Heterophily is a prevalent property of real-world graphs and is well known to impair the performance of homophilic Graph Neural Networks (GNNs). Prior work has attempted to adapt GNNs to heterophilic graphs through non-local neighbor extension or architecture refinement. However, the fundamental reasons behind misclassifications remain poorly understood. In this work, we take a novel perspective by examining recurring inductive subgraphs, empirically and theoretically showing that they act as spurious shortcuts that mislead GNNs and reinforce non-causal correlations in heterophilic graphs. To address this, we adopt a causal inference perspective to analyze and correct the biased learning behavior induced by shortcut inductive subgraphs. We propose a debiased causal graph that explicitly blocks confounding and spillover paths responsible for these shortcuts. Guided by this causal graph, we introduce Causal Disentangled GNN (CD-GNN), a principled framework that disentangles spurious inductive subgraphs from true causal subgraphs by explicitly blocking non-causal paths. By focusing on genuine causal signals, CD-GNN substantially improves the robustness and accuracy of node classification in heterophilic graphs. Extensive experiments on real-world datasets not only validate our theoretical findings but also demonstrate that our proposed CD-GNN outperforms state-of-the-art heterophily-aware baselines.

Traffic prediction plays a pivotal role in contemporary web technologies, motivating various intelligent web services such as route planning and remote traffic management. Many recent proposals that target deep learning for traffic prediction solely leverage historical traffic observations to predict future ones. However, traffic prediction is always susceptible to different factors such as road networks and social events, exhibiting different modalities. Most existing methods focus on a single modality, failing to capture the comprehensive traffic patterns among various factors, resulting in sub-optimal performance. Web-sourced geo-images, e.g., satellite imagery, encompass comprehensive contextual information and offer an effective way to represent diverse modalities. To unleash the power of such geo-images, we propose VisionST, a Vision-augmented Spatial-Temporal Neural Network, which coordinates cross-modal traffic prediction with interactive geo-image encoding. To bolster resilience against highly intricate and overlapping traffic patterns, VisionST features a visual semantic extraction mechanism and a pattern-guided aggregation mechanism. The former extracts node-level visual tokens and node-to-node visual relation patterns from geo-referenced images. The latter generates relation patterns that encompass visual, spatial, and temporal aspects, constraining nodes to interact with these relation patterns for contextual information interaction. Extensive experiments on real large-scale datasets offer insight into the effectiveness of the proposed solutions, showing that VisionST consistently outperforms state-of-the-art baselines.

Urban region profiling, the task of characterizing geographical areas, is crucial for urban planning and resource allocation. However, existing research in this domain faces two significant limitations. First, most methods are confined to single-task prediction, failing to capture the interconnected, multi-faceted nature of urban environments where numerous indicators are deeply correlated. Second, the field lacks a standardized experimental benchmark, which severely impedes fair comparison and reproducible progress. To address these challenges, we first establish a comprehensive benchmark for multi-task urban region profiling, featuring multi-modal features and a diverse set of strong baselines to ensure a fair and rigorous evaluation environment. Concurrently, we propose UrbanMoE, the first sparse multi-modal, multi-expert framework specifically architected to solve the multi-task challenge. Leveraging a sparse Mixture-of-Experts architecture, it dynamically routes multi-modal features to specialized sub-networks, enabling the simultaneous prediction of diverse urban indicators. We conduct extensive experiments on three real-world datasets within our benchmark, where UrbanMoE consistently demonstrates superior performance over all baselines. Further in-depth analysis validates the efficacy and efficiency of our approach, setting a new state-of-the-art and providing the community with a valuable tool for future research in urban analytics.

Time series anomaly detection aims to identify samples that deviate from a normal sample distribution in a time series, enabling various web-centric applications. Most existing approaches are static, targeting pre-defined types of anomalies. These methods thus fail to work well on streaming time series with changing data distributions and anomaly formats. To contend with such streaming time series and to accommodate memory constraints, we propose the first data-efficient streaming time series anomaly detection framework, called DESS. To accumulate historical knowledge, DESS includes a novel evolving proxy generation module to synthesize a small but informative proxy summarizing the historical data, facilitating data efficiency. Next, DESS employs an innovative heterogeneous temporal feature extraction module to explicitly capture correlations of multi-level time series semantics. Finally, DESS enables fast streaming anomaly detection by employing a parameter-efficient training scheme that only activates a subset of lightweight parameters while ensuring performance. Extensive experiments on real data offer insight into the effectiveness and efficiency of DESS, showing that it is able to outperform the best baselines by up to 17.53% while reducing the training time by up to 64.88%.

