Search

Traffic flow forecasting plays a crucial role in applications such as intelligent transportation systems. Despite significant research in this field, the current methods have limitations that hinder the realization of highly accurate predictions. Existing GCN-based approaches typically rely on a definite graph structure derived from a physical topology or learned from node features, which is insufficient for building intricate spatial relationships among nodes. To address this challenge, we propose an adaptive spatial–temporal graph convolutional network for traffic forecasting. Our approach exploits a multi-head attention mechanism to construct multi-view feature graphs. We then introduce an adaptive graph convolution method to dynamically aggregate and propagate information from both the topology graph and multi-view feature graphs, which are capable of capturing complex spatial correlations across diverse proximity ranges. Furthermore, we designed a cascaded structural framework that combines temporal information with node features using gated dilated causal convolution to ensure the integrated modeling of spatial–temporal dynamics in traffic flow. Experiments on real-world datasets demonstrate that our proposed method outperforms the current mainstream methods, achieving better performance in traffic flow forecasting. The code is available at https://github.com/dhxdla/AdpSTGCN.git. [ABSTRACT FROM AUTHOR]

Data center infrastructures require constant monitoring to ensure stable and reliable operation and time-series forecasting plays an indispensable role in intelligent operations and maintenance in data centers. However, the potential for accurate time-series predictions is often limited due to the overlooked relationships between data records from independent sensors. Inferring relationships for a potential graph representation of a data center is challenging due to complex relationships between nodes and multiple factors that may cause connections between them. Moreover, graphs change dynamically in long-term predictions, but current methods do not account for future graph changes. To address these challenges, we propose a long-term time-series forecasting framework called Multi-factor Separation Evolutionary Spatial–Temporal Graph Neural Networks (MSE-STGNN). Our framework considers edge diversity, graph changes and spatial–temporal architecture in long-term prediction processes and proposes three modules. Specifically, we propose a Multi-factor Separation (MS) module to separate the factors influencing node connectivity, enabling the acquisition of a graph more closely aligned with actual circumstances; then we propose a Graph Prediction (GP) module to incorporate future graphs to correct errors in the graph on which multi-step predictions depend. Moreover, we propose an Attention-enhanced Spatial–temporal dilated causal convolution module (AS-Conv) to more effectively leverage information pertaining to spatial and historical events. Our experimental results on datasets comprising of temperature and IT power data collected from real-world data centers show that the proposed method outperforms other advanced prediction methods in terms of prediction accuracy, and the learned latent graphs are explainable. [ABSTRACT FROM AUTHOR]

Network representation learning has achieved significant success in homogeneous network data analysis in recent years. Nevertheless, they cannot be directly applied in multiplex networks. To overcome the characteristic of heterogeneity in multiplex networks, several emerging methods utilize the concept of meta-path to denote different types of relations and obtain the node representations for each type of meta-path individually. Despite the remarkable progress, there still exist two important issues in the previous approaches. First, the complementary information between different types of meta-paths that may make the representations more discriminative, is rarely investigated. Second, current studies generally learn multiplex node representations based on the original graph structure, while overlooking the latent relations between nodes. To address the aforementioned issues, in this paper, we propose a novel model with M eta-path I nfomax joint S tructure E nhancement (MISE) for multiplex network representations. Specifically, we first develop a meta-path infomax mechanism, which maximizes the mutual information between local and global meta-path representations, making the node representation contain more complementary information. Additionally, we propose a graph structure learning module that captures the implicit correlations between nodes to construct the latent graph structure. Such structure enhancement is a simple yet surprisingly effective technique to learn high-quality representations. We sufficiently evaluate the performance of our proposal on both supervised and unsupervised downstream tasks. Comprehensive experimental results show that our MISE achieves a promising boost in performance on a variety of real-world datasets for multiplex network representation learning. • Proposing a meta-path infomax module to extract complementary information. • Introducing graph structure learning to extract the implicit semantic correlations. • Evaluating on four datasets sufficiently to verify the superiority of our proposal. [ABSTRACT FROM AUTHOR]

LiDAR-based 3D detection is critical in autonomous driving perception systems. However, point-based 3D object detection that directly learns from point clouds is challenging owing to the sparsity and irregularity of LiDAR point clouds. Existing point-based methods are limited by fixed local relationships and the sparsity of distant and occluded objects. To address these issues, we propose a dynamic graph transformer 3D object detection network (DGT-Det3D) based on a dynamic graph transformer (DGT) module and a proposal-aware fusion (PAF) module. The DGT module is built on a dynamic graph and graph-aware self-attention module, which adaptively concentrates on the foreground points and encodes the graph to capture long-range dependencies. With the DGT module, DGT-Det3D has better capability to detect distant and occluded objects. To further refine the proposals, our PAF module fully integrates the proposal-aware spatial information and combines it with the point-wise semantic features from the first stage. Extensive experiments on the KITTI dataset demonstrate that our approach achieves state-of-the-art accuracy for point-based methods. In addition, DGT brings significant improvements when combined with state-of-the-art methods on the Waymo open dataset. [ABSTRACT FROM AUTHOR]

Feature engineering relies on domain knowledge and human intervention. To automate the process of feature engineering, automated feature construction methods use deep neural networks to capture feature interactions and attention coefficients to quantify the relationship between features. However, these methods ignore the influence of insignificant features that introduce noise and degrade the performance of the model. In this paper, we study the problem of feature interactions from the perspective of graph and propose a novel Automatic Feature Generation model based on Graph Structure Learning, called AFGSL. In this model, the adjacency matrix reflects the relationships between features, so that it can be used to filter out insignificant features. Furthermore, Q-learning is used to train the policy of stacking interaction layers, which enables it to make full use of both local and global information in the process of feature generation. The results of experiments on four real-world datasets show that AFGSL outperforms the state-of-the-art methods. [ABSTRACT FROM AUTHOR]