spectrum energy distribution and heterophily. Then in order to make full use of the node label information, a local environmental constraint module is adaptively designed. The comprehensive experimental results on four real-world fraud detection datasets denote that SEC-GFD outperforms other competitive graph-based fraud detectors. We release our code at https://github.com/Sunxkissed/SEC-GFD.

Methodology of ICG marking technique in modern pelvic surgery: applied aspects (Literature review)

The article is devoted to a review of scientific publications on the use of the Indocyanine green (ICG) labeling technique in modern pelvic surgery. This dye was developed by Kodak Research Laboratories in 1955 and approved for clinical use in 1959, but has only recently come into widespread use. ICG-enhanced near-infrared fluorescence is introduced to improve visualization and provide more detailed anatomical localization during surgery.The dye binds to blood plasma proteins, it is quickly excreted from the organism and is non-toxic and has no found complications after its administration. Many studies have been conducted on the use of this technique: visualization of sentinel lymph nodes (SLN) in various oncogynecological diseases; visualization of ureters, lymphatic drainage, tissue perfusion, vessels of various sizes, colorectal blood circulation, etc. The high detection rate of SLN (unilateral and bilateral) with ICG compared with other dyes such as radiocolloid and patent blue suggests its effectiveness.The condition of SLN has an important prognostic and therapeutic value, based on the obtained results, a more correct further adjuvant treatment can be prescribed. A sentinel lymph node labeling procedure was performed as an alternative to systemic lymphadenectomy to assess the status of lymphovascular invasion. The use of ICG-marking allows to reduce the risk of intraoperative and postoperative complications. Many experimental studies have been conducted in relation to the assessment of uterine vascularization, preservation of individual vessels, their identification and measurement of blood flow.The article provides data from scientists on the use of the ICG-marking technique in endometrial, cervical, vulva, vaginal, ovarian cancer, as well as in endometriosis and for visualization of the ureters. This technique causes active discussions in the scientific community, requires further research and should be more widely used in various directions of surgical interventions.

Evolutionary multi-objective overlapping community detection based on fusion of internal and external connectivity and correction of node intimacy

In the field of community detection, node attribute information plays an important role in community division. Existing methods use topology structure and node attribute information to discover non-overlapping communities. However, so far, attribute information has not been fully utilized in overlapping community detection. To address this, we propose a new overlapping community detection method called “evolutionary multi-objective overlapping community detection based on Fusion of internal and external Connectivity and Correction of Node Intimacy” (FCCNI). Firstly, we propose a fusion strategy based on internal and external connectivity, which integrates some communities with sparse intra-connections and dense inter-connections. This automatically determines, reconfirms, and corrects the number of communities. Secondly, a function is designed to calculate the intimacy between nodes, and the node label with the highest intimacy is selected to correct the current wrong node. The correction strategy is used in two stages of initialization and multi-objective evolution to obtain a more accurate node label. Finally, a method that considers not only the connections of the community, but also the node attribute, is designed to obtain the overlapping community indirectly from the non-overlapping community. The experimental results on five real-life networks and four classical synthetic networks show that FCCNI achieves better overlapping community division, compared with six state-of-the-art comparison algorithms from the literature.

A Scalable Method for Readable Tree Layouts.

Large tree structures are ubiquitous and real-world relational datasets often have information associated with nodes (e.g., labels or other attributes) and edges (e.g., weights or distances) that need to be communicated to the viewers. Yet, scalable, easy to read tree layouts are difficult to achieve. We consider tree layouts to be readable if they meet some basic requirements: node labels should not overlap, edges should not cross, edge lengths should be preserved, and the output should be compact. There are many algorithms for drawing trees, although very few take node labels or edge lengths into account, and none optimizes all requirements above. With this in mind, we propose a new scalable method for readable tree layouts. The algorithm guarantees that the layout has no edge crossings and no label overlaps, and optimizes one of the remaining aspects: desired edge lengths and compactness. We evaluate the performance of the new algorithm by comparison with related earlier approaches using several real-world datasets, ranging from a few thousand nodes to hundreds of thousands of nodes. Tree layout algorithms can be used to visualize large general graphs, by extracting a hierarchy of progressively larger trees. We illustrate this functionality by presenting several map-like visualizations generated by the new tree layout algorithm.