Spatio-temporal data proliferates in numerous real-world domains, such as transportation, weather, and energy. Spatio-temporal deep learning models aims to utilize useful patterns in such data to support tasks like prediction, imputation, and anomaly detection. However, previousone-to-onedeep learning models designed for specific tasks typically require separate training for each use case, leading to increased computational and storage costs. To address this issue,one-to-manyspatio-temporal foundation models have emerged, offering a unified framework capable of solving multiple spatio-temporal tasks. These foundation models achieve remarkable success by learning general knowledge with spatio-temporal data or transferring the general capabilities of pre-trained language models. While previous surveys have explored spatio-temporal data and methodologies separately, they have ignored a comprehensive examination of how foundation models are designed, selected, pre-trained, and adapted. As a result, the overall pipeline for spatio-temporal foundation models remains unclear. To bridge this gap, we innovatively provide an up-to-date review of previous spatio-temporal foundation models from the pipeline perspective. The pipeline begins with an introduction to different types of spatio-temporal data, followed by details of data preprocessing and embedding techniques. The pipeline then presents a novel data property taxonomy to divide existing methods according to data sources and dependencies, providing efficient and effective model design and selection for researchers. On this basis, we further illustrate the training objectives of primitive models, as well as the adaptation techniques of transferred models. Overall, our survey provides a clear and structured pipeline to understand the connection between core elements of spatio-temporal foundation models while guiding researchers to get started quickly. Additionally, we introduce emerging opportunities such as multi-objective training in the field of spatio-temporal foundation models, providing valuable insights for researchers and practitioners.

With the proliferation of GPS-equipped edge devices, huge trajectory data are generated and accumulated in various domains, driving numerous urban applications. However, due to the limited data acquisition capabilities of edge devices, many trajectories are often recorded at low sampling rates, reducing the effectiveness of these applications. To address this issue, we aim to recover high-sample-rate trajectories from low-sample-rate ones enhancing the usability of trajectory data. Recent approaches to trajectory recovery often assume centralized data storage, which can lead to catastrophic forgetting, where previously learned knowledge is entirely forgotten when new data arrives. This not only poses privacy risks but also degrades performance in decentralized settings where data streams into the system incrementally. To enable decentralized training and streaming trajectory recovery, we propose aLightweight incremental framework for federatedTrajectoryRecovery, called LightTR+, which is based on a client-server architecture. Given the limited processing capabilities of edge devices, LightTR+ includes a lightweight local trajectory embedding module that enhances computational efficiency without compromising feature extraction capabilities. To mitigate catastrophic forgetting, we propose an intra-domain knowledge distillation module. Additionally, LightTR+ features a meta-knowledge enhanced local-global training scheme, which reduces communication costs between the server and clients, further improving efficiency. Extensive experiments offer insight into the effectiveness and efficiency of LightTR+.

Time series anomaly detection aims to identify samples that deviate from a normal distribution in a time series, which is practically important to a variety of real-world applications. Existing approaches are mostly centralized and domain-specific, and thus they are hard to generalize to time series of different domains that are decentralized due to the privacy concerns and the resulting data silos across institutions. To bridge this gap, we propose FAST-MAD, the first resource-aware framework for efficient federated time series anomaly detection. Operating under a client-server architecture, different clients in FAST-MAD can handle time series from distinct domains. In particular, FAST-MAD first employs a multi-resolution transformation module to capture hierarchical local semantics, frequency-oriented patching as well as inter-time-series interaction. An LLM serves as the main body of the local model for each client, owing to its strong knowledge transfer capabilities. Further, an adaptive modularized separation mechanism is integrated with sharded federated training to reduce computational costs, which innovatively splits the LLM into a U-shaped architecture. To address data heterogeneity across different clients, we propose a decomposed client-server alignment mechanism, featuring a tailored low-rank parameter decomposition that extracts domain-common knowledge. Extensive experiments on multiple cross-domain time series datasets offer insight into the effectiveness and efficiency of FAST-MAD, which outperforms SOTA baselines by up to 10.25% in terms of F1-score and reduces the training time by 40.93%.

