examples

Examples

This directory contains examples and demonstrations of AI applications in finance.

Main Example: RAG_quantum.ipynb

Overview

The RAG_quantum.ipynb notebook demonstrates a Retrieval-Augmented Generation (RAG) system for IBM Quantum Computing documentation using IBM Cloud services. This example showcases how to build an intelligent question-answering system that can retrieve and synthesize information from quantum computing research papers.

What It Does

The notebook implements a complete RAG pipeline that:

1. Document Processing: Uses Docling to convert and parse PDF research papers from arXiv about quantum computing
2. Document Chunking: Employs HybridChunker to split documents into semantically meaningful chunks for efficient retrieval
3. Embedding Generation: Utilizes IBM Granite's lightweight embedding model (granite-embedding-30m-english) to convert text chunks into vector representations
4. Vector Storage: Stores embeddings in a Milvus vector database for fast similarity search
5. Question Answering: Integrates with IBM Watsonx AI (using granite-4-h-small model) to generate accurate answers based on retrieved context

Key Technologies Used

• IBM Watsonx AI: Large language model for generating responses
• IBM Granite Embeddings: Lightweight, efficient embedding model for semantic search
• Milvus: Vector database for storing and retrieving document embeddings
• Docling: Advanced document parsing and conversion library
• LangChain: Framework for building LLM applications with RAG capabilities
• Hugging Face Transformers: For tokenization and model management

Use Case

This notebook is particularly valuable for:

• Building question-answering systems for technical documentation
• Demonstrating quantum computing concepts in finance and other domains
• Showcasing integration between IBM Cloud AI services
• Learning how to implement production-ready RAG systems

Source Documents

The example processes quantum computing research papers from arXiv, including:

• Papers on quantum algorithms and applications
• Quantum computing in various domains
• Recent advances in quantum technology

Requirements

To run this notebook, you need:

• IBM Cloud account with Watsonx AI access
• IBM Watsonx API key and project ID
• Environment variables: WATSONX_APIKEY, WATSONX_PROJECT_ID, WATSONX_URL
• Python packages: langchain-ibm, langchain-huggingface, langchain-milvus, docling, transformers

Output

The notebook creates an interactive RAG system where users can:

• Ask questions about quantum computing concepts
• Get answers grounded in the source research papers
• Retrieve relevant context from the knowledge base

---

Additional Files

Other Python scripts in the extra folder contain various experiments and utilities related to quantum computing financial analysis and web scraping. See the extra folder for more details.