Physical Trajectory Inference Attack and Defense in Decentralized POI Recommendation

Jing Long, Tong Chen, GUANHUA YE, Kai Zheng, Quoc Viet Hung Nguyen, Hongzhi Yin

Keywords: Point-of-Interest Recommendation; Decentralized collaborative Learning; Trajectory Inference Attack and Defense

TL;DR: We propose a novel attack namely Physical Trajectory Inference Attack (PTIA) to reveal users’ interacted POIs in decentralized POI recommendations, followed by an effective defender.

Abstract: As an indispensable personalized service within Location-Based Social Networks (LBSNs), the Point-of-Interest (POI) recommendation aims to assist individuals in discovering attractive and engaging places. However, the accurate recommendation capability relies on the powerful server collecting a vast amount of users' historical check-in data, posing significant risks of privacy breaches. Although several collaborative learning (CL) frameworks for POI recommendation enhance recommendation resilience and allow users to keep personal data on-device, they still share personal knowledge to improve recommendation performance, thus leaving vulnerabilities for potential attackers. Given this, we design a new Physical Trajectory Inference Attack (PITA) to expose users' historical trajectories. Specifically, for each user, we identify the set of interacted POIs by analyzing the aggregated information from the target POIs and their correlated POIs. We evaluate the effectiveness of PITA on two real-world datasets across two types of decentralized CL frameworks for POI recommendation. Empirical results demonstrate that PITA poses a significant threat to users' historical trajectories. Furthermore, Local Differential Privacy (LDP), the traditional privacy-preserving method for CL frameworks, has also been proven ineffective against PITA. In light of this, we propose a novel defense mechanism (AGD) against PITA based on an adversarial game to eliminate sensitive POIs and their information in correlated POIs. After conducting intensive experiments, AGD has been proven precise and practical, with minimal impact on recommendation performance.

Track: User Modeling and Recommendation

Submission Guidelines Scope: Yes

Submission Guidelines Blind: Yes

Submission Guidelines Format: Yes

Submission Guidelines Limit: Yes

Submission Guidelines Authorship: Yes

Student Author: Yes

Submission Number: 539