Spatio-temporal data intelligence, which includes sensing, managing, and mining large-scale data across space and time, plays a pivotal role in understanding complex systems in real-world applications, such as urban computing and smart cities. With the rapid evolution of foundation models and their growing potential to transform spatio-temporal analytics, we propose a comprehensive half-day workshop (with at least 5 accepted papers, 3 keynote talks, 1 panel discussion, and over 50 attendees) at CIKM 2025, catering to professionals, researchers, and practitioners who are interested in spatio-temporal data intelligence and foundation models to address real-world challenges. The workshop will not only offer a platform for knowledge exchange but also acknowledge outstanding contributions through a distinguished Best Paper Award. A dedicated panel discussion will explore recent advances, emerging trends, and open challenges in integrating spatio-temporal data and emerging machine learning techniques, fostering dialogue between academia and industry. Note that this will be the eleventh time that our core members have organized a similar workshop. The previous 10 workshops were hosted in top-tier data mining and management venues, e.g., SIGKDD, WWW, and IJCAI, each of which attracted over 60 participants and 25 submissions on average.

Traffic accidents are one of the leading causes of death in the world. As an important part of the design of traffic roads, tunnels bring convenience but also have huge safety risks. To monitor road safety in real time and give timely warnings for drivers in tunnels, where the light is dark, the space is limited, and the signal is unstable, we study the problem of traffic safety evaluation based on macroscopic traffic features in road tunnels. In particular, we transform the problem into a four-classification problem. To overcome the long collection cycle of traffic crash data, we use the time-to-collision index as the standard for dividing safety levels of road sections in tunnels. To achieve the goal of collecting data in real time under the environment constraints of tunnels, we use macroscopic traffic features as input in our model. Specifically, we design a deep learning model, where the lane block can extract the interaction information of sequential road segments in the same lane, and the prediction block can integrate the results of the individual prediction of each lane and the overall prediction. An extensive emprical study with real data offers insight into the effectiveness and efficiency of the proposed model.

With the rapid evolution of sensing techniques and the proliferation of mobile devices, spatial crowdsourcing (SC) has gained significant attention in both academia and industry. SC involves assigning location-based tasks to mobile workers, with task recommendation playing a key role in helping workers identify suitable and appealing tasks. However, most existing studies focus on task completion rate, worker satisfaction, or efficiency, without consideration of the environmental impact, e.g., pollutant emissions from the increased vehicle usage associated with SC applications like Uber, Lyft, and FoodPanda. In this study, we consider a novel problem of sustainable task recommendation in SC, which aims to minimize the environmental footprint (i.e., pollution) while maintaining acceptable levels of task completion, worker satisfaction, and overall task recommendation efficiency. We develop an innovative Sustainability-Oriented Task Recommendation framework encompassing two major components: speed-driven pollutant emission estimation and task recommendation. Specifically, the pollutant emission estimation component aims to estimate future pollutant emissions based on worker trajectories and speeds, using a context-enhanced spatio-temporal network for road speed prediction. In the task recommendation component, we provide a completion-sensitive recommendation algorithm to maximize the expected number of completed tasks. Further, we design an efficient emission-optimized KM ranking algorithm to minimize emissions. Experiments on real data offer insight into the effectiveness and efficiency of the proposals, providing valuable insights into its potential for sustainable spatial crowdsourcing.

