Region representation learning plays a pivotal role in urban computing by extracting meaningful features from unlabeled urban data. Analogous to how perceived facial age reflects an individual's health, the visual appearance of a city serves as its "portrait", encapsulating latent socio-economic and environmental characteristics. Recent studies have explored leveraging Large Language Models (LLMs) to incorporate textual knowledge into imagery-based urban region representation learning. However, two major challenges remain: i) difficulty in aligning fine-grained visual features with long captions, and ii) suboptimal knowledge incorporation due to noise in LLM-generated captions. To address these issues, we propose a novel pre-training framework called UrbanLN that improves Urban region representation learning through Long-text awareness and Noise suppression. Specifically, we introduce an information-preserved stretching interpolation strategy that aligns long captions with fine-grained visual semantics in complex urban scenes. To effectively mine knowledge from LLM-generated captions and filter out noise, we propose a dual-level optimization strategy. At the data level, a multi-model collaboration pipeline automatically generates diverse and reliable captions without human intervention. At the model level, we employ a momentum-based self-distillation mechanism to generate stable pseudo-targets, facilitating robust cross-modal learning under noisy conditions. Extensive experiments across four real-world cities and various downstream tasks demonstrate the superior performance of our UrbanLN.

Accurate travel time estimation (TTE) plays a crucial role in intelligent transportation systems. However, it remains challenging due to heterogeneous data sources and complex traffic dynamics. Moreover, traditional approaches typically convert trajectory data into fixed-length representations. This overlooks the inherent variability of real-world motion patterns, often resulting in information loss and redundancy. To address these challenges, this paper introduces the Multimodal Dynamic Trajectory Integration (MDTI) framework--a novel multimodal trajectory representation learning approach that integrates GPS sequences, grid trajectories, and road network constraints to enhance the performance of TTE. MDTI employs modality-specific encoders and a multimodal fusion module to capture complementary spatial, temporal, and topological semantics, while a dynamic trajectory modeling mechanism adaptively regulates information density for trajectories of varying lengths. Two self-supervised pretraining objectives, named contrastive alignment and masked language modeling, further strengthen multimodal consistency and contextual understanding. Extensive experiments on three real-world datasets demonstrate that MDTI consistently outperforms state-of-the-art baselines, confirming its robustness and strong generalization abilities. The code is publicly available at: https://github.com/City-Computing/MDTI.

Temporal Knowledge Graph (TKG) reasoning seeks to predict future events by analyzing historical data, where the effective leverage of both local and global historical facts proves crucial. Existing approaches employ graph neural networks (GNNs) and recurrent neural networks (RNNs) for local evolution patterns, complemented by statistical methods to enhance attention to global facts, demonstrating efficient predictive capabilities. However, traditional GNNs, constrained by their low-order neighborhood aggregation design, inherently fail to model potential high-order dependencies among facts. Furthermore, existing global history modeling approaches may introduce irrelevant historical information that interferes with prediction tasks. To address these limitations, we propose a Dual History-aware HyperGraph Network for TKG reasoning, namely DHHGN. Specifically, for local history modeling, we design a hybrid hypergraph-graph joint recurrent convolution module that simultaneously captures low-order neighborhood information and high-order interaction patterns among entities, employing a gating mechanism to adaptively blend their contributions. For global history modeling, we propose a dual history enhancement module that amplifies attention on pivotal historical facts while ensuring holistic integration of all historical contexts. Extensive experiments on four public benchmarks validate that DualHist-HGN consistently outperforms existing state-of-the-art methods across TKG reasoning tasks.

Temporal Knowledge Graph Completion (TKGC) aims to predict missing entities or relations based on historical facts, thereby facilitating the understanding of dynamic system evolution and supporting downstream reasoning tasks. However, existing methods predominantly focus on modeling sequential and structural dependencies, often overlooking the rich semantic information embedded in entities and relations, as well as the higher-order interactions among them, which limits their ability to handle complex, evolving scenarios effectively. To address these limitations, we propose DSTAG, a novel TKGC approach based on a semantic tag-enhanced dual-graph convolutional network. Our method leverages large language models to generate contextualized semantic multi-tags for both entities and relations (e.g., ''political event,'' ''economic activity''), thereby enriching their semantic representations. Furthermore, we introduce a semantic tag representation mechanism that captures higher-order dependencies during the aggregation and propagation of semantic tag information across graphs. DSTAG adopts a dual-graph convolutional network architecture, where the relation graph convolution extracts semantic features between temporal relationships and injects this information into the entity graph convolution, enabling joint modeling of entities and relations. We evaluate DSTAG on three widely used TKG benchmarks: ICEWS14, ICEWS18, and ICEWS05-15. Experimental results show that DSTAG achieves substantial MRR improvements over state-of-the-art baselines by 8.64%, 9.81% and 4.56%, respectively.

Real-time traffic forecasting acts as a critical enabling service for IoT-driven Intelligent Transportation Systems (ITS). While existing Spatiotemporal Graph Neural Networks (STGNNs) achieve superior forecasting accuracy, their intensive computational complexity and high latency create a deployment bottleneck for resource-constrained IoT edge devices. To address this resource-accuracy mismatch, we propose a novel framework termed Dynamic Hub-Aware Knowledge Distillation (DHKD). Unlike traditional uniform distillation paradigms, DHKD introduces a topology-aware strategy to transfer knowledge from a complex teacher to a lightweight Spatiotemporal Multi-Layer Perceptron (STMLP) student model. Specifically, we design a dynamic hub-aware gating (DHAG) mechanism that adaptively identifies time-varying pivotal sensing nodes (hubs), ensuring that the student model prioritizes the most information-dense spatiotemporal regions. Furthermore, we develop a multi-level distillation strategy that aligns both intermediate features and final predictions through contrastive learning and attention mechanisms. Extensive experiments conducted on four real-world datasets demonstrate that DHKD significantly reduces inference latency while maintaining state-of-the-art performance, validating its viability for deployment on resource-constrained Internet of Vehicles (IoV) edge devices. The implementation code is available at https://anonymous.4open.science/r/DHKD.

