Graph Generation Models Research Articles

SkyMap: a generative graph model for GNN benchmarking

SkyMap: a generative graph model for GNN benchmarking

Method for Locating Secondary Circuit Faults in Substations Based on Graph Neural Networks

To improve the accuracy and portability of fault location in secondary circuits of smart substations, a fault location method for secondary circuits based on graph neural networks is proposed. First, a fault analysis of the secondary circuits in smart substations is conducted, and a graph database model of the secondary circuits is constructed, revealing the connection relationships between the secondary devices. Then, the fault location problem in secondary circuits is defined as a graph classification problem, and a fault graph generation model is proposed based on a representation method for fault information in secondary circuits. Finally, a fault location model for secondary circuits based on graph neural networks is constructed, and a performance optimization method based on binary cross-entropy is proposed to achieve accurate fault location in substation secondary circuits. Using the secondary circuits of a 220 kV smart substation as a reference, different case studies of secondary circuit faults are generated by altering the network topology, connection relationships, and network configurations through the fault graph generation model. Experiments comparing the proposed location method with other models show that this method has higher location accuracy and robustness.

Generation Method for HVAC Systems Design Schemes in Office Buildings Based on Deep Graph Generative Models

The design process of heating, ventilation, and air conditioning (HVAC) systems is complex and time consuming due to the need to follow design codes. Since the design standards are not fixed, the final outcome often depends on the designer’s experience. The development of building information modeling (BIM) technology has made information throughout the building lifecycle more integrated. BIM-based forward design is now widely used, providing a data foundation for combining HVAC system design with machine learning. This paper proposes an unsupervised learning method based on deep graph generative models to uncover hidden design patterns and optimization strategies from the design results. We trained and validated four deep graph generative models—GAE, GNF, GAN, and diffusion—using HVAC system terminal pipeline layout data. Accuracy and precision metrics were used to compare the generated designs with automated forward design solutions, assessing the models’ ability to capture both local variations and broader changes in design logic. A graph-neural-network-based evaluation method was employed to measure the models’ capacity to detect changes. The results indicate that all four models achieved prediction accuracies exceeding 90% and precision rates above 75%. The models effectively captured both local modifications made by designers and global design changes, showing greater sensitivity to global layout adjustments than to local updates. When comparing the results generated by deep graph generative models and the actual design, it is obvious that the accuracy of the predictions varies significantly due to the complexity of the test buildings.

Decomposed Prototype Learning for Few-Shot Scene Graph Generation

Today's scene graph generation (SGG) models typically require abundant manual annotations to learn new predicate types. Therefore, it is difficult to apply them to real-world applications with massive uncommon predicate categories whose annotations are hard to collect. In this paper, we focus on Few-Shot SGG (FSSGG) , which encourages SGG models to be able to quickly transfer previous knowledge and recognize unseen predicates well with only a few examples. However, current methods for FSSGG are hindered by the high intra-class variance of predicate categories in SGG: On one hand, each predicate category commonly has multiple semantic meanings under different contexts. On the other hand, the visual appearance of relation triplets with the same predicate differs greatly under different subject-object compositions. Such great variance of inputs makes it hard to learn generalizable representation for each predicate category with current few-shot learning (FSL) methods. However, we found that this intra-class variance of predicates is highly related to the composed subjects and objects. To model the intra-class variance of predicates with subject-object context, we propose a novel Decomposed Prototype Learning (DPL) model for FSSGG. Specifically, we first construct a decomposable prototype space to capture diverse semantics and visual patterns of subjects and objects for predicates by decomposing them into multiple prototypes. Afterwards, we integrate these prototypes with different weights to generate query-adaptive predicate representation with more reliable semantics for each query sample. We conduct extensive experiments and compare with various baseline methods to show the effectiveness of our method.

Relation-Specific Feature Augmentation for unbiased scene graph generation

Scene Graph Generation (SGG) models suffer from the long-tailed distribution of relations, which results in biased predictions that favor head relations (e.g., on) over informative tail ones (e.g., sitting on, laying on, standing on). Existing solutions typically adopt class re-balancing strategies to balance data distribution. However, they do not essentially solve the lack of information due to insufficient tail data. To this end, we propose a Relation-Specific Feature Augmentation (RSFA) framework to mitigate the long-tailed bias by augmenting relations in the feature space. To perform augmentation effectively, we design an augmentation scheme and a novel Dual Attention Network (DAN). The augmentation scheme augments each relation uniformly based on the reciprocal number of samples to avoid over-fitting. By extracting relation-specific information from new object features generated by a Conditional Variational AutoEncoder (CVAE), DAN generates reliable virtual relation representations to provide useful information to guide optimizing relation classifier. Extensive ablation studies and comprehensive analysis demonstrate the effectiveness of our method in debiasing. And results on the Visual Genome benchmark show that our method significantly outperforms the existing state-of-the-art methods.

