Incremental Update Research Articles

An Online Active Broad Learning Approach for Real-Time Safety Assessment of Dynamic Systems in Nonstationary Environments.

Real-time safety assessment of the complex dynamic systems in nonstationary environments is of great significance for avoiding the potential hazards. In this case, the update procedure with high assessment accuracy and training speed is crucial and meaningful in the dynamic streaming setting. Generally, the performance of most online learning approaches will be negatively affected by limited annotated samples in such a setting. Moreover, the time cost of advanced conventional methods with retaining procedures is relatively high, constraining their practicality. In this article, a novel online active broad learning approach, termed OABL, is proposed. In detail, the effectiveness of the broad learning system in the framework of online active learning is first revealed and verified. A reasonable dynamic asymmetric query strategy is then designed with a limited annotation budget to actively annotate the relatively valuable samples, which is beneficial to mitigating the negative effects of class imbalance. In this context, the advantage of the human-in-the-loop characteristic is also effectively used to control the evolution direction of the learner during the incremental update, which makes it better able to adapt to complex and nonstationary environments. Several related experiments are conducted with the realistic data of JiaoLong deep-sea manned submersible. Results show the effectiveness and practicality of the proposal compared with the existing advanced approaches.

Fast Radio Map Construction with Domain Disentangled Learning for Wireless Localization

The accuracy of wireless fingerprint-based indoor localization largely depends on the precision and density of radio maps. Although many research efforts have been devoted to incremental updating of radio maps, few consider the laborious initial construction of a new site. In this work, we propose an accurate and generalizable framework for efficient radio map construction, which takes advantage of readily-available fine-grained radio maps and constructs fine-grained radio maps of a new site with a small proportion of measurements in it. Specifically, we regard radio maps as domains and propose a Radio Map construction approach based on Domain Adaptation (RMDA). We first employ the domain disentanglement feature extractor to learn domain-invariant features for aligning the source domains (available radio maps) with the target domain (initial radio map) in the domain-invariant latent space. Furthermore, we propose a dynamic weighting strategy, which learns the relevancy of the source and target domain in the domain adaptation. Then, we extract the domain-specific features based on the site's floorplan and use them to constrain the super-resolution of the domain-invariant features. Experimental results demonstrate that RMDA constructs a fine-grained initial radio map of a target site efficiently with a limited number of measurements. Meanwhile, the localization accuracy of the refined radio map with RMDA significantly improved by about 41.35% after construction and is comparable with the dense surveyed radio map (the reduction is less than 8%).

An incremental photovoltaic power prediction method considering concept drift and privacy protection

Concept drift (CD) and regional data sharing were considered potential factors critical to affecting data-driven model accuracy. In this paper, an incremental photovoltaic (PV) power prediction method is proposed for addressing the CD of streaming data and the privacy protection of regionally shared data. In the proposed scheme, a novel CD detection method is developed using Energy distance as a virtual drift criterion to measure the data distribution change and prediction error change as a real drift criterion to determine the drift time. In order to incorporate richer data features, this paper jointly trains a PV prediction model based on a broad learning system (BLS) through regional data sharing. Further, an incremental BLS model under the distributed federated learning (FL) framework is designed to maintain the privacy of each PV station's data. Finally, an incremental online model update strategy is proposed to learn new features continuously and without forgetting under the FL framework after CD occurs. The test results on the public dataset from DKASC, Alice Springs PV system, present that the proposed method can effectively detect the CD of streaming data and realize the fast online model update. Compared with the state-of-the-art techniques, the proposed method can significantly improve prediction accuracy while preserving privacy.

Distraction descriptor for brainprint authentication modelling using probability-based Incremental Fuzzy-Rough Nearest Neighbour