Multivariate time series forecasting (MTSF) endeavors to predict future observations given historical data, playing a crucial role in time series data management systems. With advancements in large language models (LLMs), recent studies employ textual prompt tuning to infuse the knowledge of LLMs into MTSF. However, the deployment of LLMs often suffers from low efficiency during the inference phase. To address this problem, we introduce TimeKD, an efficient MTSF framework that leverages the calibrated language models and privileged knowledge distillation. TimeKD aims to generate high-quality future representations from the proposed cross-modality teacher model and cultivate an effective student model. The cross-modality teacher model adopts calibrated language models (CLMs) with ground truth prompts, motivated by the paradigm of Learning Under Privileged Information (LUPI). In addition, we design a subtractive cross attention (SCA) mechanism to refine these representations. To cultivate an effective student model, we propose an innovative privileged knowledge distillation (PKD) mechanism including correlation and feature distillation. PKD enables the student to replicate the teacher's behavior while minimizing their output discrepancy. Extensive experiments on real data offer insight into the effectiveness, efficiency, and scalability of the proposed TimeKD.

Movement trajectory similarity computation is important when supporting functionalities such as outlier detection and prediction that may, in turn, fuel a variety of transportation-related applications. Recent trajectory similarity learning solutions often assume that trajectories are available at a central location. Yet, we are witnessing the decentralized collection of increasingly massive volumes of trajectories due to the deployment of edge devices. To enable decentralized training and improved privacy, we propose a federated trajectory similarity learning framework that features privacy-preserving clustering based on a client-server architecture. The framework encompasses local, client-side trajectory preprocessing and representation learning. This is combined with a novel privacy-preserving clustering mechanism that ensures consistent model updates between clients and the server, thus alleviating the effects of trajectory heterogeneity across clients. In addition, the framework features a hierarchical central aggregation mechanism that supports clustered federated learning. Experiments on real data offer evidence that the effectiveness of the proposed framework performs as intended.

We study the problem of time series anomaly prediction, which is relevant to a range of real-world applications. Existing anomaly prediction methods rely on labeled training data for achieving acceptable accuracy. However, such data may be difficult to obtain; and in real-time deployments, anomalies can occur that were not seen in labeled data, thus making them difficult to predict. We provide a theoretical analysis and propose an Importance-based Generative Contrastive Learning method (IGCL) for unsupervised anomaly prediction. IGCL employs a controlled diffusion module to produce anomaly precursor patterns. Next, ICGL learns contextual representations to extract temporal dependencies from pairs of normal time series and anomaly precursors. IGCL is then able to predict anomalies by identifying anomaly precursors that will evolve into future anomalies. To address challenges caused by potentially complex precursor combinations involving multiple variables, we propose a memory bank with importance scores that stores representative samples adaptively and generates more complex anomaly precursors. Extensive experiments on nine benchmark datasets offer evidence that the proposed method is able to outperform state-of-the-art baselines.

Large Language Models (LLMs) have emerged as a promising paradigm for time series analytics, leveraging their massive parameters and the shared sequential nature of textual and time series data.However, a cross-modality gap exists between time series and textual data, as LLMs are pre-trained on textual corpora and are not inherently optimized for time series.In this tutorial, we provide an up-to-date overview of LLM-based cross-modal time series analytics.We introduce a taxonomy that classifies existing approaches into three groups based on cross-modal modeling strategies, e.g., conversion, alignment, and fusion, and then discuss their applications across a range of downstream tasks.In addition, we summarize several open challenges.This tutorial aims to expand the practical application of LLMs in solving real-world problems in cross-modal time series analytics while balancing effectiveness and efficiency.Participants will gain a thorough understanding of current advancements, methodologies, and future research directions in cross-modal time series analytics.

The widespread deployment of wireless and mobile devices results in a proliferation of decentralized spatio-temporal data. Many recent proposals that target deep learning for spatio-temporal prediction assume that all data is available at a central location and suffers from so-called catastrophic forgetting, where previously learned knowledge is entirely forgotten when new data arrives. Such proposals may face data privacy concerns and may experience deteriorating prediction performance when applied in decentralized settings where data streams into the system. To bridge the gap between decentralized training and spatio-temporal prediction on streaming data, we propose a unified federated continuous learning framework, which uses a horizontal federated learning mechanism for protecting data privacy and includes a global replay buffer with synthetic spatio-temporal data generated by the previously learned global model. For each client, we fuse the current training data with synthetic spatio-temporal data using a spatio-temporal mixup mechanism to preserve historical knowledge effectively, thus avoiding catastrophic forgetting. To enable holistic representation preservation, the local models at clients each integrates a general spatio-temporal autoencoder with a spatio-temporal simple siamese network that aims to ensure prediction accuracy and avoid holistic feature loss. Extensive experiments on real data offer insight into the effectiveness of the proposed framework.