Coupling Hyperbolic GCN with Graph Generation for Spatial Community Detection and Dynamic Evolution Analysis

Spatial community detection is a method that divides geographic spaces into several sub-regions based on spatial interactions, reflecting the regional spatial structure against the background of human mobility. In recent years, spatial community detection has attracted extensive research in the field of geographic information science. However, mining the community structures and their evolutionary patterns from spatial interaction data remains challenging. Most existing methods for spatial community detection rely on representing spatial interaction networks in Euclidean space, which results in significant distortion when modeling spatial interaction networks; since spatial community detection has no ground truth, this results in the detection and evaluation of communities being difficult. Furthermore, most methods usually ignore the dynamics of these spatial interaction networks, resulting in the dynamic evolution of spatial communities not being discussed in depth. Therefore, this study proposes a framework for community detection and evolutionary analysis for spatial interaction networks. Specifically, we construct a spatial interaction network based on network science theory, where geographic units serve as nodes and interaction relationships serve as edges. In order to fully learn the structural features of the spatial interaction network, we introduce a hyperbolic graph convolution module in the community detection phase to learn the spatial and non-spatial attributes of the spatial interaction network, obtain vector representations of the nodes, and optimize them based on a graph generation model to achieve the final community detection results. Considering the dynamics of spatial interactions, we analyze the evolution of the spatial community over time. Finally, using taxi trajectory data as an example, we conduct relevant experiments within the fifth ring road of Beijing. The empirical results validate the community detection capabilities of the proposed method, which can effectively describe the dynamic spatial structure of cities based on human mobility and provide an effective analytical method for urban spatial planning.

Information-enhanced deep graph clustering network

Graph clustering is a significant task in complex network research. Deep graph clustering aims to uncover the potential community structure in graph data using the powerful feature extraction capability of deep learning, garnering much attention in recent decades. However, existing graph clustering methods fall short in fully utilizing available information, particularly in effectively fusing structural and attribute information, as well as utilizing coarse-grained data. Consequently, learned node representations remain limited, leading to suboptimal clustering results. To address these challenges, we propose Information-Enhanced Deep Graph Clustering Network (IEDGCN) for unsupervised attribute graphs. IEDGCN introduces key components to enhance information utilization and improve clustering performance. Firstly, we design a new higher-order neighborhood-weighted attribute matrix, effectively integrating higher-order neighborhood information with attributes. Secondly, a graph generation model guides the learning of the structural feature space more effectively. Additionally, IEDGCN captures more coarse-grained information by utilizing community and higher-order neighborhood features to refine clustering results. Finally, the proposed method is uniformly guided through a jointly supervised strategy for representation learning and cluster assignment. Experimental results on different benchmark datasets demonstrate the effectiveness of IEDGCN compared to state-of-the-art methods, emphasizing the importance of information enhancement for graph clustering.

A novel transformer‐based graph generation model for vectorized road design

AbstractRoad network design, as an important part of landscape modeling, shows a great significance in automatic driving, video game development, and disaster simulation. To date, this task remains labor‐intensive, tedious and time‐consuming. Many improved techniques have been proposed during the last two decades. Nevertheless, most of the state‐of‐the‐art methods still encounter problems of intuitiveness, usefulness and/or interactivity. As a rapid deviation from the conventional road design, this paper advocates an improved road modeling framework for automatic and interactive road production driven by geographical maps (including elevation, water, vegetation maps). Our method integrates the capability of flexible image generation models with powerful transformer architecture to afford a vectorized road network. We firstly construct a dataset that includes road graphs, density map and their corresponding geographical maps. Secondly, we develop a density map generation network based on image translation model with an attention mechanism to predict a road density map. The usage of density map facilitates faster convergence and better performance, which also serves as the input for road graph generation. Thirdly, we employ the transformer architecture to evolve density maps to road graphs. Our comprehensive experimental results have verified the efficiency, robustness and applicability of our newly‐proposed framework for road design.

Utilizing Artificial Intelligence in the Development of Medicinal Drugs: Practical Applications

The COVID-19 pandemic has underscored the critical need for innovative drug discovery methods. The journey from conceptualizing a drug to its clinical application is fraught with potential pitfalls, including extensive complexity, significant expenses, and a high risk of failure. Recent years have witnessed remarkable advancements in technologies such as cloud computing, GPUs, and TPUs. These developments, coupled with the surge in medical data availability and the emergence of deep learning, present an unprecedented opportunity to enhance drug discovery processes. By leveraging artificial intelligence (AI) to analyze vast amounts of data-from extensive molecular screening outcomes to individual health records and public health data-the efficiency of the drug discovery pipeline could be significantly improved, minimizing the likelihood of failure. This paper explores the application of AI in various phases of drug development, including the use of computational strategies for de novo drug design and the prediction of drug properties. We address challenges associated with molecular representation, data acquisition, complexity, and the inconsistencies in labeling across open-source databases and AI-powered tools that support drug discovery efforts. Furthermore, we examine the role of advanced AI techniques, such as graph neural networks, reinforcement learning, generative models, and structure-based methods like molecular docking and dynamics simulations, in enhancing drug discovery and evaluating drug efficacy.

