Edges Of Graph Research Articles

GDMol: Generative Double-Masking Self-Supervised Learning for Molecular Property Prediction.

Effective molecular feature representation is crucial for drug property prediction. Recent years have seen increased attention on graph neural networks (GNNs) that are pre-trained using self-supervised learning techniques, aiming to overcome the scarcity of labeled data in molecular property prediction. Traditional GNNs in self-supervised molecular property prediction typically perform a single masking operation on the nodes and edges of the input molecular graph, masking only local information and insufficient for thorough self-supervised training. Hence, we propose a model for molecular property prediction based on generative double-masking self-supervised learning, termed as GDMol. This integrates generative learning into the self-supervised learning framework for latent representation, and applies a second round of masking to these latent representations, enabling the model to better capture global information and semantic knowledge of the molecules for a richer, more informative representation, thereby achieving more accurate and robust molecular property prediction. Our experiments on 5 datasets demonstrated superior performance of GDMol in predicting molecular properties across different domains. Moreover, we used the masking operation to traverse through the gradient changes of each node, the magnitude and sign of which reflect the positive and negative contribution respectively of the local structure in the molecule to the prediction outcome. This in-depth interpretative analysis not only enhances the model's interpretability, but also provides more targeted insights and direction for optimizing drug molecules. In summary, this research offers novel insights on improving molecular property prediction tasks, and paves the way for further research on the application of generative learning and self-supervised learning in the field of chemistry.

MGHCN: Multi-graph structures and hypergraph convolutional networks for traffic flow prediction

Accurate and timely traffic flow predictions are essential for effective traffic management and congestion reduction. However, most traditional prediction methods often fail to capture the complex dynamics and correlations within traffic flows due to insufficient processing of spatiotemporal data. Specifically, these methods struggle to integrate and analyze the multi-layered spatial and temporal interactions inherent in traffic data, leading to suboptimal prediction accuracy and robustness. To address this limitation, this paper presents a Multi-Graph Structures and Hypergraph Convolutional Network (MGHCN) that combines diverse graphs and hypergraphs. The MGHCN simplifies the predictive framework by integrating key components that improve its robustness and accuracy. One of the most critical components is the dual hypergraph structure, which captures edge correlations by converting traditional graph edges into hypergraph nodes. To better capture the spatiotemporal correlation of traffic data, a Graph Convolutional Network (GCN) is employed to analyze these hypergraphs in depth. Finally, a novel adjacency matrix and a dynamic graph module are used to accurately simulate interactions between spatiotemporal features, thereby enhancing the accuracy and robustness of predictions. Experimental validation on four distinct real-world traffic datasets shows that MGHCN outperforms existing state-of-the-art traffic prediction methods.

MosGraphGen: A novel tool to generate multi-omics signaling graphs to facilitate integrative and interpretable graph AI model development

Abstract Motivation Multi-omics data, i.e., genomics, epigenomics, transcriptomics, proteomics, characterize cellular complex signaling systems from multi-level and multi-view and provide a holistic view of complex cellular signaling pathways. However, it remains challenging to integrate and interpret multi-omics data for mining critical biomarkers. Graph AI models have been widely used to analyze graph-structure datasets, and are ideal for integrative multi-omics data analysis because they can naturally integrate and represent multi-omics data as a biologically meaningful multi-level signaling graph and interpret multi-omics data via graph node and edge ranking analysis. Nevertheless, it is non-trivial for graph-AI model developers to pre-analyze multi-omics data and convert the data into biologically meaningful graphs, which can be directly fed into graph-AI models. Results To resolve this challenge, we developed mosGraphGen (multi-omics signaling graph generator), generating Multi-omics Signaling graphs (mos-graph) of individual samples by mapping multi-omics data onto a biologically meaningful multi-level background signaling network with data normalization by aggregating measurements and aligning to the reference genome. With mosGraphGen, AI model developers can directly apply and evaluate their models using these mos-graphs. In the results, mosGraphGen was used and illustrated using two widely used multi-omics datasets of The Cancer Genome Atlas (TCGA) and Alzheimer’s disease (AD) samples. Availability The code of mosGraphGen is open-source and publicly available via GitHub: Https://github.com/FuhaiLiAiLab/mosGraphGen Supplementary information Supplementary data are available at Bioinformatics Advances online.

