Training-free video object editing aims to achieve precise object-level manipulation, including object insertion, swapping, and deletion. However, it faces significant challenges in maintaining fidelity and temporal consistency. Existing methods, often designed for U-Net architectures, suffer from two primary limitations: inaccurate inversion due to first-order solvers, and contextual conflicts caused by crude "hard" feature replacement. These issues are more challenging in Diffusion Transformers (DiTs), where the unsuitability of prior layer-selection heuristics makes effective guidance challenging. To address these limitations, we introduce ContextFlow, a novel training-free framework for DiT-based video object editing. In detail, we first employ a high-order Rectified Flow solver to establish a robust editing foundation. The core of our framework is Adaptive Context Enrichment (for specifying what to edit), a mechanism that addresses contextual conflicts. Instead of replacing features, it enriches the self-attention context by concatenating Key-Value pairs from parallel reconstruction and editing paths, empowering the model to dynamically fuse information. Additionally, to determine where to apply this enrichment (for specifying where to edit), we propose a systematic, data-driven analysis to identify task-specific vital layers. Based on a novel Guidance Responsiveness Metric, our method pinpoints the most influential DiT blocks for different tasks (e.g., insertion, swapping), enabling targeted and highly effective guidance. Extensive experiments show that ContextFlow significantly outperforms existing training-free methods and even surpasses several state-of-the-art training-based approaches, delivering temporally coherent, high-fidelity results.

This article provides a novel and practical approach for investigating the characteristics of intercity telecommunication network whose overall and complete information is unavailable. Using a mobile phone call data set covering 4.39 million subscribers registered in a particular region, we construct two intercity mobile communication subnets and infer characteristics of the whole intercity mobile communication network of China. Results confirm that intercity communication intensity is characterized by the gravity model. The communication intensity based on mobile call number decreases along the distance with a scaling exponent 0.5, whereas the scaling exponent for the communication intensity based on mobile call duration is 0.4. Moreover, we uncover the rank-size distribution of tie strength (mobile call number and duration) between a city and its neighbours. The rank-size law of tie strengths between cities is mainly determined by the rank-size distribution of cities. The distance between cities plays a less decisive role than the size distribution in the network, but significantly impacts mobile communication patterns. The call duration of individual intercity mobile communication is generally positively correlated to the communication distance, explaining why the distance decay of communication intensity based on call durations is slower than that based on call numbers. The contribution of this research is twofold. First, we identify the distance decay effect in intercity mobile communications of China and uncover the dominant impact of the rank-size distribution of cities. Second, a method for estimating the properties of the whole network according to the observed interactions of its subnets is developed.

Traffic flow prediction is a basic function of Intelligent Transportation System. Due to the complexity of traffic phenomenon, most existing methods build complex models such as neural networks for traffic flow prediction. As a model may lose effect with time lapse, it is important to update the model on line. However, the high computational cost of maintaining a complex model puts great challenge for model updating. The high computation cost lies in two aspects: computation of complex model coefficients and huge amount training data for it. In this paper, we propose to use a nonparametric approach based on locally weighted learning to predict traffic flow. Our approach incrementally incorporates new data to the model and is computationally efficient, which makes it suitable for online model updating and predicting. In addition, we adopt wavelet analysis to extract the periodic characteristic of the traffic data, which is then used for the input of the prediction model instead of the raw traffic flow data. The primary experiments on real data demonstrate the effectiveness and efficiency of our approach.

With the rapid advance of location-based services (LBS) and online map services, it is now more feasible than before to collect the geo-referenced individual level data. However, privacy is always an issue whenever personal location is traced and recorded. This paper proposes a reactive location-based service to collect and process individual location data by different privacy policies. The reactive LBS provides user with an active pull mode to collect his/her location information. With different privacy policies, the user can enter his/her current address manually to the server or automatically generated by the LBS location provider. In order to gain the accurate reconstruction of individuals' activity-travel patterns with considerable space-time details, the LBS server invokes an online map service to geo-reference the location data and to derive the route between locations. In this paper, a household's daily activity survey scenario is showed in Beijing city.

GIS Web service turns out to be one of the new directions of the paradigm of GIS, due to the popular use of the Internet and the dramatic progress of telecommunications technology. However, focused on the description of the interface of functions by Web service description language (WSDL), the general Web service technologies are not sufficient to model GIS Web service with data-oriented characteristic, especially those data provider services and data transform services. In this paper, we propose geographic-Web service description language (abbr. G-WSDL) which enhances the ability of describing such data-driven services by extending the current WSDL with service metadata. Two application scenarios will be presented at last to show its strength in automatic GIS Web service discovery and dynamic services combination.