The widespread deployment of wireless and mobile devices results in a proliferation of spatio-temporal data that is used in applications, e.g., traffic prediction, human mobility mining, and air quality prediction, where spatio-temporal prediction is often essential to enable safety, predictability, or reliability. Many recent proposals that target deep learning for spatio-temporal prediction suffer from so-called catastrophic forgetting, where previously learned knowledge is entirely forgotten when new data arrives. Such proposals may experience deteriorating prediction performance when applied in settings where data streams into the system. To enable spatio-temporal prediction on streaming data, we propose a unified replay- based continuous learning framework. The framework includes a replay buffer of previously learned samples that are fused with training data using a spatio-temporal mixup mechanism in order to preserve historical knowledge effectively, thus avoiding catastrophic forgetting. To enable holistic representation preservation, the framework also integrates a general spatio-temporal autoencoder with a carefully designed spatio-temporal simple siamese (STSimSiam) network that aims to ensure prediction accuracy and avoid holistic feature loss by means of mutual information maximization. The framework further encompasses five spatio-temporal data augmentation methods to enhance the performance of STSimSiam. Extensive experiments on real data offer insight into the effectiveness of the proposed framework.

With the proliferation of GPS-equipped edge devices, huge trajectory data is generated and accumulated in various domains, motivating a variety of urban applications. Due to the limited acquisition capabilities of edge devices, a lot of trajectories are recorded at a low sampling rate, which may lead to the effectiveness drop of urban applications. We aim to recover a high-sampled trajectory based on the low-sampled trajectory in free space, i.e., without road network information, to enhance the usability of trajectory data and support urban applications more effectively. Recent proposals targeting trajectory recovery often assume that trajectories are available at a central location, which fail to handle the decentralized trajectories and hurt privacy. To bridge the gap between decentralized training and trajectory recovery, we propose a lightweight framework, LightTR, for federated trajectory recovery based on a client-server architecture, while keeping the data decentralized and private in each client/platform center (e.g., each data center of a company). Specifically, considering the limited processing capabilities of edge devices, LightTR encompasses a light local trajectory embedding module that offers improved computational efficiency without compromising its feature extraction capabilities. LightTR also features a meta-knowledge enhanced local-global training scheme to reduce communication costs between the server and clients and thus further offer efficiency improvement. Extensive experiments demonstrate the effectiveness and efficiency of the proposed framework.

With the proliferation of mobile sensing techniques, huge amounts of time series data are generated and accumulated in various domains, fueling plenty of real-world applications. In this setting, time series anomaly detection is practically important. It endeavors to identify deviant samples from the normal sample distribution in time series. Existing approaches generally assume that all the time series is available at a central location. However, we are witnessing the decentralized collection of time series due to the deployment of various edge devices. To bridge the gap between the decentralized time series data and the centralized anomaly detection algorithms, we propose a Parameter-efficient Federated Anomaly Detection framework named PeFAD with the increasing privacy concerns. PeFAD for the first time employs the pre-trained language model (PLM) as the body of the client's local model, which can benefit from its cross-modality knowledge transfer capability. To reduce the communication overhead and local model adaptation cost, we propose a parameter-efficient federated training module such that clients only need to fine-tune small-scale parameters and transmit them to the server for update. PeFAD utilizes a novel anomaly-driven mask selection strategy to mitigate the impact of neglected anomalies during training. A knowledge distillation operation on a synthetic privacy-preserving dataset that is shared by all the clients is also proposed to address the data heterogeneity issue across clients. We conduct extensive evaluations on four real datasets, where PeFAD outperforms existing state-of-the-art baselines by up to 28.74%.

