Handbook of Geospatial Artificial Intelligence

6Spatial Representation Learning in GeoAI

Gengchen Mai

Spatially Explicit Artificial Intelligence Lab, Department of Geography, University of Georgia

Ziyuan Li

School of Business, University of Connecticut

Ni Lao

Google

DOI: 10.1201/9781003308423-6

CONTENTS

• 6.1 Introduction

• 6.2 Spatially Explicit Artificial Intelligence

• 6.3 Spatial Representation Learning on Various Spatial Data Types

  1. 6.3.1 Points and Location Encoders

  2. 6.3.2 Polylines and Polyline Encoder

  3. 6.3.3 Polygons and Polygon Encoder

• 6.4 Conclusions

  1. Bibliography

6.1 INTRODUCTION

The choice of data representation and features is usually a decisive factor in the performance of machine learning (ML) models (Bengio et al., 2013). Traditionally, ML features are manually picked and ...