HRNet: Differentially Private Hierarchical and Multi-Resolution Network for Human Mobility Data Synthesization

TL;DR

This paper presents HRNet, a deep generative model that synthesizes realistic human mobility data while ensuring differential privacy, addressing key challenges with innovative hierarchical, multi-resolution, and private pre-training components.

Contribution

Introduction of HRNet, a novel deep generative model combining hierarchical encoding, multi-resolution learning, and private pre-training for privacy-preserving human mobility data synthesis.

Findings

HRNet outperforms existing methods in utility-privacy trade-off.

The model effectively captures complex mobility patterns.

Extensive experiments validate its superior performance.

Abstract

Human mobility data offers valuable insights for many applications such as urban planning and pandemic response, but its use also raises privacy concerns. In this paper, we introduce the Hierarchical and Multi-Resolution Network (HRNet), a novel deep generative model specifically designed to synthesize realistic human mobility data while guaranteeing differential privacy. We first identify the key difficulties inherent in learning human mobility data under differential privacy. In response to these challenges, HRNet integrates three components: a hierarchical location encoding mechanism, multi-task learning across multiple resolutions, and private pre-training. These elements collectively enhance the model's ability under the constraints of differential privacy. Through extensive comparative experiments utilizing a real-world dataset, HRNet demonstrates a marked improvement over existing methods in balancing the utility-privacy trade-off.