Graph Neural Network with curriculum learning for imbalanced node classification

Graph Neural Network (GNN) stands as an emerging methodology for graph-based learning tasks, particularly for node classification. This study elucidates the susceptibility of GNN to discrepancies arising from imbalanced node labels. Conventional solutions for imbalanced classification, such as resampling, falter in node classification task, primarily due to their negligence of graph structure. Worse still, they often exacerbate the model’s inclination towards overfitting or underfitting, especially in the absence of adequate priori knowledge. To circumvent these limitations, we introduce a novel Graph Neural Network framework with Curriculum Learning (GNN-CL). This framework integrates two pivotal components. Initially, leveraging the principles of smoothness and homophily, we endeavor to procure dependable interpolation nodes and edges via adaptive graph oversampling. For another, we combine the Graph Classification Loss with the Metric Learning Loss, thereby refining the spatial proximity of nodes linked to the minority class in the feature space. Drawing inspiration from curriculum learning, the parameters of these components are dynamically modulated during the training phase to accentuate generalization and discrimination capabilities. Comprehensive evaluations on several widely used graph datasets affirm the superiority of our proposed model, which consistently outperforms the existing state-of-the-art methods.

Noise-resistant graph neural networks with manifold consistency and label consistency

Previous graph neural networks (GNNs) consider the consistency in either the feature space or the label space to deal with the issue of noisy labels by ignoring their complementary information. To address this issue, in this paper, we propose a new method to simultaneously consider them in a unified framework. To do this, we design a manifold regularization to enable the embeddings of every node and its neighbors similar. Moreover, we also design a label consistency regularization to make pseudo labels of augmentation data consistent to predicted labels of unlabeled nodes, as well as make predicted labels of every unlabeled node consistent to pseudo labels of its neighborhood. Furthermore, we prove that these two regularizations provide complementary information to each other for handling label noise. Finally, experimental results on multiple real-world datasets demonstrate that our proposed method achieves superior results in terms of node classification with different label noise ratios, compared to SOTA methods.

Structure Seer – a machine learning model for chemical structure elucidation from node labelling of a molecular graph

We introduce a graph convolution based machine-learning model which provides a quantitative probabilistic prediction on the atom connectivity based on the composition of the molecule along with a list of atom-attributed isotropic shielding constants.

TAMNR: a network embedding learning algorithm using text attention mechanism.

Because many existing algorithms are mainly trained based on the structural features of the networks, the results are more inclined to the structural commonality of the networks. These algorithms ignore the rich external information and node attributes (such as node text content, community and labels, etc.) that have important implications for network data analysis tasks. Existing network embedding algorithms considering text features usually regard the co-occurrence words in the node's text, or use an induced matrix completion algorithm to factorize the text feature matrix or the network structure feature matrix. Although this kind of algorithm can greatly improve the network embedding performance, they ignore the contribution rate of different co-occurrence words in the node's text. This article proposes a network embedding learning algorithm combining network structure and co-occurrence word features, also incorporating an attention mechanism to model the weight information of the co-occurrence words in the model. This mechanism filters out unimportant words and focuses on important words for learning and training tasks, fully considering the impact of the different co-occurrence words to the model. The proposed network representation algorithm is tested on three open datasets, and the experimental results demonstrate its strong advantages in node classification, visualization analysis, and case analysis tasks.

Network Dynamic GCN Influence Maximization Algorithm With Leader Fake Labeling Mechanism

Influence maximization is an important technique for its significant value on various social network applications, such as viral marketing, advertisement, and recommendation. Traditional heuristic algorithms for influence maximization suffer from different issues, such as accuracy descent and information loss during network exploration. Besides, recently proposed deep learning-based approaches cannot extract the in-depth structural information of social networks well. In light of these problems, we propose a novel heuristic method called the network dynamic GCN influence maximization algorithm based on the leader fake labeling mechanism. To exploit the in-depth network topology information for the influence maximization task, we design a network dynamic GCN that owns adaptive layer numbers in terms of different network scales to obtain node representations. Then, considering that there are few labels in the network or the labels are irrelevant to the task of influence maximization, we establish a leader fake labeling mechanism to automatically generate node labels that are helpful to seed nodes selecting for model training. Finally, a heuristic method based on the Mahalanobis distance is developed to quickly select influential seed nodes with the learned node representations. Three real-world datasets are used in our experiments, and the experimental results demonstrate that our algorithm has a better performance for seed set identification under the premise of high efficiency compared with some latest heuristic influence maximization algorithms.

