Semantic similarity measure of natural language text through machine learning and a keyword-aware cross-encoder-ranking summarizer—A case study using UCGIS GIS&T body of knowledge

Initiated by the University Consortium of Geographic Information Science (UCGIS), the GIS&T Body of Knowledge (BoK) is a community-driven endeavor to define, develop, and document geospatial topics related to geographic information science and technologies (GIS&T). In recent years, GIS&T BoK has undergone rigorous development in terms of its topic re-organization and content updating, resulting in a new digital version of the project. While the BoK topics provide useful materials for researchers and students to learn about GIS, the semantic relationships among the topics, such as semantic similarity, should also be identified so that a better and automated topic navigation can be achieved. Currently, the related topics are either defined manually by editors or authors, which may result in an incomplete assessment of topic relationships. To address this challenge, our research evaluates the effectiveness of multiple natural language processing (NLP) techniques in extracting semantics from text, including both deep neural networks and traditional machine learning approaches. Besides, a novel text summarization—KACERS (Keyword-Aware Cross-Encoder-Ranking Summarizer)—is proposed to generate a semantic summary of scientific publications. By identifying the semantic linkages among key topics, this work guides the future development and content organization of the GIS&T BoK project. It also offers a new perspective on the use of machine learning techniques for analyzing scientific publications and demonstrates the potential of the KACERS summarizer in semantic understanding of long text documents.     

Toward an Immersive 3D Virtual BoK Exploratorium: A Proof of Concept

The UCGIS GIS&T Body of Knowledge document provided an opportunity for the GIS educational community to link course content and curricular sequencing to a catalog of subject matter. Focusing on learning objectives, a selection of relevant citations, and basic background for each topic, it has been used to a limited degree for both course topic selection and curriculum development. However, the static format of the document, lack of an index, and dated nature of the material limit its utility for education. Based on the success of our research team in developing a virtual platform for a new, more interactive, and collaborative environment to catalogue and interactively add to the body knowledge (Ahearn et al. 2013), this article describes efforts to develop a multi‐user virtual user environment that will add social presence to the experience. It describes the successes and failures of using Second Life as the initial platform for this work, illustrates the available interactions and limitations, and depicts ongoing efforts to move beyond Second Life for this development. Finally it discusses a possible methodology to leverage the power of virtual crowd sourcing within competitive gaming environments such as Unity to allow for the creation of on‐demand virtual 3‐D visualizations of GIS&T concepts in a digital Exploratorium.     

Assessment of Workforce Demands to Shape GIS&T Education

The Body of Knowledge for Geographic Information Science and Technology (GIS&T BoK) has been the main reference document for curriculum design in the geospatial domain. Today, the BoK is supposed to have fallen short in adequately covering the domain due to significant conceptual and technological advances in the field. Thus, several initiatives around the globe work towards an update of the GIS&T BoK. This research assesses the demands of today's GIS&T workforce across Europe to contribute to the effort of a demand‐oriented update. We assessed the workforce demand by means of a Europe‐wide distributed online questionnaire and complementary expert interviews. The results show that the BoK still is a comprehensive reference base for the geospatial domain that is generally deemed relevant by the European workforce. However, workforce demands point to three main topics that need to be addressed by an update of the BoK: (1) the shift from primary data acquisition to the handling of highly abundant spatial data; (2) a lack of competences in programming and application development; and (3) a poor coverage of web‐related aspects. Future research should complement workforce demands with a review of the scientific literature to identify additional shortcomings related to conceptual advances.     

Whither goes the “maps” course? Maintaining map-use concepts,skills, and appreciation in GIS&T curricula

Popular and professional interest in maps and map-making has never been greater. Despite this engagement, introductory skills courses in “map use” have largely disappeared in undergraduate geography and cartography programs at US universities, with research indicating that both map-use course enrollments and the existence of such courses themselves are on the decline. While the map-use course and its content may have been replaced by, repackaged as, or incorporated into an integrated geographical information systems-cartography–remote sensing foundations course, how do we ensure that an understanding of maps and their unique characteristics is not lost in the mix? This essay describes a recent course redesign experience incorporating information literacy, technology fluency, and spatial thinking principles to maintain map-use concepts, skills, and appreciation in an introductory geographic information science and technology (GIS&T) foundations course, meeting both GIS&T and general education curricular goals.     