The Transformer model, with its ability to capture long-term dependencies, has demonstrated significant potential in enhancing long-term traffic flow prediction for effective urban transportation management. However, most existing Transformer-based methods adhere to a centralized approach, failing to address privacy concerns and to optimize computational resource utilization. Although emerging federated learning paradigms offer privacy protection, the average aggregation still overlooks client heterogeneity and lacks the synchronous efficiency required for traffic flow prediction tasks. Consequently, we introduce FedTFormer, a hierarchical federated learning framework tailored to boost the performance of Transformer models in decentralized environments, involving clients, edge servers, and a central server. Initially, clients are organized into clusters through a sophisticated static clustering mechanism anchored in bipartite graph theory. FedTFormer enhances robustness of Transformer by facilitating synchronous average aggregation within clusters. Additionally, it performs asynchronous fine-tuning of cluster-specific parameters, leveraging hypernetwork constructed on the central server. Clients utilize an optimized Transformer model for localized training, harnessing its proficiency in capturing long-term spatio-temporal dependencies. Ultimately, we conduct extensive experiments across three datasets, comparing our method against ten sophisticated approaches and demonstrating the effectiveness and robustness of FedTFormer.

Spatiotemporal graph modeling aims to model temporally varying and spatially correlated information sequences in the traffic network, serving as a significant approach to capture the complex, dynamic, and nonlinear changes of traffic systems. Consequently, it plays a crucial role in enhancing the reliability of intelligent transportation systems. Existing methods extract traffic data’s features by designing sophisticated spatiotemporal graph neural network models, and have reached a bottleneck. Furthermore, these spatiotemporal graph models adjust their parameters globally across different datasets, limiting their generalizability and universality. Meanwhile, the pre-training and fine-tuning paradigm has greatly progressed in natural language processing. This paper proposes a Generative Adversarial Pre-trained Spatial-Temporal Graph Convolutional Network (GAPSTGCN) to break through the bottleneck of spatiotemporal modeling of traffic data. Specifically, by developing a spatiotemporal graph encoder that incorporates domain knowledge about traffic system behavior and network science. The module can extract the dynamic change trend of data from the fusion process of spatiotemporal features, while introducing a time-gating mechanism that enhances its effectiveness in medium and long-term predictions. Moreover, we embed the encoding network into the generator and discriminator. Subsequently, we employ the concept of adversarial data generation to pre-train the spatiotemporal graph neural networks, enhancing the model’s knowledge emergence capabilities. Besides, we design an adapter for downstream task prediction to bridge the gap between pre-training and prediction tasks, aligning pre-trained features with downstream task features. Extensive experiments indicate that our model outperforms the state-of-the-art on the datasets. Experiments conducted on various types of traffic datasets further demonstrate the model’s generality.

The rapid advancement of Internet of Vehicles technology has led to massive growth in vehicular data generation, imposing strict computational demands for latency-sensitive applications. By offloading computational tasks to Road-side Units (RSUs), Vehicular Edge Computing (VEC) offers an efficient solution for those latency-sensitive applications. However, current task offloading schemes generally ignore the time-varying topology caused by vehicle mobility, which poses a risk of task interruption. Moreover, existing task offloading models primarily focus on serial tasks processing and fail to adequately account for the relationships among parallel tasks, leading to inefficient resource utilization and potential latency accumulation in multi-task VEC scenarios. To this end, we propose a mobility-awareCollaborativeTaskOffloading scheme forParallel tasks (CoTOP) in VEC. Integrated with vehicle mobility detection, a collaborative task offloading model is designed based on deep reinforcement learning, achieving effective coordination among RSUs to reduce task processing latency. Additionally, a task prioritization algorithm is incorporated to optimize resource allocation. Experimental results show that CoTOP significantly outperforms existing schemes in terms of task processing latency, energy consumption and completion ratio.

Edge intelligence capitalizes on heterogeneous edge resources to empower artificial intelligence–driven applications while mitigating dependence on centralized cloud infrastructures. In mobile edge computing (MEC) systems, resource scheduling constitutes a latency-critical task within inherently dynamic environments. Deep reinforcement learning (DRL) has been widely employed to address this challenge, as it enables agents to derive task-specific policies through continual interactions with the environment. Nevertheless, MEC environments exhibit substantial heterogeneity, and the relative importance of various resource types may fluctuate as the system evolves, thereby altering task objectives. Consequently, agents often necessitate retraining across new domains, the process that is computation ally intensive and time inefficient. To surmount these challenges, transfer reinforcement learning (TRL) has emerged as a promising paradigm that facilitates the transfer of agents' acquired knowledge from the source domain to the target domain. To the best of our knowledge, this is the first comprehensive survey dedicated to TRL applications in MEC systems. Specifically, we first elucidates the foundational principles of TRL, encompassing problem formulation and DRL methodologies. Subsequently, we systematically reviews and critically analyzes existing research on TRL and its applications in MEC systems. Finally, we discuss persistent technical limitations and delineates promising avenues for future exploration.

Deep unfolding network (DUN) has shed new light on multi-sequence MRI reconstruction, providing both high interpretability and acceptable performance. However, current approaches still suffer from the plight of information isolation, i.e., learning features of multi-suquences individually and leaving the mask departed from model updating. In this work, we propose a new unfolding solution, namely Information-coupled MRI Acceleration (IMA), to address the isolation issue. Concretely, two specific mechanisms are presented. On the one hand, the latent connections across different sequences are explicitly molded via two auxiliary matrices. While the first matrix is meticulously engineered to assemble the spatial details, the second one hammers at capturing the depth information conditioned on the enriched channels. On the other hand, following a deep analysis on the non-uniform distribution in low- and high-frequency components of the given mask, we elaborate a new unfolding flow using a progressive masking scheme, featuring a dilation-contraction mechanism during forward propagation of successive stages. Massive experiments are conducted under various sampling patterns and acceleration rates, whose results demonstrate that, without any sophisticated architectures, our IMA outperforms the current cutting-edge methods both visually and numerically.

As urban residents demand higher travel quality, vehicle dispatch has become a critical component of online ride-hailing services. However, current vehicle dispatch systems struggle to navigate the complexities of urban traffic dynamics, including unpredictable traffic conditions, diverse driver behaviors, and fluctuating supply and demand patterns. These challenges have resulted in travel difficulties for passengers in certain areas, while many drivers in other areas are unable to secure orders, leading to a decline in the overall quality of urban transportation services. To address these issues, this paper introduces GARLIC: a framework of GPT-Augmented Reinforcement Learning with Intelligent Control for vehicle dispatching. GARLIC utilizes multiview graphs to capture hierarchical traffic states, and learns a dynamic reward function that accounts for individual driving behaviors. The framework further integrates a GPT model trained with a custom loss function to enable high-precision predictions and optimize dispatching policies in real-world scenarios. Experiments conducted on two real-world datasets demonstrate that GARLIC effectively aligns with driver behaviors while reducing the empty load rate of vehicles.