Label-guided graph contrastive learning for semi-supervised node classification

Semi-supervised node classification is a task of predicting the labels of unlabeled nodes using limited labeled nodes and numerous unlabeled nodes. Recently, Graph Neural Networks (GNNs) have achieved remarkable success in this task. However, GNNs typically have shallow architectures and only consider labeled nodes and their low-order neighbors during training. As a result, the supervision signals from the massive unlabeled nodes remain underutilized. To address this limitation, graph contrastive learning has been applied in semi-supervised node classification, which pulls together positive nodes and pushes away negative nodes in the embedding space. Nevertheless, existing node-level contrastive learning methods usually sample the same node from two augmented views as positive nodes and all different nodes as negative nodes. Consequently, many semantically similar nodes are not sampled as positive nodes but are mistakenly sampled as negative nodes. To tackle this issue, we propose a novel Label-guided Graph Contrastive Learning (LGGCL) training algorithm for semi-supervised node classification. Specifically, we first propose a Label-guided Graph Contrastive Learning framework as the basis of LGGCL training algorithm. Then we incorporate a self-checking mechanism based on deep clustering to ensure the authenticity of the sampled positive nodes. Moreover, we design a reweighting strategy based on the probability distribution of the anchor node to enhance the effect of hard negative nodes. Finally, experimental results on various graph benchmarks demonstrate the superiority of our LGGCL.

GraphHop: An Enhanced Label Propagation Method for Node Classification.

A scalable semisupervised node classification method on graph-structured data, called GraphHop, is proposed in this work. The graph contains all nodes' attributes and link connections but labels of only a subset of nodes. Graph convolutional networks (GCNs) have provided superior performance in node label classification over the traditional label propagation (LP) methods for this problem. Nevertheless, current GCN algorithms suffer from a considerable amount of labels for training because of high model complexity or cannot be easily generalized to large-scale graphs due to the expensive cost of loading the entire graph and node embeddings. Besides, nonlinearity makes the optimization process a mystery. To this end, an enhanced LP method, called GraphHop, is proposed to tackle these problems. GraphHop can be viewed as a smoothening LP algorithm, in which each propagation alternates between two steps: label aggregation and label update. In the label aggregation step, multihop neighbor embeddings are aggregated to the center node. In the label update step, new embeddings are learned and predicted for each node based on aggregated results from the previous step. The two-step iteration improves the graph signal smoothening capacity. Furthermore, to encode attributes, links, and labels on graphs effectively under one framework, we adopt a two-stage training process, i.e., the initialization stage and the iteration stage. Thus, the smooth attribute information extracted from the initialization stage is consistently imposed in the propagation process in the iteration stage. Experimental results show that GraphHop outperforms state-of-the-art graph learning methods on a wide range of tasks in graphs of various sizes (e.g., multilabel and multiclass classification on citation networks, social graphs, and commodity consumption graphs).

Decentralizing the lightning network: a score-based recommendation strategy for the autopilot system

The fundamental objective of the Lightning Network is to establish a decentralized platform for scaling the Bitcoin network and facilitating high-throughput micropayments. However, this network has gradually deviated from its decentralized topology since its operational inception, and its resources have quickly shifted towards centralization. The evolution of the network and the changes in its topology have been critically reviewed and criticized due to its increasing centralization. This study delves into the network’s topology and the reasons behind its centralized evolution. We explain the incentives of various participating nodes in the network and propose a score-based strategy for the Lightning Autopilot system, which is responsible for automatically establishing new payment channels for the nodes joining the network. Our study demonstrates that utilizing the proposed strategy could significantly aid in reducing the network’s centralization. This strategy is grounded in qualitative labeling of network nodes based on topological and protocol features, followed by the creation of a scoring and recommendation model. Results of the experiments indicate that in the evolved network using the proposed strategy, concentration indicators such as the Gini coefficient can decrease by up to 17%, and channels ownership of the top 1% of hubs decrease by 27% compared to other autopilot strategies. Moreover, through simulated targeted attacks on hubs and channels, it is shown that by adopting the proposed strategy, the network’s resilience is increased compared to the existing autopilot strategies for evolved networks. The proposed method from this research can also be integrated into operational Lightning clients and potentially replace the current recommendation methods used in Lightning Autopilot.