MARS: a motif-based autoregressive model for retrosynthesis prediction.

Retrosynthesis is a critical task in drug discovery, aimed at finding a viable pathway for synthesizing a given target molecule. Many existing approaches frame this task as a graph-generating problem. Specifically, these methods first identify the reaction center, and break a targeted molecule accordingly to generate the synthons. Reactants are generated by either adding atoms sequentially to synthon graphs or by directly adding appropriate leaving groups. However, both of these strategies have limitations. Adding atoms results in a long prediction sequence that increases the complexity of generation, while adding leaving groups only considers those in the training set, which leads to poor generalization. In this paper, we propose a novel end-to-end graph generation model for retrosynthesis prediction, which sequentially identifies the reaction center, generates the synthons, and adds motifs to the synthons to generate reactants. Given that chemically meaningful motifs fall between the size of atoms and leaving groups, our model achieves lower prediction complexity than adding atoms and demonstrates superior performance than adding leaving groups. We evaluate our proposed model on a benchmark dataset and show that it significantly outperforms previous state-of-the-art models. Furthermore, we conduct ablation studies to investigate the contribution of each component of our proposed model to the overall performance on benchmark datasets. Experiment results demonstrate the effectiveness of our model in predicting retrosynthesis pathways and suggest its potential as a valuable tool in drug discovery. All code and data are available at https://github.com/szu-ljh2020/MARS.

Goal-directed molecule generation with fine-tuning by policy gradient

Graph-structured drug molecule representations often struggle to generate molecules by particular intentions, which results in generated molecules without pharmacological properties. To address the problem, we propose a de novo molecular generation method that utilizes the policy gradient algorithm of reinforcement learning to fine-tune the molecular graph generation model. The training process of the method is divided into the pre-training stage and the fine-tuning stage. During the pre-training stage, it uses graph neural networks and multilayer perceptrons to train a molecule graph generation model. During the fine-tuning stage, scoring functions are devised for multiple goal-directed generation tasks, and subsequently, the policy loss function is formulated based on the reward shaping mechanism. A value network is designed to calculate the value of taking an action based on the current graph state during agent sampling to guide policy updates. To mitigate the issue of molecular uniqueness decline during the learning process, we dynamically adjust the weights of the two learning processes in the policy loss function, aiming to generate desirable molecules with high probability and reduce the descent of uniqueness. The experiments showed that after fine-tuning, the generative model exhibits a higher probability of generating molecules with desired properties compared to other models. Furthermore, our method effectively mitigates the issue of molecular uniqueness declining during the learning process when compared to alternative fine-tuning methods.

SwinRelTR: an efficient single-stage scene graph generation model for low-resolution images

SwinRelTR: an efficient single-stage scene graph generation model for low-resolution images

High-Temperature Polymer Dielectrics Designed Using an Invertible Molecular Graph Generative Model.

Generating new molecules with the desired physical or chemical properties is the key challenge of computational material design. Deep learning techniques are being actively applied in the field of data-driven material informatics and provide a promising way to accelerate the discovery of innovative materials. In this work, we utilize an invertible graph generative model to generate hypothetical promising high-temperature polymer dielectrics. A molecular graph generative model based on the invertible normalizing flow is trained on a data set containing 250k polymer molecular graphs (mostly generated by an RNN-based generative model) to learn the invertible transformations between latent distributions and molecular graph structures. When generating molecular graphs, a sample vector is drawn from the latent space, and then an adjacency tensor and node attribute matrix are generated through two invertible flows in two steps and assembled into a molecular graph. The model has the merits of exact likelihood training and an efficient one-shot generation process. The learned latent space is used to generate polymers with a high glass-transition temperature (Tg) and a wide band gap (Eg) for the application of high-temperature energy storage film capacitors. This work contributes to the efficient design of high-temperature polymer dielectrics by using deep generative models.

Deep Graph Translation.

Deep generative models for graphs have recently achieved great successes in modeling and generating graphs for studying networks in biology, engineering, and social sciences. However, they are typically unconditioned generative models that have no control over the target graphs given a source graph. In this article, we propose a novel graph-translation-generative-adversarial-nets (GT-GAN) model that transforms the source graphs into their target output graphs. GT-GAN consists of a graph translator equipped with innovative graph convolution and deconvolution layers to learn the translation mapping considering both global and local features. A new conditional graph discriminator is proposed to classify the target graphs by conditioning on source graphs while training. Extensive experiments on multiple synthetic and real-world datasets in the domain of cybernetworks, the Internet of Things, and neuroscience demonstrate that the proposed GT-GAN model significantly outperforms other baseline methods in terms of both effectiveness and scalability. For instance, GT-GAN outperforms the classical state-of-the-art (SOTA) methods in functional connectivity (FC) prediction of brain networks by at least 32.5%.