Federated Graph Learning (FGL) has emerged as a key paradigm for distributed graph machine learning, enabling cross-domain graph collaborative modeling while preserving data privacy. However, existing methods face two major bottlenecks: the structural heterogeneity discrepancy of graph data among clients weakens the generalization ability of the global model; and model heterogeneity leads to inefficient knowledge sharing and complex global aggregation. To address these issues, we propose FedGVD, an efficient framework that constructs a global perspective through data condensation and server-side virtual node generation, which not only preserves the semantic equivalence of the original data but also avoids privacy leakage. Subsequently, by distributing low-dimensional generalizable knowledge for unidirectional distillation, FedGVD enables local models to absorb global knowledge without transmitting local parameters, thus breaking through the challenges of data and structural heterogeneity as well as model heterogeneity. This innovative approach ensures privacy-preserving and efficient federated graph collaboration. Experiments show that FedGVD maintains excellent performance in heterogeneous model scenarios while significantly improving communication efficiency, offering a new approach for privacy-preserving collaborative modeling in FGL. The code is available at https://github.com/Jasonxx4/FedGVD.

Video action recognition involves interpreting both global context and specific details to accurately identify actions. While previous models are effective at capturing spatiotemporal features, they often lack a focused representation of key action details. To address this, we introduce FocusVideo, a framework designed for refining video action recognition through integrated global and local feature learning. Inspired by human visual cognition theory, our approach balances the focus on both broad contextual changes and action-specific details, minimizing the influence of irrelevant background noise. We first employ learnable action queries to selectively emphasize action-relevant regions without requiring region-specific labels. Next, these queries are learned by a local action streaming branch that enables progressive query propagation. Moreover, we introduce a parameter-free feature interaction mechanism for effective multi-scale interaction between global and local features with minimal additional overhead. Extensive experiments demonstrate that FocusVideo achieves state-of-the-art performance across multiple action recognition datasets, validating its effectiveness and robustness in handling action-relevant details.

Point-of-Interest (POI) recommendation is a critical task in location-based services, which seeks to suggest a user's next potential location by leveraging their check-in history. However, the accuracy of such data is frequently compromised by noise, arising from factors such as inadvertent check-ins and the influence of recommendation biases, which hampers the accurate modeling of user preferences. To mitigate the impact of noisy implicit feedback, we propose the Diffusion-Based Implicit Feedback Modeling for Point-of-Interest Recommendation (DIPOI). First, a hybrid hypergraph convolutional network is utilized to perform multi-hop relational reasoning, enabling the discovery of deep-seated preferences, particularly for users with low activity levels. Subsequently, DIPOI integrates the spatial learning capacity of graph neural networks with the temporal modelling strengths of attention mechanisms. This integration enables the capture of the structural dependencies among POIs and the sequential dynamics of user behaviour. Finally, a diffusion model is incorporated to progressively denoise the latent representation space via a reverse generative process based on users' historical travel preferences, thereby increasing the fidelity of the implicit feedback modelling. Experiments conducted on three real-world datasets show that DIPOI consistently achieves higher recommendation accuracy than state-of-the-art baselines, validating its effectiveness and robustness. The implementation code is available at https://github.com/dengzhaolin/DIPOI.

Recommendation systems can establish vital connections between users and information. But they are still facing data sparsity and cold start issues. Cross-domain recommendation improves performance via knowledge transfer, yet often depends on data sharing, conflicting with privacy needs. Federated learning offers a privacy-preserving alternative, but still struggles with knowledge transfer and communication overhead. In this paper, we propose FedPce, a Prompt-based Federated Cross-domain Recommendation. FedPce improves cross-domain recommendation and promotes knowledge transfer by leveraging item description texts and prompt learning, while reducing communication overhead and accelerating convergence through low-rank gradient compression and elastic aggregation strategies. Experimental results show that FedPce performs well on multidomain recommendation tasks, balancing recommendation effect and privacy protection.

Large language models have demonstrated remarkable generalization capabilities and strong performance across various fields. Recent research has highlighted their significant potential in time series forecasting. However, time series data often exhibit complex periodic characteristics, posing a substantial challenge in enabling these models to effectively capture latent patterns. To address this challenge, we propose a novel framework, LLM-TPF, which leverages individuality and commonality fusion to enhance time series forecasting. In the frequency domain, periodic features are extracted to reveal the intrinsic periodicity of the data, while textual prototypes are used to indicate temporal trends. In the time domain, carefully designed prompts are employed to guide the models in comprehending global information. A commonality fusion mechanism further aggregates heterogeneous information across dimensions, and three distinct language models are utilized to independently process different types of information. Extensive real-world experiments demonstrate that LLM-TPF is a powerful tool for time series forecasting, achieving superior performance compared to state-of-the-art specialized models and exhibiting exceptional generalization ability in zero-shot scenarios. Code is available at https://github.com/switchsky/LLM-TPF.

Spatio-temporal graph modeling is widely applied to spatio-temporal data, analyzing the relationships between data to achieve accurate predictions. However, despite the excellent predictive performance of increasingly complex models, their intricate architectures result in significant memory overhead and computational complexity when handling spatio-temporal data, which limits their practical applications. To address these challenges, we propose a plug-and-play SubGraph Learning (SGL) method to reduce the memory overhead without compromising performance. Specifically, we introduce a SubGraph Partition Module (SGPM), which leverages a set of learnable memory vectors to select node groups with similar features from the graph, effectively partitioning the graph into smaller subgraphs. Noting that partitioning the graph may lead to feature redundancy, as overlapping information across subgraphs can occur. To overcome this, we design a SubGraph Feature Aggregation Module (SGFAM), which mitigates redundancy by averaging node features from different subgraphs. Experiments on four traffic network datasets of various scales demonstrate that SGL can significantly reduce memory overhead, achieving up to a 56.4\% reduction in average GPU memory overhead, while maintaining robust prediction performance. The source code is available at https://github.com/wengwenchao123/SubGraph-Learning.