This paper aims to design distraction descriptor, elicited through the object variation, to refine the granular knowledge incrementally, using the proposed probability-based incremental update strategy in Incremental Fuzzy-Rough Nearest Neighbour (IncFRNN) technique. Most of the brainprint authentication models were tested in well-controlled environments to minimize the influence of ambient disturbance on the EEG signals. These settings significantly contradict the real-world situations. Thus, making use of the distraction is wiser than eliminating it. The proposed probability-based incremental update strategy is benchmarked with the ground truth (actual class) incremental update strategy. Besides, the proposed technique is also benchmarked with First-In-First-Out (FIFO) incremental update strategy in K-Nearest Neighbour (KNN). The experimental results have shown equivalence discriminatory performance in both high distraction and quiet conditions. This has proven that the proposed distraction descriptor is able to utilize the unique EEG response towards ambient distraction to complement person authentication modelling in uncontrolled environment. The proposed probability-based IncFRNN technique has significantly outperformed the KNN technique for both with and without defining the window size threshold. Nevertheless, its performance is slightly worse than the actual class incremental update strategy since the ground truth represents the gold standard. In overall, this study demonstrated a more practical brainprint authentication model with the proposed distraction descriptor and the probability-based incremental update strategy. However, the EEG distraction descriptor may vary due to intersession variability. Future research may focus on the intersession variability to enhance the robustness of the brainprint authentication model.

Incremental association rules update algorithm based on the sort compression matrix

Association rule algorithm has always been a research hotspot in the field of data mining, in the context of today’s big data era, in order to efficiently obtain association rules and effectively update them, based on the original fast update pruning (FUP) algorithm, an association rule incremental update algorithm (FBSCM) based on sorting compression matrix is proposed to solve the shortcomings of frequent scanning of transaction datasets. Firstly, The algorithm maps the transaction dataset as a Boolean matrix, and changes the storage mode of the matrix(that is, adding two columns and a row vector); Secondly, the matrix is compressed many times during the generation of frequent k-itemset; After that, the items in the matrix are sorted incrementally according to the support degree of the itemset; Finally, the original string comparison operation is replaced by the vector product of each column of the matrix. Experimental results and analysis show that the FBSCM algorithm has higher temporal performance than the traditional FUP algorithm in different incremental dataset sizes, different minimum support thresholds and different feature datasets, especially when the incremental transaction volume is large or the minimum support degree is small.

A Novel AUC Maximization Imbalanced Learning Approach for Predicting Composite Outcomes in COVID-19 Hospitalized Patients.

The COVID-19 patient data for composite outcome pre- diction often come with class imbalance issues, i.e., only a small group of patients develop severe composite events after hospital admission while the rest do not. An ideal COVID-19 composite outcomes prediction model should seek strong imbalanced learning capability. Such a model also should have fewer tuning hyperparameters to preserve its good usability and exhibit its potential for fast incremen- tal learning. Towards this goal, by means of the classical proximal support vector machine (PSVM), this study pro- poses a novel imbalanced learning approach called Imbalanced maximizing-Area Under the Curve (AUC) Proximal Support Vector Machine (ImAUC-PSVM) to predict the hospitalized COVID-19 patients' composite outcomes within 30 days of hospitalization. ImAUC-PSVM has the following merits: (1) it incorporates the straightforward AUC maximization into the objective function and hence has fewer parameters to tune. This makes it naturally suitable for handling imbalanced COVID-19 data with a simplified training process. (2) Our theoretical derivations reveal that it has the same form of the analytical solution as PSVM. Therefore, it inherits the advantages of PSVM in dealing with incremental COVID- 19 cases using the fast incremental updating. Our proposed classifier was built and validated internally and externally on the real COVID-19 patient data retrieved from three separate sites of Mayo Clinic in the United States and also validated on the public datasets using various performance metrics. Experi- mental results demonstrate that ImAUC-PSVM outperforms in most cases, showing its potential to assist clinicians in triaging COVID-19 patients at early stage in the hospital settings.

Interpretable Feature Construction and Incremental Update Fine-Tuning Strategy for Prediction of Rate of Penetration

Prediction of the rate of penetration (ROP) is integral to drilling optimization. Many scholars have established intelligent prediction models of the ROP. However, these models face challenges in adapting to different formation properties across well sections or regions, limiting their applicability. In this paper, we explore a novel prediction framework combining feature construction and incremental updating. The framework fine-tunes the model using a pre-trained ROP representation. Our method adopts genetic programming to construct interpretable features, which fuse bit properties with engineering and hydraulic parameters. The model is incrementally updated with constant data streams, enabling it to learn the static and dynamic data. We conduct ablation experiments to analyze the impact of interpretable features’ construction and incremental updating. The results on field drilling datasets demonstrate that the proposed model achieves robustness against forgetting while maintaining high accuracy in ROP prediction. The model effectively extracts information from data streams and constructs interpretable representational features, which influence the current ROP, with a mean absolute percentage error of 7.5% on the new dataset, 40% lower than the static-trained model. This work provides a theoretical reference for the interpretability and transferability of ROP intelligent prediction models.

