In This Article Expand or collapse the "in this article" section Geospatial Artificial Intelligence (GeoAI)

  • Introduction
  • Historical Roots and General Overviews
  • Spatially Explicit AI Models
  • Spatial Representation Learning
  • Spatial Prediction and Interpolation
  • AI in Cartography and Mapping
  • Deep Learning in Earth Observation
  • GeoAI in Social Sensing
  • Geospatial Semantics and Geo-Text Analysis

Geography Geospatial Artificial Intelligence (GeoAI)
by
Song Gao
  • LAST MODIFIED: 24 March 2021
  • DOI: 10.1093/obo/9780199874002-0228

Introduction

Nowadays, artificial intelligence (AI) is bringing tremendous new opportunities and challenges to geospatial research. Its fast development is powered by theoretical advancement, big data, computer hardware (e.g., the graphics processing unit, or GPU), and high-performance computing platforms that support the development, training, and deployment of AI models within a reasonable amount of time. Recent years have witnessed significant advances in geospatial artificial intelligence (GeoAI), which is the integration of geospatial studies and AI, especially machine learning and deep learning methods and the latest AI technologies in both academia and industry. GeoAI can be regarded as a study subject to develop intelligent computer programs to mimic the processes of human perception, spatial reasoning, and discovery about geographical phenomena and dynamics; to advance our knowledge; and to solve problems in human environmental systems and their interactions, with a focus on spatial contexts and roots in geography or geographic information science (GIScience). Thus, it would require the knowledge of AI theory, programming and computation practices as well as geographic domain knowledge to be competent in GeoAI research. There have already been increasingly collaborative GeoAI studies for GIScience, remote sensing, physical environment, and human society. It is a good time to provide a key reference list for educators, students, researchers, and practitioners to keep up with the latest GeoAI research topics. This bibliographical entry will first review the historical roots for AI in geography and GIScience and then list up to ten selective recent works with annotations that briefly describe their importance for each topic of interest in the GeoAI landscape, ranging from fundamental spatial representation learning to spatial predictions and to various advancements in cartography, earth observation, social sensing, and geospatial semantics.

Historical Roots and General Overviews

The intersection of AI and geographic studies is not completely new; its historical roots are described in Smith 1984; Couclelis 1986; Openshaw 1992; Openshaw and Openshaw 1997; and Janowicz, et al. 2020. Before the recent explosion of deep learning studies by LeCun, et al. 2015, major AI developments included theoretical speculations in the 1950s and 1960s (see Buchanan 2005); artificial neural networks (ANN), heuristic search, knowledge-based expert systems, neurocomputing and artificial life (e.g., cellular automata) in the 1980s; genetic programming, fuzzy logics, and development of hybrid intelligent systems in the 1990s; and ontology and web semantics for geographic information retrieval (GIR) in the 2000s. All of these developments have contributed to the research themes of GeoAI. One key question that drives contributions in GeoAI is why spatial is special in AI. One answer might be because geographic location is often the key for linking heterogeneous data sets that have been intensively used for training advanced AI models (more information in Hu, et al. 2019b). In addition, what are the key geographical questions that we can now address better using AI rather than traditional approaches? What are the unsolved problems that can now be solved with AI? Are there any new theories or intelligent approaches to building models and data pipelines in geographic information systems (GIS)? Geographers and computer scientists have made great efforts in contributing to these topics in recent publications, such as in a special issue on artificial intelligence techniques for geographic knowledge discovery in the International Journal of Geographical Information Science (Janowicz, et al. 2020) and in the ACM SIGSPATIAL GeoAI workshops (2017, 2018, 2019), as described in Hu, et al. 2019a, as well as discussions in the American Association of Geographers (AAG) GeoAI and Deep Learning symposiums (2018, 2019, 2020).

  • Buchanan, B. G. “A (Very) Brief History of Artificial Intelligence.” AI Magazine 26.4 (2005): 53–53.

    Profound thoughts in philosophy, fiction, and imagination and early inventions and technology advancements in electronics, engineering, and many other disciplines have influenced AI.

