Privacy-Preserving Location Recommendation via Decentralized Collaborative Learning
Summary
The widespread use of mobile Internet infrastructure in urban areas and the proliferation of location-aware mobile devices have led to a significant increase in individual-level location data, which provides new opportunities for urban mobility research and location-based services. However, the gathering and utilization of such data are typically carried out by centralized platforms, which may raise privacy concerns among the public. In this study, we propose a decentralized framework for privacy-preserving location recommendation through secure and private deep learning. Our approach enables users to retain their data on their devices and facilitates collaborative learning among them while preserving privacy through secure multi-party computation and differential privacy. We conducted a case study in the New York City (NYC) area, leveraging users’ location history and city-scale public data such as transportation infrastructure, place safety, and traffic flow data to perform location recommendation. Moreover, we explored two potential attack scenarios to evaluate the proposed framework’s privacy protection effectiveness and robustness. The results demonstrate that our framework achieves a better balance between privacy and utility compared to traditional centralized methods and exhibits robustness against certain attacks. We anticipate that our findings and subsequent discussion will provide insights into privacy-preserving geospatial artificial intelligence and promote geoprivacy in location-based services.
This seminar series is co-organized by CHUD (Center for Housing & Urban Development), GeoSAT (Center for Geospatial Sciences, Applications and Technology), and TAMIDS-DAL (Design and Analytics Lab for Urban Artificial Intelligence @ Texas A&M Institute of Data Science).
Speaker’s information
Dr. Jinmeng Rao is a Research Scientist at Mineral Earth Sciences (a Google[X] graduate project). He got his PhD in Geography at UW-Madison supervised by Prof. Song Gao. He holds a Master’s degree in Computer Sciences from UW-Madison, a Master’s degree in Cartography and Geographic Information Systems from Wuhan University supervised by Prof. Qingyun Du, and a Bachelor’s degree in Geographic Information Systems from Wuhan University. His main research interest is Geospatial Artificial Intelligence (GeoAI) and its application in geoprivacy protection. He published over 20 academic articles in peer-reviewed journals and conferences, including Transactions in GIS (TGIS), Computers, Environment and Urban Systems (CEUS), GIScience, etc. He serves as a reviewer for over 10 academic journals, including Scientific Reports, TGIS, Computers & Geosciences, etc. He is the recipient of various national/university-level scholarships or grants. He is also a volunteer for several international programs and conferences, including AAG, ERASMUS, UN Mappers, CPGIS, Wisconsin Land Information Association (WLIA), etc. He co-founded the GISphere project that collects timely and comprehensive information on global graduate programs and opportunities in GIScience and related fields.
Time: 10:00-10:30 a.m. US Central Time (Thursday, April 27th, 2023)
Zoom Meeting ID: 937 3503 2850 Passcode: 808588
Direct Link: https://tamu.zoom.us/j/93735032850?pwd=ZHU0Sm9LV2hkUHBjTUI1aUV1eFN0QT09
Host: Jiaxin Du, Data Science Ambassador@TAMIDS, PhD candidate@LAUP, TAMU