The WRF-Based Incremental Analysis Updates and Its Implementation in an Hourly Cycling Data Assimilation System

Abstract An incremental analysis update (IAU) scheme is successfully implemented into a WRF/WRFDA-based hourly cycling data assimilation system with the goal to reduce the imbalance introduced by the high-frequency intermittent data assimilation, especially when radar data are included. With the application of IAU, the analysis increment is smoothly introduced into the model integration over a time window centered at the analysis time. As in digital filter initialization (DFI), the IAU scheme is able to limit large shocks in the early part of a model forecast. Compared to DFI, IAU does better in hydrometeor spinup and produces more continuous precipitation forecasts from cycle to cycle. The run with IAU is shown to improve the precipitation forecast skills (10+% for CSI scores) compared to the regular cycling forecasts without IAU. The data assimilation system with IAU is also able to accept more observations due to balanced first-guess fields. Comparable results are obtained in IAU tests when the time-varying weights are used versus constant weights. Because of its better property, the IAU with the time-varying weights is implemented in the operational system.

Tinba: Incremental partitioning for efficient trajectory analytics

Applications with mobile and sensing devices have already become ubiquitous. In most of these applications, trajectory data is continuously growing to huge volumes. Existing systems for big trajectory data organize trajectories at distributed block storage systems. Systems like DITA that use block storage (e.g., 128 MB each) are more efficient for analytical queries, but they cannot incrementally maintain the partitioned data and do not support delete operations, resulting in difficulties in trajectory analytics. In this paper, we propose an incremental trajectory partitioning framework Tinba that enables distributed block storage systems to efficiently maintain optimized partitions under incremental updates of trajectories. We employ a data flushing technique to bulk ingest trajectory data for random writing in distributed file system (DFS). We recast the incremental partitioning problem as an optimal partitioning problem and prove its NP-hardness. A cost–benefit model is proposed to address the optimal partitioning problem. Moreover, Tinba supports most of the existing similarity measures to quantify the similarity between trajectories. A heuristic is developed to instantiate the Tinba framework. Comprehensive experiments on real-world and synthetic datasets demonstrate the advancements in ingestion performance and partition quality, as opposed to other trajectory partition methods.

An online soft sensor based on adaptive double Gaussian Bayesian network

Real-time monitoring of key performance indicators (KPI) through online soft sensors plays a crucial role in modern industrial processes to improve product quality and ensure production safety. In this paper, an Adaptive Double Gaussian Bayesian Network (ADGBN) is proposed to efficiently perform online soft sensor tasks, which can be updated according to the actual working conditions. The ADGBN is based on the concept of model bias correction and consists of an offline prediction model and a calibration model. The calibration model enables incremental learning-based parameter updates using newly generated online data. To improve the operational efficiency of the online soft sensor model, a novel dynamic variable window (DVW) is proposed to monitor model performance and enable adaptive updates. A case study on the polyester fiber polymerization process verifies the effectiveness of the proposed ADGBN online soft sensor and demonstrates its superiority over existing methods.

Flexible Data Refreshing Architecture for Health Information System Integration

Background: Having a consistent and unified view of heterogeneous distributed medical information sources is an inevitable need of health informatics. Integrating medical information of patients or about a disease, a treatment or side effects of a drug, etc, is very useful to help medical education, to achieve medical research goals and to provide the computer- based decision support systems. Contribution: This article proposes a flexible incremental update method for the materialized part of the integration system. It permits us to manage the integration system according to the characteristics of the data sources which can change. Method: This paper presents a hybrid data integration approach in which the materialized part of the system in mediator is the object indexation structure based on an instance classification of the sources objects which correspond to the global schema. The object identifier of each object in the indexation structure is materialized together with the attributes which are needed for the incremental updating of this indexation (classifying attributes). Results: The main idea of this paper is to develop a hybrid data integration framework, which represents a new aspect of a hybrid method focusing on flexible data refreshing. Conclusion: This hybrid approach implements a vertical hybrid approach. It means that at the mediator level, some data of each object are materialized and others are virtual.