The expanding instrumentation of processes throughout society with sensors yields a proliferation of time series data that may in turn enable important applications, e.g., related to transportation infrastructures or power grids. Machine-learning based methods are increasingly being used to extract value from such data. We provide means of reducing the resulting considerable computational and data storage costs. We achieve this by providing means of condensing large time series datasets such that models trained on the condensed data achieve performance comparable to those trained on the original, large data. Specifically, we propose a time series dataset condensation framework, TimeDC, that employs two-fold modal matching, encompassing frequency matching and training trajectory matching. Thus, TimeDC performs time series feature extraction and decomposition-driven frequency matching to preserve complex temporal dependencies in the reduced time series. Further, TimeDC employs curriculum training trajectory matching to ensure effective and generalized time series dataset condensation. To avoid memory overflow and to reduce the cost of dataset condensation, the framework includes an expert buffer storing pre-computed expert trajectories. Extensive experiments on real data offer insight into the effectiveness and efficiency of the proposed solutions.

Spatial Crowdsourcing (SC) is finding widespread application in today's online world. As we have transitioned from desktop crowdsourcing applications (e.g., Wikipedia) to SC applications (e.g., Uber), there is a sense that SC systems must not only provide effective task assignment but also need to ensure privacy. To achieve these often-conflicting objectives, we propose a framework, Task Assignment with Federated Preference Learning, that performs task assignment based on worker preferences while keeping the data decentralized and private in each platform center (e.g., each delivery center of an SC company). The framework includes a federated preference learning phase and a task assignment phase. Specifically, in the first phase, we build a local preference model for each platform center based on historical data. We provide means of horizontal federated learning that makes it possible to collaboratively train these local preference models under the orchestration of a central server. Specifically, we provide a practical method that accelerates federated preference learning based on stochastic controlled averaging and achieves low communication costs while considering data heterogeneity among clients. The task assignment phase aims to achieve effective and efficient task assignment by considering workers’ preferences. Extensive evaluations on real data offer insight into the effectiveness and efficiency of the paper's proposals.

With the proliferation of wireless and mobile devices, Spatial Crowdsourcing (SC) attracts increasing attention, where task assignment plays a critically important role. However, recent task assignment solutions in SC often assume that data is stored in a central station while ignoring the issue of privacy leakage. To enable decentralized training and privacy protection, we propose a federated task assignment framework with personalized location-preference learning, which performs efficient task assignment while keeping the data decentralized and private in each platform center (e.g., a delivery center of an SC company). The framework consists of two phases: personalized federated location-preference learning and task assignment. Specifically, in the first phase, we design a personalized location-preference learning model for each platform center by simultaneously considering the location information and data heterogeneity across platform centers. Based on workers' location preference, the task assignment phase aims to achieve effective and efficient task assignment by means of the Kuhn-Munkres (KM) algorithm and the newly proposed conditional degree-reduction algorithm. Extensive experiments on real-world data show the effectiveness of the proposed framework.

Fine-grained urban flow inference, which aims to infer the fine-grained urban flows of a city given the coarse-grained urban flow observations, is critically important to various smart city related applications such as urban planning and public safety. Previous works assume that the urban flow monitoring sensors are evenly distributed in space for data collection and thus the observed urban flows are complete. However, in real-world scenarios, sensors are usually unevenly deployed in space. For example, the traffic cameras are mostly deployed at the crossroads and central areas of a city, but less likely to be deployed in suburb. The data scarcity issue poses great challenges to existing methods for accurately inferring the fine-grained urban flows, because they require all urban flow observations to be available. In this paper, we make the first attempt to infer fine-grained urban flows based on the incomplete coarse-grained urban flow observations, and propose a Multi-Task urban flow Completion and Super-Resolution network (MT-CSR for short) to simultaneously complete the coarse-grained urban flows and infer the fine-grained flows. Specifically, MT-CSR consists of the data completion network (CMPNet for short) and data super-resolution network (SRNet for short). CmpNet is composed of a local spatial information based data completion module LocCmp and an auxiliary information based data completion module AuxCmp to consider both the local geographical and global semantic correlations for urban flow data completion. SRNet is designed to capture the complex associations between fine-and coarse-grained urban flows and upsample the coarse-grained data by stacking the designed super-resolution blocks. To gain an accurate inference, two parts are jointly conducted under a multi-task learning framework, and trained in an end-to-end manner using a two-stage training strategy. Extensive experiments on four large real-world datasets validate the effectiveness and efficiency of our method compared with the state-of-the-art baselines.