On the Constructor–Blocker game

AbstractIn the Constructor–Blocker game, two players, Constructor and Blocker, alternately claim unclaimed edges of the complete graph . For given graphs and , Constructor can only claim edges that leave her graph ‐free, while Blocker has no restrictions. Constructor's goal is to build as many copies of as she can, while Blocker attempts to minimize the number of copies of in Constructor's graph. The game ends once there are no more edges that Constructor can claim. The score of the game is the number of copies of in Constructor's graph at the end of the game when both players play optimally and Constructor plays first. In this paper, we extend results of Patkós, Stojaković and Vizer on to many pairs of and : We determine when and , also when both and are odd cycles, using Szemerédi's Regularity Lemma. We also obtain bounds of when and .

GazeGCN: Gaze-aware Graph Convolutional Network for Text Classification

Graph convolutional networks (GCNs) are capable of capturing contextual relationships in text classification. In this paper, we propose a novel Gaze-aware Graph Convolutional Network (GazeGCN) for text classification, where the gaze signals from humans are incorporated into the GCN architecture, empowering our GazeGCN to use gaze information that can be shared across all data to model the intricate relationships between words and documents. To be specific, we first build a gaze prediction model to obtain five gaze signals to form the gaze distribution of each word. Then, Wasserstein Distance is employed to derive the gaze-aware word–word weight by calculating the gaze distribution between words, so as to improve the capture of luxuriant and similar relationships between words. Furthermore, to enhance the extraction of contextual syntactic information, we introduce a Gaze-enhanced TF–IDF method integrating gaze signals and term frequency–inverse document frequency (TF–IDF) for gaze-aware word-document weight derivation, thus making up for the lack of TF–IDF that does not consider syntactic information. Afterward, the rich source of graph edge information including gaze-aware word–word weight and gaze-aware word-document weight is incorporated to construct graphs, so as to leverage the relationship of word–word and word-document. Experiments show encouraging results on seven benchmark datasets that our approach outperforms the state-of-the-art baseline methods.

A Coupled Spatial-Network Model: A Mathematical Framework for Applications in Epidemiology.

There is extensive evidence that network structure (e.g., air transport, rivers, or roads) may significantly enhance the spread of epidemics into the surrounding geographical area. A new compartmental modeling framework is proposed which couples well-mixed (ODE in time) population centers at the vertices, 1D travel routes on the graph's edges, and a 2D continuum containing the rest of the population to simulate how an infection spreads through a population. The edge equations are coupled to the vertex ODEs through junction conditions, while the domain equations are coupled to the edges through boundary conditions. A numerical method based on spatial finite differences for the edges and finite elements in the 2D domain is described to approximate the model, and numerical verification of the method is provided. The model is illustrated on two simple and one complex example geometries, and a parameter study example is performed. The observed solutions exhibit exponential decay after a certain time has passed, and the cumulative infected population over the vertices, edges, and domain tends to a constant in time but varying in space, i.e., a steady state solution.

Maui: Black-Box Edge Privacy Attack on Graph Neural Networks

Graphs are ubiquitous data structures with nodes representing objects and edges representing relationships between them. Graph Neural Networks (GNNs) have recently been proposed to study graph-structured data, but unfortunately, are susceptible to privacy leakage. This issue becomes more urgent as GNNs gain wide deployment in many real-world settings including social network analysis, bioinformatics, and cybersecurity. In this paper, we propose the first link inference attack that can compromise user data under the most difficult security settings, which we call Maui. We demonstrate that private edge information can be inferred by a malicious user with a black-box approach. Extensive experiments on six real-world datasets show our attacks conduct effective link inference attacks in various scopes. Our attack achieves significant performance improvements over the current state-of-the-art. When targeting 2-layer Graph Convolution Networks, for inferring edges of a single node, our attack outperforms the best existing method by 12.0%, increasing from 83.7% to 95.7%; when inferring edges of the entire graph, our attack achieves a 19.6% improvement, from 67.7% to 87.3%. Our results underscore the need for countermeasures against privacy attacks in GNNs, as they can reveal rich information about graph structures.

