Loading…
← Back to schedule
Building a Generative AI RAG Application with Vector Search in Oracle Database 23ai
Tuesday March 25, 2025 1:30pm - 2:15pm MDT
Room 224
A key AI feature introduced in Oracle Database 23ai is vector search. A vector is an array of numbers representing the semantic content of data, rather than words or image pixels. Vectors enable similarity searches based on semantic content.

In this session, we'll explore the AI vector search features in Oracle 23ai, including generating, storing, indexing, and searching vectors, as well as integrating vector search with mission-critical enterprise capabilities.

Through sample code, we'll cover vector embedding generation methods for unstructured data, including using external embedding services and Oracle's native PL/SQL procedure, VECTOR_EMBEDDING(), to import ONNX embedding models.

We'll also discuss how to perform vector similarity searches and address challenges associated with managing high-dimensional data efficiently, ensuring query performance and accuracy, and creating new vector indexes for faster, more accurate similarity searches.

Finally, we'll explore how vector search enhances Generative AI and Large Language Models (LLMs), addressing the hallucination issue by augmenting prompts with private database content. Using examples, we will demonstrate building a Retrieval-Augmented Generation (RAG) application, incorporating tracking, embedding with the LangChain framework, and LLMs.
Speakers
avatar for Kai Yu

Kai Yu

Principal Consultant, Independent
Kai Yu is an independent principal consultant with over 30 years of experience in the tech industry, specializing in Oracle Database, Oracle Applications, Cloud, and AI and machine learning. He previously served as a Distinguished Engineer at Dell Technologies, leading Oracle Database... Read More →
Tuesday March 25, 2025 1:30pm - 2:15pm MDT
Room 224
  Data (not database), Development

Sign up or log in to save this to your schedule, view media, leave feedback and see who's attending!

Tweet Share
Share Modal

Share this link via

Or copy link