EMF-Syncer: scalable maintenance of view models over heterogeneous data-centric software systems at run time

With the increasing presence of cyber-physical systems (CPSs), like autonomous vehicle systems and digital twins, the future of software engineering is predicated on the importance of designing and developing data-centric software systems that can adapt intelligently at run time. CPSs consist of complex heterogeneous software components. Model-driven engineering advocates using software models to tame such complexity, capturing the relevant design concerns of such systems at different levels of abstraction. Yet most of the existing CPSs are engineered without considering MDE practices and tools, facing fundamental challenges when working with data: monitoring the program data at run time, syncing updates between program and model, dealing with heterogeneous data sources, and representing such observed data at run time to facilitate automated analysis. In this work, we introduce the notion of view models to explicitly represent parts of the domain knowledge implicitly embedded in the source code of a software system. This notion is equipped with a scalable bidirectional syncing mechanism that extracts view model instances from program snapshots at run time. The syncing mechanism is proposed from a conceptual point of view, independently of specific implementations and supports incremental view model update and view model maintenance. We show how this syncing mechanism is flexible enough to facilitate the non-intrusive adoption of MDE technology over existing MDE-agnostic heterogeneous data-centric systems. We study the run-time cost implied by the EMF-Syncer , the tool implementing this syncing mechanism for Java applications and view models atop the eclipse modeling framework (EMF) when executing data analytic and transformation tasks over large volumes of data in the presence of data updates at run time. An empirical evaluation of the EMF-Syncer has been conducted with an industry-targeted benchmark for decision support systems, analyzing performance and scalability. The novel syncing mechanism enables new opportunities to adopt MDE technology in heterogeneous data-centric systems.

Extended Online DMD and Weighted Modifications for Streaming Data Analysis

We present novel methods for computing the online dynamic mode decomposition (online DMD) for streaming datasets. We propose a framework that allows incremental updates to the DMD operator as data become available. Due to its ability to work on datasets with lower ranks, the proposed method is more advantageous than existing ones. A noteworthy feature of the method is that it is entirely data-driven and does not require knowledge of any underlying governing equations. Additionally, we present a modified version of our proposed approach that utilizes a weighted alternative to online DMD. The suggested techniques are demonstrated using several numerical examples.

The improvement path of international Chinese education management in universities based on big data technology

Abstract The short video under the wave of the big data era is as crazy as the rain, with all kinds of online hot words abuse, subvert the three views the pompous funny video, using Chinese text stems lead to many college students and teachers Chinese culture level decline, Chinese culture awareness is weak, Chinese education management along the traditional, education management can not keep up with the innovation of the times to study the problem. In this paper, the improvement path of international Chinese education management in colleges and universities based on big data technology is studied, and the incremental update model of data is constructed by calculating the dissimilarity of users and items and getting the dissimilarity matrix. According to the neighborhood rating data, the iu-User CF algorithm provides users with more qualified items. The experimental analysis results show that the algorithm maintains more than 90% accuracy and stability. It shows that the system library of the algorithm is stable and efficient. Education management should innovate development ideas, broaden horizons, adapt to the new era, new situation, and new trends, seize new development opportunities, explore new paths for reform, unite Chinese and foreign cooperation, serve world needs, and promote digitalization, intelligence, and modernization of international Chinese education.

Privacy-Preserving Shortest Distance Query for Dynamic Large-Scale Road Network

The road network plays an important role in guiding people's daily travel. In the road network, the shortest distance query is one of the most basic query operations. With the scale of the road network continuously expanding, people choose to outsource the road network to the cloud server in the form of an encrypted graph. To the best of authors' knowledge, the existing privacy-preserving shortest distance query schemes over large-scale encrypted graphs are all designed for the static graph. Nonetheless, the graph of road network is often updated over time. These static privacy-preserving shortest distance query schemes cannot work well in this practical application. In order to address this problem, we explore how to implement the privacy-preserving shortest distance query in an incremental large-scale road network. We use the 2-hop cover label index as the basic secure index structure, which can realize the shortest distance fast query. In order to update the 2-hop cover label index safely and efficiently, we combine a special adjacency index as an auxiliary index structure with homomorphic encryption to implement a dynamic incremental update algorithm for the encrypted graph. We prove that the proposed scheme achieves adaptive semantic security and verify its practicality through experiments.