Mixed orthogonality graphs for continuous-time stationary processes

In this paper, we introduce different concepts of Granger causality and contemporaneous correlation for multivariate stationary continuous-time processes to model different dependencies between the component processes. Several equivalent characterisations are given for the different definitions, in particular by orthogonal projections. We then define two mixed graphs based on different definitions of Granger causality and contemporaneous correlation, the (mixed) orthogonality graph and the local (mixed) orthogonality graph. In these graphs, the components of the process are represented by vertices, directed edges between the vertices visualise Granger causal influences and undirected edges visualise contemporaneous correlation between the component processes. Further, we introduce various notions of Markov properties in analogy to Eichler (2012), which relate paths in the graphs to different dependence structures of subprocesses, and we derive sufficient criteria for the (local) orthogonality graph to satisfy them. Finally, as an example, for the popular multivariate continuous-time AR (MCAR) processes, we explicitly characterise the edges in the (local) orthogonality graph by the model parameters.

A short-term wind power prediction method via self-adaptive adjacency matrix and spatiotemporal graph neural networks

Graph neural networks (GNN) recently have been successfully applied in wind power prediction by employing geo-information to construct the graphs that serve as the GNN-based prediction model. However, these graphs cannot maintain the latent correlation and attribute of the wind turbines, leading to inferior performance. This research proposes an optimizing wind power prediction model through attention mechanism and spatiotemporal graph neural networks. Initially, the spectral clustering and a self-adjacency matrix to construct the graph nodes and edges. Subsequently, the proposed ASTGNN combined from graph convolutional networks and a gated recurrent unit with an attention mechanism is designed to act as the prediction model. A comprehensive set of experiments has been carried out and the proposed method not only improved prediction accuracy but also enhanced computational efficiency. According to the results obtained, the developed model demonstrated an improvement that is up to 13 % in terms of RMSE and 6.7 % in terms of computation time compared to the latest models.

D $d$‐connectivity of the random graph with restricted budget

AbstractIn this short note, we consider a graph process recently introduced by Frieze, Krivelevich and Michaeli. In their model, the edges of the complete graph are ordered uniformly at random and are then revealed consecutively to a player called Builder. At every round, Builder must decide if they accept the edge proposed at this round or not. We prove that, for every , Builder can construct a spanning ‐connected graph after rounds by accepting edges with probability converging to 1 as . This settles a conjecture of Frieze, Krivelevich and Michaeli.

ES-GNN: Generalizing Graph Neural Networks Beyond Homophily With Edge Splitting.

While Graph Neural Networks (GNNs) have achieved enormous success in multiple graph analytical tasks, modern variants mostly rely on the strong inductive bias of homophily. However, real-world networks typically exhibit both homophilic and heterophilic linking patterns, wherein adjacent nodes may share dissimilar attributes and distinct labels. Therefore, GNNs smoothing node proximity holistically may aggregate both task-relevant and irrelevant (even harmful) information, limiting their ability to generalize to heterophilic graphs and potentially causing non-robustness. In this work, we propose a novel Edge Splitting GNN (ES-GNN) framework to adaptively distinguish between graph edges either relevant or irrelevant to learning tasks. This essentially transfers the original graph into two subgraphs with the same node set but complementary edge sets dynamically. Given that, information propagation separately on these subgraphs and edge splitting are alternatively conducted, thus disentangling the task-relevant and irrelevant features. Theoretically, we show that our ES-GNN can be regarded as a solution to a disentangled graph denoising problem, which further illustrates our motivations and interprets the improved generalization beyond homophily. Extensive experiments over 11 benchmark and 1 synthetic datasets not only demonstrate the effective performance of ES-GNN but also highlight its robustness to adversarial graphs and mitigation of the over-smoothing problem.

