AI 合作伙伴解决方案

使用 Oracle Database 23ai 保护 AI 文档搜索

Introduction

Building a highly secure and scalable chatbot driven by retrieval-augmented generation (RAG) to query your organization’s documents is easier than you think. By using Oracle APEX, a low-code application platform for Oracle Database 23ai, there’s no complex infrastructure to set up and no Python libraries to manage.

We start by uploading an embedding model in ONNX format. Then, using a drag-and-drop interface developed in APEX, a set of PDFs is selected to be stored and searched using vectors in Oracle Database 23ai. A simple RAG chatbot application, also built using APEX, is created to accept natural language queries. Oracle AI Vector Search retrieves relevant document chunks, and Oracle Cloud Infrastructure (OCI) Generative AI crafts precise answers—all within the safety of your OCI tenancy, with data that can't be seen by other customers.

Thanks to our Oracle partner Miracle Finland Oy for sharing this simple, yet effective, AI solution.

Demo

Demo: Secure AI Document Search with Oracle Database 23ai (2:11)

Block diagram

Oracle Apex AI Vector Search block diagram
Block diagram of a secure AI document search solution using Oracle technologies.

注:为免疑义,本网页所用以下术语专指以下含义:

  1. 除Oracle隐私政策外,本网站中提及的“Oracle”专指Oracle境外公司而非甲骨文中国 。
  2. 相关Cloud或云术语均指代Oracle境外公司提供的云技术或其解决方案。