Scalable multi‐site photovoltaic power forecasting based on stream computing

AbstractPhotovoltaic (PV) is essential for global carbon neutrality, it is imperative to forecast PV generation accurately for power operation. With the rapid growth of distributed PV sites, a scalable cloud service tends to play a vital role in PV forecasting to address the increasing cost of computing resources and data subscriptions. Such a scheme creates a possibility to further enhance forecasting performance by re‐using forecasting model, data, and computing resources all in the cloud. In order to achieve this goal, this work proposes a multi‐site PV forecasting system design with a message queue (MQ) and stream computing engine, where a hybrid neural network model is trained and continuously updated using real‐time data. A performance benchmark with up to 60 sites served simultaneously was performed to verify the scalability of the stream computing based approach. Moreover, after incremental updating of the forecasting model, a decrease in normalized root mean square error and normalized mean absolute error of PV forecasting were observed, demonstrating that better short‐term forecasting accuracy was achieved.

Global ocean reanalysis CORA2 and its inter comparison with a set of other reanalysis products

We present the China Ocean ReAnalysis version 2 (CORA2) in this paper. We compare CORA2 with its predecessor, CORA1, and with other ocean reanalysis products created between 2004 and 2019 [GLORYS12v1 (Global Ocean reanalysis and Simulation), HYCOM (HYbrid Coordinate Ocean Model), GREP (Global ocean Reanalysis Ensemble Product), SODA3 (Simple Ocean Data Assimilation, version 3), and ECCO4 (Estimating the Circulation and Climate of the Ocean, version 4)], to demonstrate its improvements and reliability. In addition to providing tide and sea ice signals, the accuracy and eddy kinetic energy (EKE) of CORA2 are also improved owing to an enhanced resolution of 9 km and updated data assimilation scheme compared with CORA1. Error analysis shows that the root-mean-square error (RMSE) of CORA2 sea-surface temperature (SST) remains around 0.3°C, which is comparable to that of GREP and smaller than those of the other products studied. The subsurface temperature (salinity) RMSE of CORA2, at 0.87°C (0.15 psu), is comparable to that of SODA3, smaller than that of ECCO4, and larger than those of GLORYS12v1, HYCOM, and GREP. CORA2 and GLORYS12v1 can better represent sub-monthly-scale variations in subsurface temperature and salinity than the other products. Although the correlation coefficient of sea-level anomaly (SLA) in CORA2 does not exceed 0.8 in the whole region, as those of GREP and GLORYS12v1 do, it is more effective than ECCO4 and SODA3 in the Indian Ocean and Pacific Ocean. CORA2 can reproduce the variations in steric sea level and ocean heat content (OHC) on the multiple timescales as the other products. The linear trend of the steric sea level of CORA2 is closer to that of GREP than that of the other products, and the long-term warming trends of global OHC in the high-resolution CORA2 and GLORYS12v1 are greater than those in the low-resolution EN4 and GREP. Although CORA2 shows overall poorer performance in the Atlantic Ocean, it still achieves good results from 2009 onward. We plan to further improve CORA2 by assimilating the best available observation data using the incremental analysis update (IAU) procedure and improving the SLA assimilation method.

RD-FCA: A resilient distributed framework for formal concept analysis

Formal Concept Analysis (FCA) is a data analysis technique with applications in data mining, artificial intelligence, software engineering, etc. The algorithms for FCA are computationally expensive, and their recursion tree is highly irregular and dynamic in nature. Several distributed FCA algorithms have been proposed to exploit parallelism within and across machines. However, none of the distributed approaches are able to recover from failures in the system. We propose RD-FCA, the first resilient distributed framework for FCA that uses novel load-balancing strategy, handles fail-stop failures and provides at-least-once semantics for concept discovery. Our asynchronous snapshot mechanism with incremental updates reduces the snapshot overhead and minimizes recalculation of concepts during recovery. RD-FCA also supports dynamic addition of workers to accelerate performance. Compared to MapReduce based approaches, RD-FCA performs an order of magnitude faster. We show through extensive evaluation that RD-FCA recovers efficiently from single, multiple, independent and cascading failures.