Unsupervised multi-view graph representation learning with dual weight-net

Unsupervised multi-view graph representation learning (UMGRL) aims to capture the complex relationships in the multi-view graph without human annotations, so it has been widely applied in real-world applications. However, existing UMGRL methods still face the issues as follows: (i) Previous UMGRL methods tend to overlook the importance of nodes with different influences and the importance of graphs with different relationships, so that they may lose discriminative information in nodes with large influences and graphs with important relationships. (ii) Previous UMGRL methods generally ignore the heterophilic edges in the multi-view graph to possibly introduce noise from different classes into node representations. To address these issues, we propose a novel bi-level optimization UMGRL framework with dual weight-net. Specifically, the lower-level optimizes the parameters of encoders to obtain node representations of different graphs, while the upper-level optimizes the parameters of the dual weight-net to adaptively and dynamically capture the importance of node level, graph level, and edge level, thus obtaining discriminative fused representations for downstream tasks. Moreover, theoretical analysis demonstrates that the proposed method shows a better generalization ability on downstream tasks, compared to previous UMGRL methods. Extensive experimental results verify the effectiveness of the proposed method on public datasets, in terms of different downstream tasks, compared to numerous comparison methods.

Routing Task in Dynamic Fog Computing Network

Relevance. In the context of traffic growth, transition to IMT-2030 networks and Telepresence services, the tasks of efficient management of network and computing resources occupy a special place. Fog computing as the next stage of decomposition of the architecture of multi access edge cloud computing is designed to radically change the models and methods of distributing computing tasks, influencing, among other things, the user-operator interaction models. At the moment, there is a whole layer of scientific problems for revealing the possibilities of fog computing. They can be divided into a number of areas, such as: study of models and methods for implementing services of ultra-reliable and ultra-low latency communications, defined in IMT-2020 networks; study of models and methods for ensuring quality of service, including quality of experience; study of methods for live migration of microservices, as well as groups of typical microservices; study of models and methods for distributing resources of dynamic fog computing while ensuring the stability of fog computing forms (clusters, nebulae); one of the potentially effective areas is research in the field of combining federated learning with dynamic fog computing. This paper solves a routing problem that can be attributed to the direction of infrastructure research in dynamic fog computing.Problem statement: research and develop the effective methods for routes determination in a dynamic fog computing network, including tasks of migrating microservices of telepresence services. Goal of the work: research and development of an effective method for ways determination to migrate microservices in communication networks using fog computing technologies, which could take into account not only the characteristics of connections (edges of the network graph), but also the computing capabilities and limitations of fog computing devices, as well as their features - the dynamics of computing devices. Methods: in order to test the proposed method, the program model was developed in the NS-3 modeling environment. Result. Analysis of the results showed the effectiveness of the proposed method within the framework of the task and various application scenarios. Novelty. A microservice migration method has been developed as a new routing protocol in a dynamic fog computing environment, which differs from the known ones in that this method ensures the interaction of fog computing devices for migrating microservices, while achieving a reduction in energy consumption by fog computing devices by 41% and reducing the share of lost packages on average up to 34%. Practical significance: The developed method can be used to implement fog computing in conditions of mobility of end devices in order to achieve the requirements of promising services of IMT-2030 networks.

Enhancing Knowledge-Aware Recommendation with Dual-Graph Contrastive Learning

Incorporating knowledge graphs as auxiliary information to enhance recommendation systems can improve the representations learning of users and items. Recommendation methods based on knowledge graphs can introduce user–item interaction learning into the item graph, focusing only on learning the node vector representations within a single graph; alternatively, they can treat user–item interactions and item graphs as two separate graphs and learn from each graph individually. Learning from two graphs has natural advantages in exploring original information and interaction information, but faces two main challenges: (1) in complex graph connection scenarios, how to adequately mine the self-information of each graph, and (2) how to merge interaction information from the two graphs while ensuring that user–item interaction information predominates. Existing methods do not thoroughly explore the simultaneous mining of self-information from both graphs and effective interaction information, leading to the loss of valuable insights. Considering the success of contrastive learning in mining self-information and auxiliary information, this paper proposes a dual-graph contrastive learning recommendation method based on knowledge graphs (KGDC) to explore a more accurate representations of users and items in recommendation systems based on external knowledge graphs. In the learning process within the self-graph, KGDC strengthens and represents the information of different connecting edges in both graphs, and extracts the existing information more fully. In interactive information learning, KGDC reinforces the interaction relationship between users and items in the external knowledge graph, realizing the leading role of the main task. We conducted a series of experiments on three standard datasets, and the results show that the proposed method can achieve better results.

