llmdb

Databases and Large Language Models (LLMDB) 2023

1st International Workshop on Databases and Large Language Models at VLDB 2023

The emergent abilities of LLMs represent a turning point that is poised to significantly alter the technological and industrial landscape.

There is an urgent and exciting need to integrate relational databases and LLMs. This workshop’s objective is to draw attention to this emerging topic, which has the potential to improve the performance of LLMs as well as deepen their impact in real-world applications where relational databases are used in data management.

LLMs are trained on large corpora of natural language content and provide responses in natural language, whereas databases manage structured data with a strict schema. So far, they operate in separate spaces. On the one hand, the vast amount of knowledge in relational databases is not used to train and tune LLMs, and on the other hand, relational databases are not able to access and operate on the facts contained in the LLMs.

The workshop will solicit work and facilitate discussions on topics including but not limited to the following:

  1. How do we incorporate structured or semi-structured data into LLMs?
  2. How do we model knowledge in LLMs as a database so that we can query it using SQL?
  3. How do we use relational databases to improve the performance of LLMs?

Workshop Program

We have invited distinguished researchers in this domain to deliver keynote talks. Here are the confirmed keynote speakers.

Dan Roth, UPenn

Wang-Chiew Tan, Meta

Laurel Orr, Numbers Station

Tao Yu, HKU

The LLMDB 2023 workshop will take place on September 1 at 10:30 AM as a half day workshop.

Time Agenda
10:30 - 11:00 Invited talk: The Science of Generative AI: Focus on Reasoning (Dan Roth / UPenn)
11:00 - 11:30 Invited talk: Using LLMs to Query Unstructured and Structured Data (Wang-Chiew Tan / Meta)
11:30 - 11:50 Paper: Proportionate Diversification of Top-k LLM Results using Database Queries (Thinh On et al)
11:50 - 12:10 Paper: Towards Consistent Language Models Using Declarative Constraints (Jasmin Mousavi et al)
12:10 - 13:30 Lunch break
13:30 - 14:00 Invited talk: Deploying LLMs on Structured Data Tasks: Lessons from the Trenches (Laurel Orr / Numbers Station)
14:00 - 14:30 Invited talk: Language Model Agents for Building Natural Language Interfaces to Data (Tao Yu / HKU)
14:30 - 14:50 Paper: Generating Data Augmentation Queries Using Large Language Models (Christopher Buss et al)

Paper Submission

Please submit your manuscript in pdf to llmdb.workshop@gmail.com. Please ensure your submissions adhere to the VLDB format. Papers should be no more than 10 pages (including all figures, tables, and references).

Submission deadline: Jun 30, 2023, 9pm PST

Notification: July 20, 2023

Final version: Aug 20, 2023

Workshop: Sept 1, 2023

Workshop Organizers