Exploring the anatomy of Geographic Information Systems and Technology (GIS&T) textbooks

Choosing a textbook is among the most important decisions instructors make in preparation for an academic term. Geographic Information Systems and Technology (GIS&T) textbook development has been influenced by a unique set of circumstances, mainly the rapid development of the discipline within an interdisciplinary environment, which has resulted in a continuous state of evolution. We examine the anatomy of GIS&T textbooks through a comparison of their organization, content, and depth of coverage. Specifically, utilizing the Geographic Information Science and Technology Body of Knowledge (BoK) as a comprehensive reference, we categorize the content of 26 of the most widely used introductory GIS textbooks. Our results show that there has been consistent coverage of topics over time, with analytical methods and geospatial data being the most prominent topics covered in texts. However, individual textbooks place varying emphasis on the BoK knowledge areas, which is potentially useful to instructors seeking books that emphasize particular knowledge areas. Additionally, long‐term trends indicate a shift toward an emphasis on new forms of geospatial data (e.g., social media). Ongoing efforts to expand and revise the BoK reflect how the discipline continues to manage its own evolution as new geographic research linked to GIS and GIScience emerges.     

在特殊疫情背景下《城市地理资讯学》MOOC课程教学实施效果分析

在新冠肺炎疫情期间,线上教学大规模开展,相关线上教学问题逐渐成为研究热点。传统的测绘地理类课程的教学方式主要是上课时以老师传授知识为主,课后以学生完成作业为辅的方式进行。但是,这种教学方式比较单一,并且在科学化考核和促进学生学习主动性等方面存在诸多问题。然而,线上网络课程是通过互联网以电子设备为教学平台,具有不受教学时空限制、教学方法灵活多样和丰富学生的学习方式等优点。因此,对《城市地理资讯学》MOOC(massive online open courses)课程实施效果进行分析,为其他测绘地理类网络课程设计提供优化思路,同时为无法进行面授教学的特殊时期提供一种有效的教学方案。     

翻转课堂在风景园林专业“工程测量”课程中的教学设计探析

风景园林专业"工程测量"课程中运用翻转课堂是我国高校传统教学模式的重要变革。翻转课堂的本质是教师与学生角色的转变,教师转变为专业知识的领学者和答疑者,而学生则是专业知识的学习主体。结合风景园林专业中"工程测量"课程,应用翻转课堂的基本理论与内在实施逻辑,对风景园林专业中的"工程测量"课程进行翻转教学设计,并构建了以学生为中心的互动式教学评价体系,以提高学生学习的主观能动性和创新性,从而带动高校传统教学模式的变革。     

误差理论与测量平差基础课程教学改革

文章针对测绘工程专业基础核心课程"误差理论与测量平差基础"的教学内容、教学手段和方法以及考试评价体系进行改革和探索。发挥"测量平差"国家精品课程及国家精品资源共享课程的优质教学资源的优势,丰富教学内容,融合程序设计,拓展训练平台,创新考核模式。重点提高课堂和实践教学效果,试行四段式考核和评价系统,并以课外导学、项目引导、交互沟通为补充,探索开放式的学习养成系统。     

Web‐based GIS Approaches to Enhance Public Participation in Wind Farm Planning

Planning information pertaining to the potential visual impacts of proposed construction developments is particularly important in the case of wind farm planning, given the high levels of concern amongst members of the public regarding the perceived negative visual impacts of wind turbines on the landscape. Previous research has highlighted the shortcomings associated with traditional visualization techniques used to assess these impacts, and also the means by which such information is then disseminated to the wider public during the consultation stages of the wind farm planning process. This research is concerned with examining the potential of Web‐based mapping and digital landscape visualization techniques for addressing some of these shortcomings. This article reports the findings of a Web‐based survey study designed to evaluate the potential of online GIS‐based approaches for improving the effectiveness and dissemination of wind farm visualizations and enhancing public participation in the wind farm planning process. Results from the survey study add to the research literature by demonstrating how innovative Web‐based approaches have real potential for augmenting existing methods of information provision and public participation in the planning process. The findings of this study are also potentially transferrable to other landscape planning scenarios.     