Crowd flow prediction, which aims to predict the in/out flows of different areas of a city, plays an important role in various applications like intelligent transportation. The challenges of this problem lie in both dynamic mobility patterns of crowds and complex spatial-temporal correlations. Meanwhile, crowd flow is highly correlated to and affected by the Origin-Destination (OD) locations of the flow trajectories, which is largely ignored by existing works. In this paper, we study the novel problem of predicting the crowd flow and flow OD simultaneously, and propose a multi-task bayes-enhanced adversarial spatial temporal network entitled MBA-STNet. MBA-STNet adopts a shared-private framework that contains private spatial-temporal encoders, a shared spatial-temporal encoder, and decoders to learn the task-specific features and shared features. To effectively extract discriminative shared features, an adversarial loss on shared feature extraction is incorporated to reduce information redundancy. A Bayesian Heterogeneous Spatio-temporal Attention Network is designed to learn complex spatio-temporal correlations and alleviate data uncertainty. We also design an attentive temporal queue to capture the complex temporal dependency automatically without domain knowledge. Extensive evaluations are conducted over the bike and taxicab trip datasets in New York. The results demonstrate that the proposed MBA-STNet is superior to state-of-the-art methods.

Abnormal traffic incidents such as traffic accidents have become a significant health and development threat with the rapid urbanization of many countries. Thus it is critically important to accurately forecast the traffic accident risks of different areas in a city, which has attracted increasing research interest in the research area of urban computing. The challenges of accurate traffic risk forecasting are three-fold. First, traffic accident data in some areas of a city is sparse, especially for a fine-grained prediction, which may cause the zero inflation problem during model training. Second, the spatio-temporal correlations of the traffic accidents occurring in different areas are rather complex and non-linear, which is difficult to capture by existing shallow models like regression. Third, the occurrence of traffic accidents can be significantly affected by various context features including weather, POI and road network features. It is non-trivial to capture the complex associations between the diverse context features and traffic accident risks for building an accurate prediction model. To address the above challenges, this paper proposes a Multi-View Multi-Task Spatio-Temporal Networks (MVMT-STN) model to forecast fine- and coarse-grained traffic accident risks of a city simultaneously. Specifically, to address the data sparsity issue in a fine-grained prediction, we adopt a multi-task learning framework to jointly forecast both fine- and coarse-grained traffic accident risks by considering their spatial associations. For each granularity prediction, we design the channel-wise CNN and multi-view GCN to capture the local geographic dependency and global semantic dependency, respectively. In order to obtain the diverse impacts of the context features on traffic accidents, we also introduce a fusion learning module that integrates the channel-wise and multi-view features learned from different types of the external factors. We conduct extensive experiments over two large real traffic accident datasets. The results show that MVMT-STN improves the performance of traffic accident risk prediction in both fine- and coarse-grained prediction by a large margin compared with existing state-of-the-art methods.

Spatial Crowdsourcing (SC) is ubiquitous in the online world today. As we have transitioned from crowdsourcing applications (e.g., Wikipedia) to SC applications (e.g., Uber), there is a substantial precedent that SC systems have a responsibility not only to effective task assignment but also to privacy protection. To address these often-conflicting responsibilities, we propose a framework, Task Assignment with Federated Preference Learning, which performs task assignment based on worker preferences while keeping the data decentralized and private in each platform center (e.g., each delivery center of an SC company). The framework includes two phases, i.e., a federated preference learning and a task assignment phase. Specifically, in the first phase, we design a local preference model for each platform center based on historical data. Meanwhile, the horizontal federated learning with a client-server structure is introduced to collaboratively train these local preference models under the orchestration of a central server. The task assignment phase aims to achieve effective and efficient task assignment by considering workers' preferences. Extensive evaluations over real data show the effectiveness and efficiency of the paper's proposals.