Elastic founder graphs improved and enhanced

Indexing labeled graphs for pattern matching is a central challenge of pangenomics. Equi et al. (2022) [14] developed the Elastic Founder Graph (EFG) representing an alignment of m sequences of length n, drawn from alphabet Σ plus the special gap character: the paths spell the original sequences or their recombination. By enforcing the semi-repeat-free property, the EFG admits a polynomial-space index for linear-time pattern matching, breaking through the conditional lower bounds on indexing labeled graphs (Equi et al. [13]). In this work, we improve the space of the EFG index answering pattern matching queries in linear time, from linear in the length of all strings spelled by three consecutive node labels, to linear in the size of the edge labels. Then, we develop linear-time construction algorithms optimizing for different metrics: we improve the existing linearithmic construction algorithms to O(mn), by solving the novel exclusive ancestor set problem on trees; we propose, for the simplified gapless setting, an O(mn)-time solution minimizing the maximum block height, that we generalize by substituting block height with prefix-aware height. Finally, to show the versatility of the framework, we develop a BWT-based EFG index and study how to encode and perform document listing queries on a set of paths of the graphs, reporting which paths present a given pattern as a substring. We propose the EFG framework as an improved and enhanced version of the framework for the gapless setting, along with construction methods that are valid in any setting concerned with the segmentation of aligned sequences.

Community perceptions about causes of suicide among young men in Botswana: an analysis based on fuzzy cognitive maps

ABSTRACT Suicide is common in Botswana, particularly among young men. Fuzzy cognitive mapping (FCM) can support participatory research by depicting local stakeholder knowledge about causes of health outcomes. This study used FCM to explore local perceptions about causes of suicide among young men in rural communities close to the capital, Gaborone. In nine sessions, groups of young men, young women, older men, and older women separately mapped their knowledge of factors related to suicide among young men (46 people in total). Two trained facilitators, fluent in the local language, led the group sessions. The maps depicted risk and protective factors as nodes connected by arrows to show causal relationships. Participants also ranked the strength of each link on a scale of one (weakest) to five (strongest). Fuzzy transitive closure calculated the maximum influence of each factor, taking into account all other influences on the map. We combined maps by different stakeholders and grouped the 130 unique factors across the maps into 17 broader categories which emerged from an inductive thematic analysis of all the node labels. Financial difficulties, relationship problems, and family issues were the strongest categories of perceived causes of suicide by young men. Mental health problems played an intermediary role between more distal causes and suicide. There were differences in maps of different gender and age groups, but the strongest influences were consistent across groups. Young women, but not young men, identified men’s lack of self-esteem as a strong cause of suicide. The FCM findings offer a starting point for community discussions to seek local solutions to youth suicide.

A Multi-Agent System in Education Facility Design

This paper deals with a multi-agent system which supports the designer in solving complex design tasks. The behaviour of design agents is modelled by sets of grammar rules. Each agent uses a graph grammar or a shape grammar and a database of facts concerning the subtask it is responsible for. The course of the design process is determined by the interaction between specialised agents. Space layouts of designs are represented by attributed graphs encoding both topological structures and semantic properties of solutions. The agents work in parallel on the common graph, independently generating layouts of different design components while specified node labels evoke agents using shape grammars. The agents’ cooperation allows them to combine a form-oriented approach with a functional-structural one in the design process, where the agents generate the general 3D form of the object based on design requirements together with the space layout based on the functional aspects of the solution. Based on the given design criteria, the agents search for admissible solutions within the design space that constitutes their operating environment. The proposed approach is illustrated by the example of designing kindergarten facilities.