DBGroup @ SUSTech

The Database Research Group is founded in 2017 and affiliated to Department of Computer Science and Engineering, Southern University of Science and Technology. We conduct in-depth, interesting and insightful researches on data science and engineering, which cover the following aspects:

(1) System: Architect novel data-intensive systems from 0 to 1 (lead by Dr. Bo Tang); (2) Algorithm: Support and accelerate advanced data analytics (lead by Dr. Xiao Yan); (3) VIS: Enable intuitive visual understanding of data and systems (lead by Dr. Qiaomu Shen); (4) AI: Improve data analytics and system performance with deep learning (lead by Dr. Dan Zeng).

System

Algorithm

VIS

AI

For more details about our researches, please refer to our publications. We always welcome brilliant people to join DBGroup.

News

2022-10

Congratulations to Jiarui, who won 2022 National Scholarship!

2022-09

Congratulations to Dr. Bo Tang, who won 2022 Shenzhen Excellent Teaching Award!

2022-09

Our system work GHive has been accepted by ACM SoCC 2022. Congratulations to GHive team, which is leading by Haotian. GHive will open source soon, stay tuned!

2022-08

Our work on "Weighted Terrain" got accepted to SIGMOD 2023. Congratulations to Bo Huang and Bo Tang, special thanks to Victor (PolyU) and Raymond (HKUST)!

2022-08

Our work on "Learning-based Cardinality Estimation" got accepted to SIGMOD 2023. Congratulations to Fang, Xiao and Bo, special thanks to Ken (PolyU)!

See more news

Seminars

Date

Title

Speaker

2022-05-28

Tailoring Abstractions for Emerging Machine Learning Workloads

谢志强,斯坦福大学博士生

Performing emerging machine learning computation on hardware accelerators efficiently is challenging. Existing frameworks often adopt several standard intermediate representation layers to bridge the gap between model description and hardware execution units, including high level interfaces, computational graphs and kernel implementation. While this kind of design simplifies the system stack implementation, it can incur fundamental performance limitations. In this talk, I will present Rammer (OSDI ‘20) and Graphiler (GNNSys ‘21), two compiler stacks for ordinary Deep Neural Networks (DNNs) and Graph Neural Networks (GNNs) respectively.

2022-05-19

边缘计算场景下的高效时空分位数监测

易岚菁, 南科大本科生

分位数计算 (Quantile Computation) 是了解数据集分布的基本手段,也是诸多数据查询处理、分析挖掘以及机器学习任务的基础模块。当下,越来越多的物理信息 (如温湿度、车速、噪声等) 被传感器所采集,形成了高速高密度的时空大数据。目前,集中式的分位数计算方法难以实现对高速高密度物联网数据的高响应能力 (responsiveness),而少数的分布式的、针对无线传感网络的方案则未考虑弹性多变的多查询用户场景。本工作提出了一个适应物联网环境的、基于边缘计算 (edge computing) 和数据摘要技术(data sketching) 的处理框架,以高效完成时空分位数监测的查询应答。

2022-04-14

Resource Management and Templates in C++

李润中, 南科大-香港理工大学联合培养博士生

在使用C++时,程序员往往需要对内存进行手动管理,这正是C++有别于其他具有高级特性的编程语言之处。这为C++带来了灵活的资源控制粒度和较低的资源消耗,但同时为程序设计引入了不小的麻烦。本次报告主要通过对C++现代资源管理设施的介绍,展示现代C++的资源管理理念,以及消除不安全资源管理的手段。本次报告同时简要地介绍如何利用C++模板元编程的高级新特性来减少重复代码的编写。

2022-04-07

Plotting in Python

张家澍, 香港理工大学博士生

在学术论文中,我们常常需要绘制大量的点图,线图和柱图来呈现实验结果,好的作图方法允许我们在更好地呈现实验结果的同时提高工作的效率。我们有许多方式绘制论文中所需要的点图,线图和柱图,如通过Office软件,MATLAB和Origin。在本次学术论坛中,我们介绍一种跨平台的作图方法,使用Python进行作图。我们结合一些例子,向大家展示如何使用Python进行作图,如何微调图,以及如何将图插入到论文中。
See more seminars

Publications

𝜏-LevelIndex: Towards Efficient Query Processing in Continuous Preference Space

ACM Conference on Management of Data (SIGMOD, CCF-A), 2022

Authors: Jiahao Zhang, Bo Tang, Man Lung Yiu, Xiao Yan, Keming Li.

abstract pdf

Efficient and Error-bounded Spatiotemporal Quantile Monitoring in Edge Computing Environments

Proceedings of the VLDB Endowment (PVLDB, CCF-A), 2022

Authors: Huan Li, Lanjing Yi, Bo Tang, Hua Lu, Christian Jensen.

abstract pdf

Face2Exp: Combating Data Biases for Facial Expression Recognition

IEEE Conference on Computer Vision and Pattern Recognition (CVPR, CCF-A), 2022

Authors: Dan Zeng, Zhiyuan Lin, Xiao Yan, Yuting liu, Fei Wang and Bo Tang.

abstract pdf
See more publications

Members

Dr. Bo Tang

Dr. Xiao Yan

Dr. Qiaomu Shen

Dr. Dan Zeng

Fang Wang

Ph.D. Candidate

Class of 2019

Shiqi Zhang

Ph.D. Candidate

Class of 2019

Jiaping Cao

Ph.D. Candidate

Class of 2020

Haotian Liu

Ph.D. Candidate

Class of 2020

Wentao Ning

Ph.D. Candidate

Class of 2020

Zheng Bian

Ph.D. Candidate

Class of 2021

Runzhong Li

Ph.D. Candidate

Class of 2021

Jiashuo Lin

Ph.D. Candidate

Class of 2021

Long Xiang

Ph.D. Candidate

Class of 2021

Zhengxin You

Ph.D. Candidate

Class of 2022

Keming Li

MPhil Candidate

Class of 2020

Qian Li

MPhil Candidate

Class of 2020

Zelin Li

MPhil Candidate

Class of 2021

Chaozu Zhang

MPhil Candidate

Class of 2021

Xiangyu Zhi

MPhil Candidate

Class of 2021

Yangshen Deng

MPhil Candidate

Class of 2022

Qilong Li

MPhil Candidate

Class of 2022

Shangxuan Wu

MPhil Candidate

Class of 2022

See all members

Activities

课题组成员曾剑顺利通过博士学位论文答辩

CCF信息系统专委走进高校系列——走进南方科技大学

2021年表彰大会 照片墙

2020粤港澳大湾区数据库研讨会

DBGroup协助组织南科大附中-计算机系AI夏令营

唐教授与2016级毕业生《我与南科大畅享会》

Collaborators

Contact

We always welcome brilliant people to join our group. If you want to join DBGroup, please fill this form and drop an email to us as soon as possible.

Email

dbgroup_AT_sustech_DOT_edu_DOT_cn

Address

DBGroup, South Tower, CoE Building

Southern University of Science and Technology

Shenzhen, China