New Dynamic Programming algorithm for the Multiobjective Minimum Spanning Tree Problem

The Multiobjective Minimum Spanning Tree (MO-MST) problem generalizes the Minimum Spanning Tree problem by weighting the edges of the input graph using vectors instead of scalars. In this paper, we design a new Dynamic Programming MO-MST algorithm. Dynamic Programming for a MO-MST instance requests solving a One-to-One Multiobjective Shortest Path (MOSP) instance and both instances have equivalent solution sets. The MOSP instance is defined on a so called transition graph. We study the original size of this graph in detail and reduce its size using cost-dependent arc pruning criteria. To solve the MOSP instance on the reduced transition graph, we design the Implicit Graph Multiobjective Dijkstra Algorithm (IG-MDA), exploiting recent improvements on MOSP algorithms from the literature. All in all, the new IG-MDA outperforms the current state of the art on a big set of instances from the literature. Our code and results are publicly available.

MosGraphGen: a novel tool to generate multi-omics signaling graphs to facilitate integrative and interpretable graph AI model development.

Multi-omics data, i.e., genomics, epigenomics, transcriptomics, proteomics, characterize cellular complex signaling systems from multi-level and multi-view and provide a holistic view of complex cellular signaling pathways. However, it remains challenging to integrate and interpret multi-omics data for mining key disease targets and signaling pathways. Graph AI models have been widely used to analyze graph-structure datasets, and are ideal for integrative multi-omics data analysis because they can naturally integrate and represent multi-omics data as a biologically meaningful multi-level signaling graph and interpret multi-omics data via graph node and edge ranking analysis. However, it is non-trivial for graph-AI model developers to pre-analyze multi-omics data and convert the data into biologically meaningful graphs, which can be directly fed into graph-AI models. To resolve this challenge, we developed mosGraphGen (multi-omics signaling graph generator), generating Multi-omics Signaling graphs (mos-graph) of individual samples by mapping multi-omics data onto a biologically meaningful multi-level background signaling network with data normalization by aggregating measurements and aligning to the reference genome. With mosGraphGen, AI model developers can directly apply and evaluate their models using these mos-graphs. In the results, mosGraphGen was used and illustrated using two widely used multi-omics datasets of TCGA and Alzheimer's disease (AD) samples. The code of mosGraphGen is open-source and publicly available via GitHub: https://github.com/FuhaiLiAiLab/mosGraphGen.

Distilling interaction knowledge for semi-supervised egocentric action recognition

Egocentric action recognition, the identification of actions within video content obtained from a first-person perspective, is receiving increasing attention due to the widespread adoption of wearable camera technology. Nonetheless, the task of annotating actions within a video characterized by a cluttered background and the presence of various objects is labor-intensive. In this paper, we consider learning for egocentric action recognition in a semi-supervised manner. Inspired by the fact that videos captured from first-person viewpoint usually contain rich contents about how human hands interact with objects, we thus propose to employ a popular teacher–student framework and distill the interaction knowledge between hand and objects for semi-supervised egocentric action recognition. We refer to the proposed method as Interaction Knowledge Distillation or IKD. Specifically, the teacher network takes hands and action-related objects in the labeled videos as input, and uses graph neural networks to capture their spatial–temporal relations as graph edge features. The student network then takes the detected hands/objects from both labeled and unlabeled videos as input and mimics the teacher network to learn from the interactions to improve model performance. Experiments are performed on two popular egocentric action recognition datasets, Something-Something-V2 and EPIC-KITCHENS-100, which show that our proposed approach consistently outperforms recent state-of-the-art methods in typical semi-supervised settings.