Urban flow imputation, which aims to infer the missing flows of some locations based on the available flows of surrounding areas, is critically important to various smart city related applications such as urban planning and public safety. Although many methods are proposed to impute time series data, they may not be feasible to be directly applied on urban flow data due to the following reasons. First, urban flows have the complex spatial and temporal correlations which are much harder to be captured compared with time series data. Second, the urban flow data can be random missing (i.e., missing randomly in terms of times and locations) or block missing (i.e., missing for all locations in a particular time slot). Thus it is difficult for existing methods to work well on both scenarios. In this paper, we for the first time study the urban flow imputation problem and propose a generative-free Attention-based Spatial-Temporal Combine and Mix Completion Network model (AST-CMCN for short) to effectively address it. Specifically, AST-CMCN consists of a Spatial and Temporal Completion Network (SATCNet for short) and a Spatial-Temporal Mix Completion Network (STMCNet for short). SATCNet is composed of stacked GRUAtt modules to capture the geographical and temporal correlations of the urban flows, separately. STMCNet is designed to capture the complex spatial-temporal associations jointly between historical urban flows and current data. A Message Passing module is also proposed to capture new spatial-temporal patterns that never appear in the historical data. Extensive experiments on two large real-world datasets validate the effectiveness and efficiency of our method compared with the state-of-the-art baselines.

Traffic flow prediction based on vehicle trajectories collected from the installed GPS devices is critically important to Intelligent Transportation Systems (ITS). One limitation of existing traffic prediction models is that they mostly focus on predicting road-segment level traffic conditions, which can be considered as a fine-grained prediction. In many scenarios, however, a coarse-grained prediction, such as predicting the traffic flows among different urban areas covering multiple road links, is also required to help government have a better understanding on traffic conditions from the macroscopic point of view. This is especially useful in the applications of urban planning and public transportation planning. Another limitation is that the correlations among different types of traffic-related features are largely ignored. For example, the traffic flow and traffic speed are usually negatively correlated. Existing works regard these traffic-related features as independent features without considering their correlations. In this article, we for the first time study the novel problem of multivariate correlation-aware multi-scale traffic flow predicting, and we propose a feature correlation-aware spatio-temporal graph convolutional networks named MC-STGCN to effectively address it. Specifically, given a road graph, we first construct a coarse-grained road graph based on both the topology closeness and the traffic flow similarity among the nodes (road links). Then a cross-scale spatial-temporal feature learning and fusion technique is proposed for dealing with both the fine- and coarse-grained traffic data. In the spatial domain, a cross-scale GCN is proposed to learn the multi-scale spatial features jointly and fuse them together. In the temporal domain, a cross-scale temporal network that is composed of a hierarchical attention is designed for effectively capturing intra- and inter-scale temporal correlations. To effectively capture the feature correlations, a feature correlation learning component is also designed. Finally, a structural constraint is introduced to make the predictions on the two scale traffic data consistent. We conduct extensive evaluations over two real traffic datasets, and the results demonstrate the superior performance of the proposal on both fine- and coarse-grained traffic predictions.

Crowd flow prediction, which aims to predict the in-out flows (e.g. the traffic of crowds, taxis and bikes ) of different areas of a city, is critically important to many real applications including public safety and intelligent transportation systems. The challenges of this problem lie in both the dynamic mobility patterns of crowds and the complex spatial-temporal correlations. Meanwhile, crowd flow is highly correlated to and affected by the Origin-Destination (OD) locations of the flow trajectories, which is largely ignored by existing works. In this paper, we study the novel problem of predicting the crowd flow and flow OD simultaneously, and propose a multi-task adversarial spatial-temporal network model entitled MT-ASTN to effectively address it. As a multi-task learning model, MT-ASTN adopts a shared-private framework which contains private spatial-temporal encoders, a shared spatial-temporal encoder, and decoders to learn the task-specific features and shared features. To effectively extract high quality shared features, a discriminative loss on task classification and an adversarial loss on shared feature extraction are incorporated to reduce information redundancy. We also design an attentive temporal queue to automatically capture the complex temporal dependency without the help of domain knowledge. Extensive evaluations are conducted over the bike and taxicab trip datasets in New York. The results demonstrate that our approach significantly outperforms state-of-the-art methods by a large margin on both tasks.