Out-of-town trip recommendation aims to generate a sequence of Points of Interest (POIs) for users traveling from their hometowns to previously unvisited regions based on personalized itineraries, e.g., origin, destination, and trip duration. Modeling the complex user preferences--which often exhibit a two-fold nature of static and dynamic interests--is critical for effective recommendations. However, the sparsity of out-of-town check-in data presents significant challenges in capturing such user preferences. Meanwhile, existing methods often conflate the static and dynamic preferences, resulting in suboptimal performance. In this paper, we for the first time systematically study the problem of out-of-town trip recommendation. A novel framework SPOT-Trip is proposed to explicitly learns the dual static-dynamic user preferences. Specifically, to handle scarce data, we construct a POI attribute knowledge graph to enrich the semantic modeling of users’ hometown and out-of-town check-ins, enabling the static preference modeling through attribute relation-aware aggregation. Then, we employ neural ordinary differential equations (ODEs) to capture the continuous evolution of latent dynamic user preferences and innovatively combine a temporal point process to describe the instantaneous probability of each preference behavior. Further, a static-dynamic fusion module is proposed to merge the learned static and dynamic user preferences. Extensive experiments on real data offer insight into the effectiveness of the proposed solutions, showing that SPOT-Trip achieves performance improvement by up to 17.01%.

Temporal knowledge graph completion (TKGC) seeks to develop more comprehensive knowledge representations by addressing missing relationships and entities within temporal knowledge graphs (TKGs), thereby enhancing reasoning and predictive capabilities in downstream tasks. Nonetheless, real-world knowledge-such as the progression of social network interactions and the unfolding of news events-is inherently dynamic, resulting in substantial sparsity issues in TKGs that profoundly impair the performance of TKGC models. To overcome this challenge, we introduce the Adaptive Neighborhood Enhancement Layer (ANEL), a novel module that can be effortlessly integrated into existing TKGC models to substantially elevate the representation quality of sparse entities. ANEL first derives initial entity embeddings through a base model and then uncovers concealed semantic relationships between entities via a latent relation module, enriching the explicit relationships within the knowledge graph. Furthermore, ANEL incorporates an adaptive latent information adjustment component, which dynamically calibrates the influence of latent information based on the entity's relational structure: entities with fewer connections derive greater benefit from latent information, while entities with denser connections become less dependent on latent augmentation, ensuring precise and resilient representations. We conducted comprehensive experiments on four prominent benchmark datasets, and the results underscore the effectiveness and superiority of ANEL in TKGC tasks.

Missing traffic data is a common phenomenon in practice, which has serious implications for the development of downstream data-driven traffic research. Although numerous methods have been proposed to impute missing values, two limitations still exist: lack of complete ground truth and reliance on carefully constructed model structures. In this paper, we propose a novel deep neural imputation network framework named STIseq2seq, based on self-supervised contrastive learning. It follows a sequence-to-sequence architecture with attention mechanism to address the traffic data imputation problem. Specifically, the model consists of two key components: encoder and decoder, both composed of ST-block. Initially, we utilize dynamic graph convolution (DGC) and gated recurrent unit (GRU) to construct the fundamental unit, ST-block, for learning spatiotemporal dependencies. Then, a bidirectional ST-block is employed to model incomplete input data in the encoder; while a forward ST-block combined with an attention mechanism is used to accurately impute missing data in the decoder. Lastly, we devise a training method based on self-supervised contrastive learning, achieved through randomly shuffling the masking matrix, to enhance training effectiveness. Extensive experiments conducted on two available real-world traffic datasets demonstrate that our proposed STIseq2seq has superior and stable performance compared to multiple baseline models under different missing data scenarios with missing rates ranging from 20% to 80%. Visualization analysis of the imputation results further provides insights into the advantages of STIseq2seq.

The primary objective of graph node anomaly detection is to pinpoint rare patterns that display marked deviations from the typical one. Existing methods utilize Graph Convolutional Networks (GCNs) to model complex interactions in Heterogeneous Information Networks (HINs), typically homogenizing HINs using meta-paths to effectively focus on particular semantic scenarios. However, meta-paths excessively emphasize specific nodes and their connections on predefined paths, leading to the neglect of one-hop context-rich neighbors. Furthermore, the conversion from heterogeneous to homogeneous structures disrupts inherent relationships, resulting in an irreversible loss of direct links. Thus, we propose a dual-view-based H eterogeneous I nformation N etworks Node Ano maly Detection framework, HINAno, to mitigate structural loss. HINAno adopts a synergetic approach that balances local structural information with semantic richness, drawing from both the one-hop neighbor view and the meta-path view. Specifically, this dual-view utilizes hierarchical fusion mechanisms at node, type, and semantic levels to capture one-hop and multi-hop neighborhoods in a level-wise manner. In addition, HINAno adopts self-supervised contrastive learning and GCNs to amplify the gap between normal and abnormal nodes, thereby reducing the reliance on anomalous labels and enhancing the capability of anomaly detection. Finally, we successfully verify that the HINAno framework is effective and superior on four real-world datasets.

Multivariate time series anomaly detection plays a crucial role in industrial production. However, the inherent complexity and randomness of time series pose significant challenges. Furthermore, existing detection methods struggle to provide reliable explanations for outliers. To address these issues, this paper presents an unsupervised multivariate time series anomaly detection model named Point-Correlate Adversarial Transformer (PCAT). In this work, we leverage Transformer networks to capture the underlying correlations between different points in a time series and reconstruct the original sequence. By analyzing the correlation differences and reconstruction errors, we identify anomalies at the point level. Our model incorporates an adversarial structure, enabling unsupervised learning and enhancing the learning capability and robustness of the detection network. Experimental evaluations on four real-world datasets demonstrate the superiority of our approach over other state-of-the-art models in terms of detection delay and accuracy.

Knowledge graph completion (KGC) is dedicated to deducing absent facts within incomplete knowledge graphs (KGs). The majority of antecedent studies exclusively address the transductive scenario, wherein all entities are observed throughout the training process, which is impractical for the everyday reality with constantly emerging entities. Recent works that concern this issue mainly focus on representing emerging entities with seen neighbors while ignoring the semantics in the queried relations. In this paper, we introduce a model named DKGC, which incorporates two purposefully designed entity encoding modules to capture the local and global entity representations in a relation-specific manner and a sampling module with a diffusion process to explore the potential semantics further. Extensive experiments on benchmark datasets indicate the superiority of our model for inductive KGC.