A graph residual generation network for node classification based on multi-information aggregation

The key to improving the performance of graph convolutional networks (GCN) is to fully explore the correlation between neighboring and distant information. Aiming at the over-smoothing problem of GCN, in order to make full use of the relationship among features, graphs and labels, a graph residual generation network based on multi-information aggregation (MIA-GRGN) is proposed. Firstly, aiming at the defects of GCN, we design a deep initial residual graph convolution network (DIRGCN), which connects the initial input through residuals, so that each layer node retains part of the information of the initial features, ensuring the localization of the graph structure and effectively alleviating the problem of over-smoothing. Secondly, we propose a random graph generation method (RGGM) by utilizing graph edge sampling and negative edge sampling, and optimize the supervision loss function of DIRGCN in the form of generation framework. Finally, applying RGGM and DIRGCN as inference modules for modeling hypotheses and obtaining approximate posterior distributions of unknown labels, an optimized loss function is obtained, we construct a multi-information aggregation MIA-GRGN that combines graph structure, node characteristics and label joint distribution. Experiments on benchmark graph classification datasets show that MIA-GRGN achieves better classification results compared with the benchmark models and mainstream models, especially for datasets with less dense edge relationships between nodes.

Improved time complexity for spintronic oscillator ising machines compared to a popular classical optimization algorithm for the Max-Cut problem.

Solving certain combinatorial optimization problems like Max-Cut becomes challenging once the graph size and edge connectivity increase beyond a threshold, with brute-force algorithms which solve such problems exactly on conventional digital computers having the bottleneck of exponential time complexity. Hence currently, such problems are instead solved approximately using algorithms like Goemans-Williamson (GW) algorithm, run on conventional computers with polynomial time complexity. Phase binarized oscillators (PBOs), also often known as oscillator Ising machines, have been proposed as an alternative to solve such problems. In this paper, restricting ourselves to the combinatorial optimization problem Max-Cut solved on three kinds of graphs (Mobius Ladder, random cubic, Erdös Rényi) up to 100 nodes, we empirically show that computation time/time to solution (TTS) for PBOs (captured through Kuramoto model) grows at a much lower rate (logarithmically:O(log(N)), with respect to graph sizeN) compared to GW algorithm, for which TTS increases as square of graph size (O(N2)). However, Kuramoto model being a physics-agnostic mathematical model, this time complexity/ TTS trend for PBOs is a general trend and is device-physics agnostic. So for more specific results, we choose spintronic oscillators, known for their high operating frequency (in GHz), and model them through Slavin's model which captures the physics of their coupled magnetization oscillation dynamics. We thereby empirically show that TTS of spintronic oscillators also grows logarithmically with graph size (O(log(N)), while their accuracy is comparable to that of GW. So spintronic oscillators have improved time complexity over GW algorithm. For large graphs, they are expected to compute Max-Cut values much faster than GW algorithm, as well as other oscillators operating at lower frequencies, while maintaining the same level of accuracy.

A data-driven approach for volume feature recognition based on cell graph

ABSTRACT Machining feature recognition is one of the key technologies for realizing automated process planning. The machining feature represented by volume contains more complete information due to its property of forming a closed region which facilitates subsequent process planning. The existing volume feature recognition approaches mainly rely on expert-defined rules to combine volumes into the desired feature. When dealing with some complex parts, volume feature recognition faces combinatorial explosion, and it is difficult to formulate complete combination rules manually. This research proposes a novel data-driven approach for volume machining feature recognition. The proposed approach does not require manually defined rules and can learn various complex volume combination rules from the data, which can be implicitly represented by the structure and parameters of the data-driven model. The volume machining feature recognition is first converted to machining feature subgraph (MFS) detection by creating the zone graph. MFS detection can then be achieved by graph edge cut. Finally, the edge cut is modeled as a binary classification deep learning problem, which is solved by graph neural network. Two datasets are created, and case studies are carried out on them. The results show that the proposed approach is effective for volume machining feature recognition.