Seminar on Recent Developments in Databases and Information Systems

This semester, the focus of the seminar is on Databases for AI and AI for Databases. To attend this seminar, students need to register according to the registration procedure explained below. After the end of the registration process, participants are selected and informed. Each selected student will get assigned one research paper that needs to be presented in a talk at the end of the semester. Finally, also a short report about the presented paper needs to be submitted, the latest two weeks after the presentation.


  • Participants should have successfully attended the core lecture Datenbanksysteme (database systems) or equivalent.
  • Having attended the beginner’s course Informationssysteme (information systems), or equivalent, is assumed anyway.

Dates and Deadlines:

  • 08.10.2023 Registration deadline for the Seminar
  • 25.10.2023 10:00am Kickoff Meeting in 36-336 (Slides, Summary Template)
  • 20.11.23 First Meeting with Tutor & Summary
  • 14.01.24 No-Left-TODO slides
  • 16.02.24 9:00am: Presentations
  • 03.03.24: Final report deadline (find the template here)

Information on Registration Process:

  • The number of participants is limited.
  • Registration is not done on a first-come, first-served basis.
  • In order to register, download this JSON template registration file rename it to yourmatriculationnumber.json, edit it to reflect your information, and send it as an attachment via email to Patrick Hansert ( Make sure the file is valid JSON and ASCII or UTF-8 encoded, the latter without byte order mark. Use the paper ids from the following table to specify your preference for at least four papers. Please use your official university email account or or or @rptu…. to register and to send the email.
  • Registration due 08.10.2023
  • Soon after the end of the registration, we will let you know whether or not you got a slot in the seminar.


Look further than only titel and abstract before deciding for a paper, since some papers might demand math or deeper ML skills you might find difficult. You have to be in the University’s VPN (full tunnel) for accessing some of the papers.

# Paper Title Weblink
1 Towards Optimizing Storage Costs on the Cloud
2 Analyzing and Mitigating Data Stalls in DNN Training
3 DeepDB: Learn from Data, not from Queries!
4 NeuroCard: One Cardinality Estimator for All Tables
5 The Case for Learned In-Memory Joins
6 Automatic Database Management System Tuning Through Large-scale Machine Learning
7 Cross Modal Data Discovery over Structured and Unstructured Data Lakes
8 DB-BERT: A Database Tuning Tool that “Reads the Manual”
9 Can Learned Models Replace Hash Functions?
10 The case for learned index structures


  • 2023-10-25 Deadlines and Tentative Schedule Announced
  • 2023-10-09 Papers Assigned, Kickoff Meeting Announced
  • 2023-09-04 List of papers added and instructions updated
  • 2023-08-17 Page created