Unmanned Aerial Vehicles (UAVs) can capture pictures of road conditions in all directions and from different angles by carrying high-definition cameras, which helps gather relevant road data more effectively. However, due to their limited energy capacity, drones face challenges in performing related tasks for an extended period. Therefore, a crucial concern is how to plan the path of UAVs and minimize energy consumption. To address this problem, we propose a multi-agent deep deterministic policy gradient-based (MADDPG) algorithm for UAV path planning (MAUP). Considering the energy consumption and memory usage of MAUP, we have conducted optimizations to reduce consumption on both fronts. Firstly, we define an optimization problem aimed at reducing UAV energy consumption. Secondly, we transform the defined optimization problem into a reinforcement learning problem and design MAUP to solve it. Finally, we optimize energy consumption and memory usage by reducing the number of neurons in the hidden layer of MAUP and conducting fine-grained pruning on connections. The final simulation results demonstrate that our method effectively reduces the energy consumption of UAVs compared to other methods.

The growing number of cars on city roads has led to an increase in traffic accidents, highlighting the need for traffic safety measures. Mobile trajectory anomaly detection is an important area of research that can identify unusual patterns or trajectories in urban environments and provide timely warnings to drivers to avoid accidents. However, there is a significant lack of research on the analysis of vehicle trajectory anomalies. To address this gap, we provide a comprehensive review of currently published papers on anomalous trajectories, highlighting important research trends and future directions. Besides, we innovatively classify trajectory anomalies into vehicle-based anomalies and driver-based anomalies according to whether they are caused by the driver’s behavior or not. The study further examines the existing challenges associated with analyzing anomalous trajectories and assesses the currently available solutions.

Unmanned aerial vehicle (UAV)-assisted mobile edge computing (MEC) has emerged as a valuable supplement to conventional MEC, offering unique advantages in temporary or emergency scenarios. However, the integration of UAVs into MEC introduces new challenges. First, the existing UAV-empowered MEC frameworks predominantly have not accounted for dependencies between the computing tasks. Second, the limited coverage area of UAVs and the sparse distribution of end devices result in a scarcity of training samples, hindering the effectiveness of data-driven approaches. To address these challenges, we introduce a novel UAV-assisted task offloading scheme designed to minimize the average task execution delay and energy consumption in vehicular networks. Initially, we craft a dependency-aware UAV-assisted MEC framework, where the task dependency and priority models are meticulously developed to illustrate the associations between the tasks. Subsequently, we devise an asynchronous federated deep reinforcement learning-based task offloading algorithm, incorporating an asynchronous federated optimization mechanism to enhance the data diversity, with a premise of ensuring data privacy. Our method has been rigorously tested using real traffic flow data and a directed acyclic graph task data set. Comparative experiments highlight the superiority of our framework, showcasing a substantial reduction in the average task delay and energy consumption.

The next point-of-interest (POI) recommendation aims to predict users’ future movements based on their historical trajectories. However, in reality, users may provide uncertain check-in records, resulting in uploaded data that lack precise location information and is instead ambiguous. Despite this challenge, only a limited number of studies have addressed this issue, often overlooking the intricate interactions among users, POIs, and POI categories. To that end, we propose a novel model called knowledge-graph-enhanced transformer (KGNext). KGNext leverages transition and interaction graphs derived from our constructed transitional-interactive knowledge graph (TIKG) to uncover both general movement patterns and varied user preferences regarding POIs and POI categories. Furthermore, KGNext integrates comprehensive contextual information from historical trajectories with TIKG to generate user trajectory embeddings. These encoded features are then utilized by a transformer model to provide fine-grained predictions of the next POI. Experimental results on three real-world datasets demonstrate the superiority of KGNext.

Urban overtourism results in heavy traffic, degraded tourist experiences, and overloaded infrastructure. Detecting urban overtourism at the early stage is important to minimize the adverse effects. However, urban overtourism detection (UOD) is a challenging task due to ambiguity, sparsity, and complex spatiotemporal relations of overtourism. In this article, we propose a novel UOD framework based on graph temporal convolutional networks (TCNs) to tackle the challenges mentioned above. More specifically, we propose the grid overtourism mode (GOM) to detect urban overtourism on a grid level and propose the overtourism detection mechanism, which gives a quantitative definition of overtourism and screens out the regions where overtourism may occur as candidate regions. Then, we construct the GOM graphs of the candidate regions. Next, we employ the graph TCNs to model the complex spatiotemporal relations of urban overtourism and predict the future GOM graph at the next time interval. Finally, we calculate the urban overtourism scores based on the prediction results. The experiments are conducted based on a real-world dataset. The evaluation results demonstrate the effectiveness of our methods.

Vehicle driving behavior analysis and detection tasks have become an indispensable part of intelligent transportation systems. Accurate pattern recognition of potential anomalies during the movement of entities is crucial for improving transportation efficiency. Current methods typically analyze vehicle trajectories independently without considering potential interactions among vehicles. To address this limitation, some studies have integrated graph attention mechanisms to capture the influence of neighboring vehicles during the aggregation process. However, Graph Attention Networks (GATs) are constrained by the univariate nature of attention heads and coefficients, thus lacking flexibility. In this work, we not only consider the social dynamics among neighboring vehicles but also delve into the limitations of GAT models. We propose a Vehicular Social Dynamics Anomaly Detection (VSD-AD) model based on the Recurrent Multi-Mask Aggregator (MMA) enabled Variational AutoEncoder (VAE) architecture to maximize the learning of relational embeddings among neighbors in a highway vehicle network. Furthermore, we apply Node Feature Quantisation (NFQ) to the encoder output to mitigate the complexity of neighbor relationships. Our model is flexible and customizable for different highway scenarios, suitable for large-scale highway vehicle video data. To validate real-world applicability, we further assess its performance on both the simulated dataset and real-world traffic dataset, where our model outperforms other mainstream methods in terms of detection performance.