  • Couclelis, H. “Artificial Intelligence in Geography: Conjectures on the Shape of Things to Come.” Professional Geographer 38.1 (1986): 1–11.

    DOI: 10.1111/j.0033-0124.1986.00001.x

    Couclelis broadens the discussion of AI in geography after Smith 1984 from more theoretical dimensions of the computational approach, and introduces the discrete-structure hierarchy as a universal framework for multilevel analytical representation.

  • Hu, Y., S. Gao, D. Lunga, W. Li, S. Newsam, and B. Bhaduri. “GeoAI at ACM SIGSPATIAL: Progress, Challenges, and Future Directions.” SIGSPATIAL Special 11.2 (2019a): 5–15.

    DOI: 10.1145/3377000.3377002

    Reviews the research articles published in the 2017, 2018, and 2019 SIGSPATIAL GeoAI workshops, and summarizes a wide range of topics, such as geospatial image processing, transportation modeling, public health, and digital humanities. A list of GeoAI research directions is also suggested.

  • Hu, Y., W. Li, D. Wright, et al. Artificial Intelligence Approaches. In The Geographic Information Science and Technology Body of Knowledge (3rd Quarter 2019 Edition). Edited by John P. Wilson. Ithaca, NY: University Consortium for Geographic Information Science, 2019b.

    Authors review recent developments of AI in geospatial studies, with a focus on machine learning and deep learning approaches, and introduce a variety of applications, such as automatic recognition of natural terrain features from remote sensing images, land cover classification, and spatiotemporal habitats modeling.

  • Janowicz, K., S. Gao, G. McKenzie, Y. Hu, and B. Bhaduri. “GeoAI: Spatially Explicit Artificial Intelligence Techniques for Geographic Knowledge Discovery and Beyond.” International Journal of Geographical Information Science 34.4 (2020): 625–636.

    DOI: 10.1080/13658816.2019.1684500

    Reviews state-of-the-art research on GeoAI for geographic knowledge discovery, explains how a change in data culture is fueling the rapid growth of GeoAI, and points to future research directions. Also calls for the development of spatially explicit models and the sharing of high-quality geospatial data sets for advancing reproducible GeoAI research.

  • LeCun, Y., Y. Bengio, and G. Hinton. “Deep Learning.” Nature 521.7553 (2015): 436–444.

    DOI: 10.1038/nature14539

    A high-level review from worldwide lead researchers on deep convolutional neural networks and recurrent neural networks that have brought about breakthroughs in processing images, video, speech, and audio.

  • Openshaw, S. “Some Suggestions concerning the Development of Artificial Intelligence Tools for Spatial Modelling and Analysis in GIS.” Annals of Regional Science 26.1 (1992): 35–51.

    DOI: 10.1007/BF01581479

    Key landmarks in the development process of AI tools for spatial modeling and analysis in GIS may include (1) demonstrations of new technology with pedagogic data support, (2) empirical evidences that new AI methods outperform traditional approaches, (3) new methods work in data rich environment in GIS, and (4) incorporation of AI procedures in general analysis and modeling.

  • Openshaw, S., and C. Openshaw. Artificial Intelligence in Geography. Chichester, UK: John Wiley & Sons, 1997.

    A milestone book project in GeoAI that introduces key principles of AI, with applications in geography, urban planning, and GIS. It mainly covers the contemporary AI methods and technologies in the 1970s–1990s, including heuristic search, expert systems, intelligent knowledge-based systems, neurocomputing, ANN, artificial life, genetic algorithms, and fuzzy systems.

  • Smith, T. R. “Artificial Intelligence and Its Applicability to Geographical Problem Solving.” Professional Geographer 36.2 (1984): 147–158.

    DOI: 10.1111/j.0033-0124.1984.00147.x

    Smith summarizes the applicability of AI to geographical problem solving, research, and practices, with a focus on individual and aggregated intelligent spatial decision-making from both cognitive and engineering perspectives.

back to top

Users without a subscription are not able to see the full content on this page. Please subscribe or login.

How to Subscribe

Oxford Bibliographies Online is available by subscription and perpetual access to institutions. For more information or to contact an Oxford Sales Representative click here.

Article

Up

Down