Traffic transformer: Capturing the continuity and periodicity of time series for traffic forecasting

Traffic forecasting is a challenging problem due to the complexity of jointly modeling spatio‐temporal dependencies at different scales. Recently, several hybrid deep learning models have been developed to capture such dependencies. These approaches typically utilize convolutional neural networks or graph neural networks (GNNs) to model spatial dependency and leverage recurrent neural networks (RNNs) to learn temporal dependency. However, RNNs are only able to capture sequential information in the time series, while being incapable of modeling their periodicity (e.g., weekly patterns). Moreover, RNNs are difficult to parallelize, making training and prediction less efficient. In this work we propose a novel deep learning architecture called Traffic Transformer to capture the continuity and periodicity of time series and to model spatial dependency. Our work takes inspiration from Google’s Transformer framework for machine translation. We conduct extensive experiments on two real‐world traffic data sets, and the results demonstrate that our model outperforms baseline models by a substantial margin.     

VGDI – Advancing the Concept: Volunteered Geo‐Dynamic Information and its Benefits for Population Dynamics Modeling

The concept of Volunteered Geographic Information (VGI) has progressed from being an exotic prospect to making a profound impact on GIScience and geography in general, as initially anticipated. However, while massive and manifold data is continuously produced voluntarily and applications are built for information and knowledge extraction, the initially introduced concept of VGI lacks certain methodological perspectives in this regard which have not been fully elaborated. In this article we highlight and discuss an important gap in this concept, i.e. the lack of formal acknowledgment of temporal aspects. By coining the proposed advanced framework ‘Volunteered Geo‐Dynamic Information’ (VGDI), we attempt to lay the ground for full conceptual and applied spatio‐temporal integration. To illustrate that integrative approach of VGDI and its benefits, we describe the potential impact on the field of dynamic population distribution modeling. While traditional approaches in that domain rely on survey‐based data and statistics as well as static geographic information, the use of VGDI enables a dynamic setup. Foursquare venue and user check‐in data are presented for a test site in Lisbon, Portugal. Two core modules of spatio‐temporal population assessment are thereby addressed, namely time use profiling and target zone characterization, motivated by the potential integration in existing population dynamics frameworks such as the DynaPop model.     

Robust and fast reconstruction of complex roofs with active sampling from 3D point clouds

This article proposes a novel method for the 3D reconstruction of LoD2 buildings from LiDAR data. We propose an active sampling strategy which applies a cascade of filters focusing on promising samples at an early stage, thus avoiding the pitfalls of RANSAC‐based approaches. Filters are based on prior knowledge represented by (nonparametric) density distributions. In our approach samples are pairs of surflets—3D points together with normal vectors derived from a plane approximation of their neighborhood. Surflet pairs provide parameters for model candidates such as azimuth, inclination and ridge height, as well as parameters estimating internal precision and consistency. This provides a ranking of roof model candidates and leads to a small number of promising hypotheses. Building footprints are derived in a preprocessing step using machine learning methods, in particular support vector machines.     

Comparison of sampling strategies for object-based classification of urban vegetation from Very High Resolution satellite images

Vegetation monitoring is becoming a major issue in the urban environment due to the services they procure and necessitates an accurate and up to date mapping. Very High Resolution satellite images enable a detailed mapping of the urban tree and herbaceous vegetation. Several supervised classifications with statistical learning techniques have provided good results for the detection of urban vegetation but necessitate a large amount of training data. In this context, this study proposes to investigate the performances of different sampling strategies in order to reduce the number of examples needed. Two windows based active learning algorithms from state-of-art are compared to a classical stratified random sampling and a third combining active learning and stratified strategies is proposed. The efficiency of these strategies is evaluated on two medium size French cities, Strasbourg and Rennes, associated to different datasets. Results demonstrate that classical stratified random sampling can in some cases be just as effective as active learning methods and that it should be used more frequently to evaluate new active learning methods. Moreover, the active learning strategies proposed in this work enables to reduce the computational runtime by selecting multiple windows at each iteration without increasing the number of windows needed.     

软件考试驱动的GIS专业教学改革与实践——以“空间数据库”课程为例

GIS专业的学生需要一定的软件设计和编程技能,传统单一的课堂教学模式难以适应当前培养GIS人才的要求。文中提出通过引导学生参加软件考试,结合考试类型和考试科目进行考试的技术指导和培训,并将其引入到"空间数据库"的课堂教学,使得课程教学与软件考试紧密结合;通过引入任务驱动、情景教学、分级和分组的教学方法,以及课程考试和软件考试的双重考核方法,促进"空间数据库"课程教学改革。课程改革的实施不仅活跃了课堂教学氛围,改善课堂秩序,而且促进学生课后的学习自主性和社会实践积极性,为培养高质量的GIS人才打下基础。     