The location technologies, such as GPS and telegraphy, are producing more and more data of moving objects. Spatio-temporal database is needed to manage these data, so as to solve the problems in spatio-temporal applications. However, there are few prototypes of spatio-temporal database systems yet. This is because attention is always paid on some specified aspects not the whole system. In this paper, we present the design and implementation of a spatio-temporal database prototype for managing moving objects. A Trajectory Model is proposed in this paper, where the moving objects' movements in 2D plane over time are treated as trajectories in the 3D spatio-temporal space. Referring to R tree and the trajectory oriented indexing method TB tree, a new indexing method: S-TB tree is proposed in the prototype. Distance based queries and trajectory topology based queries are all supported by the prototype. The distance based queries includes point queries, range queries. This paper also presents different strategies to process these queries. Finally, two application cases are presented: traffic control and historical queries of the storms on Earth.

GIS is a good spatial analysis tool, nevertheless, as the accumulation of spatial data, the functions that GIS offers are not enough. Spatial data mining, which can automatically discover implicit knowledge from spatial data, has recently received wide attention. However, the spatial data mining systems are not competent for preprocessing and presenting spatial data. Hence, the idea to integrate spatial data mining and GIS is straightforward. In this paper, we propose a novel method to integrate spatial data mining and GIS. This method is implemented with SPMML, an XML-based language that is extended from PMML. We build a prototype system with this method and it works well.

With the evolution of GIS from stand-alone systems with geo-data tightly coupled with systems to an increasingly distributed model based on independently-provided, interoperable GIS Web service, much more research has been focused on GIS Web service composition. However, little research works concern on control mechanisms of improving availability and reliability in GIS Web service composition. Considering that GIS Web services are in essence loosely-coupled and hosted by different providers. As a result, any update of any service, might affect critically the overall composition consistency and execution; Moreover because most of geographic operations are CPU-intensive, it means that GIS Web service compositions consisting of basic geographic operations are more CPU-intensive - it will perhaps take several days to finish the execution of one single GIS Web service composition! Without support of high availability and reliability, any mistake will lead to the failure of the whole composition execution and the waste of much computing resource. So availability and reliability in GIS Web service composition should be paid more attention. This paper proposes a mechanism of service compensation to achieve high availability and reliability in GIS Web service composition. The basic idea is that when one Web service of composition fails, the execution of the whole composition is not aborted immediately; on the contrary, the compensation mechanism tries to find another Web service which can provide the same function to "compensate" it. The abortion of composition only happens when the "compensating" Web services can not be found or all of them fail.

Outlier detecting is one of the most important data analysis technologies in data mining, which can be used to discover anomalous phenomena in huge dataset. Many literatures on spatial outlier detecting and time series outlier detecting have appeared, while the area of spatio-temporal outliers considering both spatial and temporal dimensions has still rarely been touched. Defining outliers in traditional dataset is more explicit because the data structure we need to focus on is very straightforward (e.g., a spatial point or a transaction record). However, it is much more difficult to give outlier a definite characterization in spatio-temporal lattice data, since there are so many data structures we can pay attention to. With the aim of detecting useful and meaningful outliers in climate dataset, we introduce a formalized way to define outliers in spatio-temporal lattice data, in which the importance of clarifying basic data structure (we call it basic element in our paper) is stressed. As a case study, we define two kinds of spatio-temporal outliers based on a global climate dataset, according to the three aspects we propose in defining an outlier. The introduction of basic element and the formulation of outlier definition process make it easier and clearer to define meaningful outliers. Thus outlier detecting in spatio-temporal lattice data will provide us with really interesting and useful knowledge.

Spatial sampling is a significant part of query processing in spatial database. In this paper, an adaptive data-driven sampling for selectivity estimation is proposed in spatial database. This technique presents an efficient sampling for spatial data, especially for two-dimensional line and polygon data, and it make the sample size fit the limit of time and memory, or a user-defined parameter. The data-driven sampling technique is compared with various techniques on different type of datasets in our experimental study, and it out outperforms the other techniques over a broad range of query workloads and datasets.

Spatial data mining is a promising technique that deals with extraction of implicit knowledge or other interesting patterns from large amount of spatial data. Though most data mining systems work with data stored in flat files or operational database, it has been recognized that mining in a data warehouse usually result in better information. Because data are usually cleansed before they are stored into data warehouse. Furthermore, data warehouse provides data with different levels of summarization for the clients, which will lead to fruitful data mining. However, current techniques of data warehouse can not handle spatial data well. Both dimensions and measures in the data model of data warehouse are nonspatial data. In this paper, we propose a new data model called spatial data cube for data warehouse. Spatial data cube supports both spatial and nonspatial data. We also introduce how to construct a spatial data cube that can answer queries efficiently by selective materialization. We believe that the spatial data cube can provide better support for spatial data mining.