Promenades à travers le monde francophone par James G. Mitchell and Cheryl Tano

Reviewed by: Promenades à travers le monde francophone par James G. Mitchell and Cheryl Tano Jeremy Patterson Mitchell, James G., and Cheryl Tano. Promenades à travers le monde francophone. Vista, 2022. ISBN 978-1-54332-689-5. Pp. 632. As language textbooks become more aligned to proficiency outcomes, it is always instructive to observe incremental updates in new editions. The fourth edition of Promenades, coming four years after the introduction of the third edition in 2018, demonstrates this approach. Whereas the third edition had more significant updates regarding technology, this edition does not present much new technology but does further align itself to communicative goals. It continues to use the vText and Supersite technologies. This edition's updates are its most promising feature—clearer connections to the World-Readiness Standards for Learning Languages. The textbook includes Communicative Goals and Can-Do Statements throughout all units and lessons. When students are writing one of their earliest compositions in French in Unit 2B, for example, they are reminded of the goal ("Write a description about yourself") and the proficiency ("I can write a description of myself including basic information about daily life and preferences"). Another extremely positive addition is the intercultural mode of communication along with the other three modes in the Can-Do Statements. When students are reading a "Panorama" about some aspect of French culture, they are reminded of their ability to "identify French cultural products and practices and relate them to perspectives in my own and other cultures" (79) or, even more specifically, to "identify cultural products and practices of the French overseas departments and reflect on attitudes around them" (285). This is helpful both to students and teachers, as sometimes the latter can forget to build their lessons explicitly around such Can-Do Statements, and especially since the intercultural mode is often taken for granted in language courses, when they should really be understood as language and culture courses—even more so as proficiency-based instruction has led teachers away from a purely linguistic model of pedagogy. Overall, the layout with all of the different sections in each unit can be a little busy from a visual perspective, but still useful and varied. The structure of the textbook remains largely the same. It has the same 13 units as the previous edition, all divided into two lessons. Each lesson follows the same structure of sections: contextes, roman-photo, lecture culturelle, structures, synthèse, and savoir-faire. This structure helpfully includes form and function in the fourth section, structures, rather than starting out with grammar in the first few parts. The last two sections are also adapted from the third edition, where Le Zapping was the last part of each lesson. Le Zapping is now part of the fifth section (synthèse) and includes some new material, like the last section, savoir-faire. Each of these final sections focuses on [End Page 180] a different Francophone country or region of France. The first seven units, or chapters, present different Francophone areas of the world, but students who study the last 6 units (8-13) will be exposed only to the different regions of France (from Paris to Normandy to Corsica to Hauts-de-France). Overall, this edition presents well-organized and pedagogically sound material for students in early semesters of university French acquisition. [End Page 181] Jeremy Patterson Bob Jones University (SC) Copyright © 2023 American Association of Teachers of French

Domain-incremental learning for fire detection in space-air-ground integrated observation network

Deep learning-based fire detection models are usually trained offline on static datasets. For continuously increasing heterogeneous sensor data, incremental learning is a resolution to enable incremental updates of models. However, it still encounters the challenge of the stability-plasticity dilemma on cross-domain data. In this paper, we propose a Dynamic Equilibrium Network (DENet) to achieve cross-domain fire detection on heterogeneous images captured by spaceborne, airborne, and terrestrial sensors. It can learn more flame and smoke features from cross-domain data in a balanced manner by imposing dynamic equilibrium constraints on the self-knowledge distillation model with deeply-supervised loss. It adopts learnable loss weights to adaptively adjust the importance of the old and new datasets and focuses on rebalancing the weight bias based on dynamic equilibrium constraints guided by the number of samples. Model performance is evaluated on the Flame And Smoke Detection Dataset (FASDD) consisting of computer vision (CV) and satellite remote sensing (RS) images, as well as the FLAME dataset consisting of Unmanned Aerial Vehicle (UAV) images. Extensive experiments demonstrated that DENet performs better generalization and equalization on CV, RS, and UAV data than other classical incremental learning methods, and significantly narrows the performance gap toward joint training. It outperforms the well-performing PODNet model by +7.71% mAP@0.5 on the old dataset while maintaining the same good performance on the new dataset. In the ablation experiments, our model outperforms the No Bias model by +27.13% mAP@0.5 on the old dataset, and by +2.77% mAP@0.5 on the new dataset. Given its excellent equalization capability on heterogeneous data, DENet can provide effective support for cross-domain fire detection. This study will provide an important reference for collaborative fire warning, emergency response, and government decision-making in a space-air-ground integrated observation network.