This repository serves as a curated collection of research papers and code related to the intersection of Artificial Intelligence (AI) and Databases (DB), focusing on "AI for DB" and "DB for AI" paradigms. It targets researchers, engineers, and students interested in autonomous database systems, learned data structures, query optimization, and leveraging LLMs for database tasks. The primary benefit is a comprehensive overview of the state-of-the-art in this rapidly evolving field.

How It Works

The repository categorizes research across various aspects of database management, including configuration tuning, query optimization, database design (learned indexes, layouts), monitoring, diagnosis, and general AI techniques applied to databases. It also includes a significant section on Large Language Models (LLMs) applied to databases, covering areas like text-to-SQL, database tuning with LLMs, and LLM-augmented DBMS. The structure facilitates easy navigation and discovery of relevant work.

Quick Start & Requirements

This repository is a collection of research papers and links to code repositories, not a single installable tool. No direct installation or execution commands are applicable.

Highlighted Details

• Extensive coverage of learned index structures, including immutable, mutable, and persistent variants.
• Detailed sections on AI-driven query optimization techniques like learned cardinality estimation, cost estimation, and plan optimization.
• A dedicated and growing section on the application of Large Language Models (LLMs) in database management, including LLM-based tuning and text-to-SQL.
• Links to numerous academic papers with associated code repositories for many entries, enabling practical exploration.

Maintenance & Community

The repository is maintained by the Tsinghua Database Group, a prominent research institution in the database field. The content is updated to reflect recent advancements, particularly in the LLM x DB space.

Licensing & Compatibility

The repository itself is a collection of links and does not have a specific license. The licensing of the linked papers and code repositories varies and must be checked individually.

Limitations & Caveats

This is a curated list of research, not a ready-to-use framework. Users must navigate to individual linked papers and code repositories to evaluate and utilize specific tools or techniques. The rapidly advancing nature of the field means some entries may become dated quickly.