The seminar named “AIxDB: On the Design and
Implementation of an AI-powered DBMS” was organized by COMP Dept., PolyU on 30th
Jul 2024. This seminar aims to explore the evolution of data systems with a
focus on deepening the fusion of AI and DB; and outline the conceptual and
architectural overview of NeurDB. Prof. Beng Chin OOI (Lee Kong Chian
Centennial Professor, COMP Dept., National University of Singapore) was the
guest speaker. He briefed the seminar outline in the beginning.
Firstly, Prof. OOI introduced the background of
database management system (DBMS) which is the cornerstone of most digital
applications and proposed to add AI system for analytics.
Then he briefed the AI background for different
applications such as healthcare and e-commerce.
Traditionally, DBMS supported online application
including Online Transactional Processing (OLTP) and Online Analytical
Processing (OLAP) but AI-powered analytics cannot be handled by DBMS alone that
needs AI system.
And then Prof. OOI mentioned that integrating AI and
DB is not new but expressed about forty years ago. Our department head Prof.
Qing Li’s PhD thesis is on this field too!
The most updated paper on the fusion of AI and DB
(AIxDB) in 2016 was introduced.
After that he briefed the development history of DBMSs
from 1960 to 2010s.
The development history of AI was also stated that “designing
and developing next-generation DBMSs is both timely and realizable” he said.
NeurDB is an AI-powered DBMS that was built and its
fundamental infrastructure was briefed.
Its overall conceptual for AI-powered application for
healthcare, e-commerce and finance were showed.
The in-database model selection is explained and they
proposed to use two-phase model selection algorithm. Its design principles
included “Filter-and-Refine (FRP)” and “Continue to exploit FRP”.
Finally, Prof. OOI showed us their development roadmap
of NeurDB.
NeurDB-PoC supports in-database AI-powered analytics
so as to enhance SQL syntax.
Q&A session
Reference
沒有留言:
發佈留言