The spatio-temporal prediction task in the transportation network is the core of the solutions for various traffic problems. On one hand, the mobility pattern in traffic can be reflected in the travel behavior of the crowd. In most traffic prediction tasks, the importance of the mobility pattern is often overlooked. On the other hand, traffic prediction also has a variety of predicting scenarios, including short-term and long-term prediction, and relevant research cannot solve the problems under the two scenarios at the same time. In view of the problem of existing work, we propose a multi-pattern traffic prediction framework, MPGNNFormer. First, we construct a new bus stop distance network to model the relationships between stops. Then, we use a graph neural network-based deep clustering method to extract the bus stop mobility pattern. Finally, we design a transformer-based spatio-temporal prediction model (STGNNFormer) to predict bus stop flow by taking full advantage of time dependency and space dependency. After that, we conduct a series of experiments to evaluate and test them on the real bus dataset, including analyzing mobility patterns and comparing prediction results. The experimental results prove that MPGNNFormer can improve the calculation efficiency in the prediction scene while ensuring prediction accuracy in the stop-flow prediction of the transportation network.

Intelligent Transportation Systems (ITS) are essential and play a key role in improving road safety, reducing congestion, optimizing traffic flow and facilitating the development of smart cities. The collection of data from ITS and its transformation into information is challenged by the presence of missing data in datasets. Timely and effective handling of missing data is crucial to facilitate intelligent decision making. In response to this need, numerous researchers have proposed various techniques for dealing with missing data, with commendable results in different scenarios. Therefore, this survey aims to provide a well-organized and thorough overview of the research related to imputation of missing data in traffic. Our purposes are at least fourfold. First, we discuss the background of missing data and highlight the value of traffic data imputation. Second, we present a comprehensive list of missing patterns, open data, widely used evaluation metrics and performance goals for this issue. Third, we categorize related studies into three parts: interpolation-based, statistical learning-based, prediction-based methods, and provide a secondary classification within each category to better understand the characteristics and limitations of each method. Finally, we identify future research directions to advance the understanding of traffic data imputation.

The rapid development of technology such as virtual reality and augmented reality, coupled with the reduced direct contact due to the COVID-19 pandemic, has led to the emergence of a more advanced mode of interaction: air handwriting. This new form of human-computer interaction allows users to input text by writing in the air freely. However, deploying and applying existing air handwriting recognition systems in real-world scenarios still presents challenges, particularly in real-time performance, privacy protection, and label scarcity. To address these challenges, we propose a federated active learning framework called FedAWR for air handwriting recognition tasks. FedAWR utilizes distributed learning to train a shared global model in the cloud from multiple user devices at the network's edge, while keeping the user's handwritten data local to ensure privacy. In addition, FedAWR employs an interactive active learning strategy to collect user-provided annotations for iterative training during the online federated learning process, bootstrapping personalized models for each client. To further enhance interactivity and real-time performance, we designed a lightweight recognition model, which is integrated into FedAWR. Finally, extensive experiments were conducted on real-world air handwritten datasets to validate the superiority of FedAWR.

The recommendation of taxi pickup locations plays an important role for drivers in carrying passengers efficiently. In addition, the emergence of the Internet of Vehicles provides technical support for it. However, existing recommendation methods do not model dynamic global positioning system information well and in real-time. In this article, we propose a spatio-temporal digraph convolutional network (STDCN) model. First, the pickup and drop-off locations are modeled into a directed spatio-temporal graph as input to the model. The correlation between each node is calculated as a unified edge weight based on the gray relational analysis. Then, the STDCN is used for dynamic spatio-temporal feature extraction. Finally, the edge-cloud collaboration framework is adopted to recommend local taxi pickup locations in real-time. The experimental results show that the proposed method is better than competing methods in terms of effectiveness and efficiency, and it shows good industrial conversion application prospects.

Predicting traffic accidents can help traffic management departments respond to sudden traffic situations promptly, improve drivers’ vigilance, and reduce losses caused by traffic accidents. However, the causality of traffic accidents is complex and difficult to analyze. Most existing traffic accident prediction methods do not consider the dynamic spatio-temporal correlation of traffic data, which leads to unsatisfactory prediction accuracy. To address this issue, we propose a multi-task learning framework (TAP) based on the Spatio-temporal Variational Graph Auto-Encoders (ST-VGAE) for traffic accident profiling. We firstly capture the dynamic spatio-temporal correlation of traffic conditions through a spatio-temporal graph convolutional encoder and embed it as a low-latitude vector. Then, we use a multi-task learning scheme to combine external factors to generate the traffic accident profiling. Furthermore, we propose a traffic accident profiling application framework based on edge computing. This method increases the speed of calculation by offloading the calculation of traffic accident profiling to edge nodes. Finally, the experimental results on real datasets demonstrate that TAP outperforms other state-of-the-art baselines.

Intelligent vehicle applications, such as autonomous driving and collision avoidance, put forward a higher demand for precise positioning of vehicles. The current widely used global navigation satellite systems (GNSS) cannot meet the precision requirements of the submeter level. Due to the development of sensing techniques and vehicle-to-infrastructure (V2I) communications, some vehicles can interact with surrounding landmarks to achieve precise positioning. Existing work aims to realize the positioning correction of common vehicles by sharing the positioning data of sensor-rich vehicles. However, the privacy of trajectory data makes it difficult to collect and train data centrally. Moreover, uploading vehicle location data wastes network resources. To fill these gaps, this article proposes a vehicle cooperative positioning (CP) system based on federated learning (FedVCP), which makes full use of the potential of social Internet of Things (IoT) and collaborative edge computing (CEC) to provide high-precision positioning correction while ensuring user privacy. To the best of our knowledge, this article is the first attempt to solve the privacy of CP from a perspective of federated learning. In addition, we take the advantages of local cooperation through vehicle-to-vehicle (V2V) communications in data augmentation. For individual differences in vehicle positioning, we utilize transfer learning to eliminate the impact of such differences. Extensive experiments on real data demonstrate that our proposed model is superior to the baseline method in terms of effectiveness and convergence speed.

