More about HKUST
On Boosting Spatial Data Management with Efficient Indexing Structures
Speaker: Dr. Victor WEI Department of Computing The Hong Kong Polytechnic University Title: "On Boosting Spatial Data Management with Efficient Indexing Structures" Date: Friday, 7 January 2022 Time: 10:00am - 11:00am (Hong Kong Local Time) Zoom link: https://hkust.zoom.us/j/928308079?pwd=b29SMXI1bHNWV1UrdjQ3UWlmUUNSdz09 Meeting ID: 928 308 079 Passcode: 20220107 Abstract: Due to the advance of geo-positioning technologies, spatial data including spatial trajectories, 3D terrain data and spatial networks, etc. becomes more and more popular and draws a lot of attention from both academia and industry. As such, the query processing in spatial data management becomes more and more important and finds wide applications in urban computing, smart cities, autonomous driving, etc. One of the fundamental queries in spatial data management is the shortest distance and shortest path query. In this talk, I will present two research works on this fundamental query which boosts the query processing with efficient indexing structures. The first work tackles an emerging type of spatial data, namely 3D terrain surface. It proposes an indexing structure, namely a Distance Oracle, to efficiently index the pairwise distances among a set of points-of-interest on the terrain surface. The oracle answers the distance query by using a small set of pre-computed distances and utilizes a tree structure to boost the search on the pre-computed distances. Besides, it guarantees the error is bounded by a user-specified error parameter. The second one studies the dynamic road networks where the traffic information changes over time. It proposes an efficient indexing structure for the shortest distance and path queries on dynamic road networks. The indexing structure is based on auxiliary edges introduced to the network, namely shortcuts, which bridge distant vertices on the network to accelerate the query processing. In this work, we propose using a randomized algorithm to generate the shortcuts. As such, we obtain that it has small space consumption and query time and it could be updated efficiently in the perspective of both theory and practice. **************** Biography: Victor Junqiu Wei obtained his PhD degree from the Department of Computer Science and Engineering, the Hong Kong University of Science and Technology in 2018. From March 2018 to May 2018, he was a visiting student of Prof. Hanan Samet and Prof. David Mount at University of Maryland in the US. He obtained his bachelor degree from Nanjing University in 2013. He is currently working as a research assistant professor in the Hong Kong Polytechnic University (jointly Appointed by Department of Computing and Research Institute for Artificial Intelligence). Prior to this, he was an AI Researcher in Huawei Noah's Ark Lab. His research interests span spatial data management, graph data analytics, random sampling and deep neural networks. He received many honors and awards including the nomination of the Best Paper Award of SIGMOD 2020, Potentially High-Value Patent Award of Huawei, Future Star Award of Huawei, etc.