主题: One Design Does Not Fit All — Adaptive Optimization Techniques in Data Management

报告人: Dr. Siqiang Luo, Harvard University

时间: 15:30 - 16:30, Thursday, January 9, 2020

地点: Boardroom, Dao Yuan Building

 

摘要:

Billion-dollar businesses, such as Uber (taxi-hailing service, valued at $82 billion in 2019), DiDi (taxi-sharing service, valued at $69 billion in 2019), and Pokemon GO (locationbased game, $1.8 billion revenue in 2 years since launch), are supported by powerful data systems that can process massive volumes of operations (e.g., queries, updates). Previous studies mostly focus on how such services can be supported by data systems with inflexible designs. Examples include building a static index structure to support data queries. However, an inflexible design can lead to unpredictable performance when the data system is applied in applications of different characteristics. My long term research goal is to design adaptive data management algorithms or systems that can be self-configured to optimize the performance under a wide spectrum of system characteristics.

In this talk, I will introduce my recent projects in designing adaptive algorithms or data systems for various applications, including the location-based services, graph analytics, vertexcentric distributed systems as well as NoSQL key-value stores. 
 

简介:

Dr. Siqiang Luo is now a postdoctoral fellow at Harvard University. Before that, he got his Ph.D. degree in computer science from the University of Hong Kong, and his master’s and bachelor’s degrees in computer science from Fudan University. He has been a visiting researcher at the University of Cambridge and Nanyang Technological University. His research interests include adaptive optimization techniques in data management, distributed computation, and efficient index structures on the graph data. He has published more than 20 papers in major database-related journals/conferences including VLDB Journal, TKDE, SIGMOD, VLDB, ICDE, EDBT, and CIKM. He is a member of SIGMOD reproducibility committee.