Mobile network operators (MNOs) allocate computing and caching resources for mobile users by deploying a central control system. Existing studies mainly use programming and heuristic methods to solve the resource allocation problem, which ignores the energy cost problem that is really significant to the MNO. To solve this problem, in this article, we design a joint computing and caching framework by integrating deep deterministic policy gradient (DDPG) algorithm. Especially, we focus on the Internet of Vehicles scenario, which needs the support of mobile network provided by MNO. We first formulate an optimization problem to minimize MNO’s energy cost by considering the computation and caching energy costs jointly. Then, we turn the formulated problem into a reinforcement learning problem and utilize DDPG methods to solve this problem. The final simulation result shows that our solution can reduce energy costs by more than 15%, while ensuring the tasks can be completed on time.

Vehicle-to-vehicle interaction and collaboration can provide us with a large number of mobile traffic trajectories that can be used to analyze driving behavior. In this article, we propose a spatio-temporal cost combination based framework for taxi driving fraud detection. First, the point of interest where taxis interact and collaborate with collaborative Internet of Vehicles participants is identified, and a baseline trajectory model is built to determine the typical trajectory distribution. Second, a statistical model is used to calculate the travel distribution, travel time, and travel cost. At the same time, the taxi trajectory points are converted into evolving graphs to detect the abnormality of the local road segment. Then, we can analyze the causes of outlier trajectories combined with the perception of abnormal road environments. Finally, the trajectories of real taxis were used to evaluate outliers, which proves the effectiveness and efficiency of the method.

Modern enterprises attach much attention to the selection of commercial locations. With the rapid development of urban data and machine learning, we can discover the patterns of human mobility with these data and technology to guide commercial district discovery. In this article, we propose an unsupervised commercial district discovery framework via embedding space clustering on graph convolution networks to solve the problem of commercial district discovery. Specifically, the proposed framework aggregates human mobility features according to geographic similarity by graph convolution networks. Based on the graph convolution network embedding space, we apply hierarchical clustering to mine the latent functional regions hidden in different human patterns. Then, with the kernel density estimation, we can obtain the semantic labels for the clustering results to discover commercial districts. Finally, we analyze the multisource data of the Xiaoshan District and Chengdu City, and experiments verify the effectiveness of our framework.

OBJECTIVE: Preterm birth is the leading cause of neonatal morbidity and mortality. Early identification of high-risk patients followed by medical interventions is essential to the prevention of preterm birth. Based on the relationship between uterine contraction and the fundamental electrical activities of muscles, we extracted effective features from EHG signals recorded from pregnant women, and use them to train classifiers with the purpose of providing high precision in classifying term and preterm pregnancies. METHODS: To characterize changes from irregularity to coherence of the uterine activity during the whole pregnancy, network representations of the original electrohysterogram (EHG) signals are established by applying the Horizontal Visibility Graph (HVG) algorithm, from which we extract network degree density and distribution, clustering coefficient and assortativity coefficient. Concerns on the interferences of different noise sources embedded in the EHG signal, we apply Short-Time Fourier Transform (STFT) to expand the original signal in the time-frequency domain. This allows a network representation and the extraction of related features on each frequency component. Feature selection algorithms are then used to filter out unrelated frequency components. We further apply the proposed feature extraction method to EHG signals available in the Term-Preterm EHG database (TPEHG), and use them to train classifiers. We adopt the Partition-Synthesis scheme which splits the original imbalanced dataset into two sets, and synthesizes artificial samples separately within each subset to solve the problem of dataset imbalance. RESULTS: The optimally selected network-based features, not only contribute to the identification of the essential frequency components of uterine activities related to preterm birth, but also to improved performance in classifying term/preterm pregnancies, i.e., the SVM (Support Vector Machine) classifier trained with the available samples in the TPEHG gives sensitivity, specificity, overall accuracy, and auc values as high as 0.89, 0.93, 0.91, and 0.97, respectively.

Congestion recognition is necessary for vehicle routing, traffic control, and many other applications in intelligent transportation systems. Besides, traffic facilities in the three-dimensional road network, which contains the fundamental spatiotemporal features for congestion recognition, provides multi-source traffic information. To exploit these traffic big data, in this paper, we propose an attention mechanism-based digraph convolution network (ADGCN) enabled framework to tackle the congestion recognition problem. It can be divided into two parts, spatial relevance modeling and temporal relevance modeling. At first, the representation incorporates spatiotemporal traffic information with the three-dimensional urban network, and partially decouples the global network topology to a single-knot digraph. Then a digraph-based convolution network is used to capture high-order spatial features. Finally, to proceed with time-series features, the multi-modal attention mechanism is introduced to catch the long-range temporal dependence and the congestion classifier is defined accordingly. This distinguishes the proposed model from the conventional congestion recognition methods. Comprehensive experiments are conducted based on real traffic data. The results demonstrate the advantages of the proposed framework over the existing spatiotemporal analysis methods.

Due to the incomplete coverage and failure of traffic data collectors during the collection, traffic data usually suffers from information missing. Achieving accurate imputation is critical to the operation of transportation networks. Existing approaches usually focus on the characteristic analysis of temporal variation and adjacent spatial representation, and the consideration of higher‐order spatial correlations and continuous data missing attracts more attentions from the academia and industry. In this paper, by leveraging motif‐based graph aggregation, we propose a spatiotemporal imputation approach to address the issue of traffic data missing. First, through motif discovery, the higher‐order graph aggregation model was presented in traffic networks. It utilized graph convolution network (GCN) to polymerize the correlated segment attributes of the missing data segments. Then, the multitime dimension imputation model based on bidirectional long short‐term memory (Bi‐LSTM) incorporated the recent, daily‐periodic, and weekly‐periodic dependencies of the historical data. Finally, the spatial aggregated values and the temporal fusion values were integrated to obtain the results. We conducted comprehensive experiments based on the real‐world dataset and discussed the case of random and continuous data missing by different time intervals, and the results showed that the proposed approach was feasible and accurate.

During the outbreak of the Coronavirus disease 2019 (COVID-19), while bringing various serious threats to the world, it reminds us that we need to take precautions to control the transmission of the virus. The rise of the Internet of Medical Things (IoMT) has made related data collection and processing, including healthcare monitoring systems, more convenient on the one hand, and requirements of public health prevention are also changing and more challengeable on the other hand. One of the most effective nonpharmaceutical medical intervention measures is mask wearing. Therefore, there is an urgent need for an automatic real-time mask detection method to help prevent the public epidemic. In this article, we put forward an edge computing-based mask (ECMask) identification framework to help public health precautions, which can ensure real-time performance on the low-power camera devices of buses. Our ECMask consists of three main stages: 1) video restoration; 2) face detection; and 3) mask identification. The related models are trained and evaluated on our bus drive monitoring data set and public data set. We construct extensive experiments to validate the good performance based on real video data, in consideration of detection accuracy and execution time efficiency of the whole video analysis, which have valuable application in COVID-19 prevention.

