Learned Spatial Keyword Indexes

The Hong Kong University of Science and Technology
Department of Computer Science and Engineering

Final Year Thesis Oral Defense

Title: "Learned Spatial Keyword Indexes"

by

WANG Aaron Si-yuan

Abstract:

Index structures are one of the most fundamental building blocks of 
database systems, as they allow for fast and efficient query processing. 
These index structures typically use traditional algorithmic approaches to 
handle these queries and have been well studied. Learned indexes have 
become a recent and active research topic within the database research 
community. In the past year, there have been numerous studies about their 
potential to improve the performance of traditional index structures for 
different types of data.

Currently, there have yet to be any published results about the potential 
of applying the learned index concept to spatial keyword indexes. As such, 
this project aims to investigate the potential of learned spatial keyword 
indexes, approaching the problem from two different angles: 1) extending 
existing learned spatial indexes to handle spatial keyword queries, 2) 
extending existing spatial keyword indexes to incorporate learned aspects. 
Two learned spatial keyword indexes have been proposed and implemented 
according to these two different approaches. The performance of these 
learned spatial keyword indexes have been evaluated against their 
traditional counterparts. Finally, the issues and challenges of these 
learned spatial keyword indexes have been discussed and explored, 
detailing rooms for improvements and areas for future work.


Date            : 6 May 2021 (Thursday)

Time            : 14:00-14:40

Zoom Link:
https://hkust.zoom.us/j/91308385105?pwd=MXcveTVXWlY0Qk5BZmFDRVpXejNJZz09

Meeting ID      : 913 0838 5105
Passcode        : 651403

Advisor         : Prof. WONG Raymond Chi-Wing

2nd Reader      : Prof. CHAN Gary Shueng-Han