C3DSG: A 3D Scene Graph Generation Model Using Point Clouds of Indoor Environment

C3DSG: A 3D Scene Graph Generation Model Using Point Clouds of Indoor Environment

Quantum walk-based protocol for secure communication between any two directly connected nodes on a network

The utilization of quantum entanglement as a cryptographic resource has superseded conventional approaches to secure communication. Security and fidelity of intranetwork communication between quantum devices is the backbone of a quantum network. This work presents an protocol that generates entanglement between any two directly connected nodes of a quantum network to be used as a resource to enable quantum communication across that pair in the network. The protocol is based on a directed discrete-time quantum walk and paves the way for private inter-node quantum communication channels in the network. We also present the simulation results of this protocol on random networks generated from various models. We show that after implementation, the probability of the walker being at all nodes other than the source and target is negligible and this holds independent of the random graph generation model. This constitutes a viable method for the practical realisation of secure communication over any random network topology.

RelTR: Relation Transformer for Scene Graph Generation.

Different objects in the same scene are more or less related to each other, but only a limited number of these relationships are noteworthy. Inspired by Detection Transformer, which excels in object detection, we view scene graph generation as a set prediction problem. In this paper, we propose an end-to-end scene graph generation model Relation Transformer (RelTR), which has an encoder-decoder architecture. The encoder reasons about the visual feature context while the decoder infers a fixed-size set of triplets subject-predicate-object using different types of attention mechanisms with coupled subject and object queries. We design a set prediction loss performing the matching between the ground truth and predicted triplets for the end-to-end training. In contrast to most existing scene graph generation methods, RelTR is a one-stage method that predicts sparse scene graphs directly only using visual appearance without combining entities and labeling all possible predicates. Extensive experiments on the Visual Genome, Open Images V6, and VRD datasets demonstrate the superior performance and fast inference of our model.

TenGAN: adversarially generating multiplex tensor graphs

In this work, we explore multiplex graph (networks with different types of edges) generation with deep generative models. We discuss some of the challenges associated with multiplex graph generation that make it a more difficult problem than traditional graph generation. We propose TenGAN, the first neural network for multiplex graph generation, which greatly reduces the number of parameters required for multiplex graph generation. We also propose 3 different criteria for evaluating the quality of generated graphs: a graph-attribute-based, a classifier-based, and a tensor-based method. We evaluate its performance on 4 datasets and show that it generally performs better than other existing statistical multiplex graph generative models. We also adapt HGEN, an existing deep generative model for heterogeneous information networks, to work for multiplex graphs and show that our method generally performs better.

Generating Activity Snippets by Learning Human-Scene Interactions

We present an approach to generate virtual activity snippets, which comprise sequenced keyframes of multi-character, multi-object interaction scenarios in 3D environments, by learning from recordings of human-scene interactions. The generation consists of two stages. First, we use a sequential deep graph generative model with a temporal module to iteratively generate keyframe descriptions, which represent abstract interactions using graphs, while preserving spatial-temporal relations through the activities. Second, we devise an optimization framework to instantiate the activity snippets in virtual 3D environments guided by the generated keyframe descriptions. Our approach optimizes the poses of character and object instances encoded by the graph nodes to satisfy the relations and constraints encoded by the graph edges. The instantiation process includes a coarse 2D optimization followed by a fine 3D optimization to effectively explore the complex solution space for placing and posing the instances. Through experiments and a perceptual study, we applied our approach to generate plausible activity snippets under different settings.

RegraphGAN: A graph generative adversarial network model for dynamic network anomaly detection

Due to the wide application of dynamic graph anomaly detection in cybersecurity, social networks, e-commerce, etc., research in this area has received increasing attention. Graph generative adversarial networks can be used in dynamic graph anomaly detection due to their ability to model complex data, but the original graph generative adversarial networks do not have a method to learn reverse mapping and require an expensive process in recovering the potential representation of a given input. Therefore, this paper proposes a novel graph generative adversarial network by adding encoders to map real data to latent space to improve the training efficiency and stability of graph generative adversarial network models, which is named RegraphGAN in this paper. And this paper proposes a dynamic network anomaly edge detection method by combining RegraphGAN with spatiotemporal coding to solve the complex dynamic graph data and the problem of attribute-free node information coding challenges. Meanwhile, anomaly detection experiments are conducted on six real dynamic network datasets, and the results show that the dynamic network anomaly detection method proposed in this paper outperforms other existing methods.