License plate is an essential characteristic to identify vehicles for the traffic management, and thus, license plate recognition is important for Internet of Vehicles. Since 5G has been widely covered, mobile devices are utilized to assist the traffic management, which is a significant part of Industry 4.0. However, there have always been privacy risks due to centralized training of models. Also, the trained model cannot be directly deployed on the mobile device due to its large number of parameters. In this article, we propose a federated learning-based license plate recognition framework (FedLPR) to solve these problems. We design detection and recognition model to apply in the mobile device. In terms of user privacy, data in individuals is harnessed on their mobile devices instead of the server to train models based on federated learning. Extensive experiments demonstrate that FedLPR has high accuracy and acceptable communication cost while preserving user privacy.

As a new computing paradigm, edge computing emerges in various fields. Many tasks previously relied on cloud computing are distributed to various edge devices that cooperate to complete the tasks. However, circumstantial factors in the edge network (e.g., functionality, transmission efficiency, and resource limitation) become more complex than those in cloud computing. Consequently, there is instability that cannot be ignored in the cooperation between the edge devices. In this article, we propose a novel framework to optimize edge cooperative network (ECN), called ECN-Opt, to improve the performance of edge computing tasks. Specifically, we first define the evaluation metrics for cooperation. Next, the cooperation of an ECN is optimized to improve the performance of specific tasks. Extensive experiments using real datasets from wearable sensors on the players in soccer teams demonstrate that our ECN-Opt framework performs well, and it also validate the effectiveness of the proposed optimization algorithm.

Aiming at the problem of intersection signal control, a method of traffic phase combination and signal timing optimization based on the improved K-medoids algorithm is proposed. Firstly, the improvement of the traditional K-medoids algorithm embodies in two aspects, namely, the selection of the initial medoids and the parameter k , which will be applied to the cluster analysis of historical saturation data. The algorithm determines the initial medoids based on a set of probabilities calculated from the distance and determines the number of clusters k based on an exponential function, weight adjustment, and elbow ideas. Secondly, a phase combination model is established based on the saturation and green split data, and the signal timing is optimized through a bilevel programming model. Finally, the algorithm is evaluated over a certain intersection in Hangzhou, and results show that this algorithm can reduce the average vehicle delay and queue length and improve the traffic capacity of the intersection in the peak hour.

The car-following model describes the microscopic behavior of the vehicle. However, the existing car-following models set the drivers’ reaction time to a fixed value without considering its dynamics. In order to improve the accuracy of car-following model, this paper proposes Deep Feature Learning-based Car-Following Model (DeepCF), a car-following model based on fatigue driving and Generative Adversarial Networks (GAN). The model is composed of the drivers’ reaction time model and the car-following decision algorithm. First, we regard driving fatigue as the starting point to study the influence of driving time and the acceleration of the preceding vehicle on the drivers’ reaction time, and develop a coarse-grained drivers’ reaction time model. Secondly, considering the impact of fatigue driving on car-following decisions, we utilize GAN to generate a driving decision database based on reaction time and use Euclidean distance as a decision search indicator. Finally, we conduct experiments on a real data set, and the results indicate that our DeepCF model is superior to baseline models.

In this paper, we consider the robust face recognition problem via iterative re-constrained group sparse classifier (IRGSC) with adaptive weights learning. Specifically, we propose a group sparse representation classification (GSRC) approach in which weighted features and groups are collaboratively adopted to encode more structure information and discriminative information than other regression based methods. In addition, we derive an efficient algorithm to optimize the proposed objective function, and theoretically prove the convergence. There are several appealing aspects associated with IRGSC. First, adaptively learned weights can be seamlessly incorporated into the GSRC framework. This integrates the locality structure of the data and validity information of the features into l2,p-norm regularization to form a unified formulation. Second, IRGSC is very flexible to different size of training set as well as feature dimension thanks to the l2,p-norm regularization. Third, the derived solution is proved to be a stationary point (globally optimal if p ≥ 1). Comprehensive experiments on representative data sets demonstrate that IRGSC is a robust discriminative classifier which significantly improves the performance and efficiency compared with the state-of-the-art methods in dealing with face occlusion, corruption, and illumination changes, and so on.

We propose a novel dynamic traffic signal coordination method that takes account of the special traffic flow characteristics of urban arterial roads. The core of this method includes a control area division module and a signal coordination control module. Firstly, we analyze and model the influences of segment distance, traffic flow density, and signal cycle time on the correlation degree between two neighboring intersections. Then, we propose a fuzzy computing method to estimate the correlation degree based on a hierarchical structure and a method to divide the control area of urban arterial roads into subareas based on correlation degrees. Subarea coordination control arithmetic is used to calculate the public cycle time of the control subarea, up-run offset and down-run offset of the section, and the split of each intersection. An application of the method in Shaoxing City, Zhejiang Province, China shows that the method can reduce the average travel time and the average stop rate effectively.

On the basis of distributed traffic control framework, fuzzy theory, and artificial neural networks technique, the road network traffic intelligent coordination control technique with bus priority was proposed. The whole road network was regarded as a large-scale system, and the subsystems were the intersections. Multiphase intelligent signal controller that controlled its own traffic and cooperated with its neighbors was installed at each intersection. By exchanging information collected from its social vehicle detectors and the bus detection and location devices, and cooperating with adjacent signal controllers, social vehicle coordination and bus priority in the whole road network were realized. Bus priority module, green observation module, and phase switch module comprised the hard core of the controller. In each module, the fuzzy rule base system was designed in detail. To improve the control system's robusticity, the fuzzy relations of the three modules were implemented by one neural network. The target of this proposed method was to maximize the possibility for vehicles to depart from the upstream intersection, and the traveling bus nearby the local intersection to pass the local intersection without stoppage while the utility efficiency of the green signal time was at a relatively high level. The actual application shows that the proposed method can decrease the average vehicle delay and average travel time effectively.