GeoAI 2017 workshop report: the 1st ACM SIGSPATIAL International Workshop on GeoAI: @AI and Deep Learning for Geographic Knowledge Discovery

Abstract

Deep Learning and Artificial Intelligence (AI) techniques are transforming a range of sectors from computer vision and natural language processing to autonomous driving and healthcare. In particular, deep learning methods achieve great success in many computer vision problems, such as image classification and object detection. Deep neural networks are very powerful to capture the hierarchical representation of features in massive and complex data by adopting multiple layers of non-linear information processing. Due to the availability of vast and high-resolution geospatial data and efficient high-performance computing architectures, deep learning techniques empower the geospatial system to provide fast and near-human level perception. For example, recent studies have shown deep learning techniques coupled with volunteered geographic information (such as OpenStreetMap data) can accurately extract buildings from satellite imagery for humanitarian mapping in rural African areas. Also, deep learning helps assimilate autonomous vehicles and intelligent transport system by incorporating a great amount of information gathered by traffic cameras and sensors. Moreover, deep learning technology facilitates the discovery of geographic information within unstructured text data across different languages. There are also many other applications of deep learning in the domain of GIS, such as the prediction for spatial diffusion patterns in epidemiology, urban expansion prediction, and hyperspectral image analysis. The 1st GeoAI workshop aims to bring geoscientists, computer scientists, engineers, entrepreneurs, and decision makers from academia, industry, and government to discuss the latest trends, successes, challenges, and opportunities in the field of deep learning for geographical data mining and knowledge discovery.