基于MOOC翻转课堂的“GPS原理与应用”课程教学改革实践

为提升"GPS原理与应用"课程教学效果和教学质量,解决"GPS原理与应用"课程教学中存在的问题,文中提出采用MOOC与翻转课堂融合的教学方式探索课程教学改革。通过构建课前设计、课前学习、课堂教学和课后答疑等四阶段的MOOC与翻转课堂融合的教学模式,探讨该教学模式在课程教学中的实施过程,对比不同班级的教学效果,最终得出MOOC与翻转课堂融合的教学模式能够有效提升课程教学效果。     

Spatial optimization for land acquisition problems: A review of models,solution methods,and GIS support

This article reviews the interdisciplinary research field of spatial optimization for land acquisition problems. We start with a theoretical framework to identify three categories of spatial optimization models: problems with aspatial constraints, location models, and problems with topological constraints. Exact, heuristic, and metaheuristic approaches to solving these problems are critically discussed. Tools that are available in commercial and open‐source GIS packages are reviewed from four aspects. We first survey the off‐the‐shelf support and then the development environments in these packages. A case study of the one‐center problem is used to illustrate the computational performance of different solution methods. Finally the advantages and disadvantages of current GIS data models are discussed. The article concludes with challenges and future directions for solving spatial optimization problems for land acquisition.     

Room semantics inference using random forest and relational graph convolutional networks: A case study of research building

Semantically rich maps are the foundation of indoor location‐based services. Many map providers such as OpenStreetMap and automatic mapping solutions focus on the representation and detection of geometric information (e.g., shape of room) and a few semantics (e.g., stairs and furniture) but neglect room usage. To mitigate the issue, this work proposes a general room tagging method for public buildings, which can benefit both existing map providers and automatic mapping solutions by inferring the missing room usage based on indoor geometric maps. Two kinds of statistical learning‐based room tagging methods are adopted: traditional machine learning (e.g., random forests) and deep learning, specifically relational graph convolutional networks (R‐GCNs), based on the geometric properties (e.g., area), topological relationships (e.g., adjacency and inclusion), and spatial distribution characteristics of rooms. In the machine learning‐based approach, a bidirectional beam search strategy is proposed to deal with the issue that the tag of a room depends on the tag of its neighbors in an undirected room sequence. In the R‐GCN‐based approach, useful properties of neighboring nodes (rooms) in the graph are automatically gathered to classify the nodes. Research buildings are taken as examples to evaluate the proposed approaches based on 130 floor plans with 3,330 rooms by using fivefold cross‐validation. The experiments conducted show that the random forest‐based approach achieves a higher tagging accuracy (0.85) than R‐GCN (0.79).     

测绘工程全英语教学模式的建立与实施策略

测绘工程全英语教学是我国测绘高等教育教学改革与实践面临的挑战。以武汉大学测绘学院为例,探讨全英语教学的内涵,分析测绘工程全英语教学的现状与存在的问题,以测绘工程专业卓越班人才培养方案为建设对象,提出测绘工程全英语教学模式的建立与实施策略,培养测绘工程专业具有国际视野与竞争力的高质量人才。     

The lay of the language: surveying the cartographic characteristics of language maps

Although visible in research in the 1980s and 1990s, works concerning language-mapping issues are recently rather absent. This is an unfortunate oversight given current GIS capability and its potential to tackle visualization issues that were previously simply acknowledged and accepted. Given that there are no established guidelines for language map construction, this work aims to renew attention to language mapping, beginning with a survey documenting the characteristics of published language maps. The survey components address the problematic aspects described in the literature, such as boundary representation and depicting linguistic diversity, and reveal their usage and frequency. The noted map characteristics include, but are not limited to: publication type, publication year, coverage area, language data or variable used, and symbology details. For consistent classification, we use a language-map symbology classification scheme found in previous research. In general, chorochromatic maps using polygonal map units dominate our survey. We also find further evidence supporting the problems outlined in language mapping literature with the widespread use of solid-line boundaries and depiction of only one language or feature per place. However, we also note some unique strategies used for handling uncertainty and linguistic plurality. Observations of tactics not captured by the existing 20-year-old typology lead us to create an updated language map symbology typology consistent with the trends observed in our survey. Overall, we document language mapping strategies in practice and provide direction for future research by highlighting the pros and cons of current cartographic approaches for depicting language.