Toward Decentralized Geospatial Applications via Secure and Private Deep Learning

Keywords: Geoprivacy, Privacy-Preserving AI, GeoAI, Decentralization, Federated Learning
Abstract Type: Paper Abstract

Abstract

The nowadays ubiquitous location-aware mobile devices have contributed to the rapid growth of individual-level location data. Such data are usually collected by centralized location-based service platforms as training data to improve their predictive models' performance, but the collection of such data may raise public concerns about privacy issues. In this work, we introduce a privacy-preserving decentralized framework based on secure and private deep learning. Compared with traditional centralized learning frameworks, we keep users' data on their own devices and train the model locally so that their data remain private. The local model parameters are aggregated and updated through secure multiple-party computation and differential privacy to achieve collaborative learning among users while preserving privacy. We also discuss potential attack cases that can be used to examine the privacy protection effectiveness and robustness of the framework. The results of case studies show that our framework achieves a better balance on the privacy–utility trade-off compared with traditional centralized learning methods. The results and ensuing discussion offer new insights into privacy-preserving geospatial artificial intelligence and promote geoprivacy in location-based services.