Enhancing Generative AI Chatbot Accuracy Using Knowledge Graph

Overview | Key Features | Quick Start

Overview

A graph-based chatbot system that uses Neo4j and LangChain to enhance generative AI responses by leveraging structured knowledge representation.

Key Features

• Graph-based knowledge representation using Neo4j
• LangChain integration for natural language understanding
• Vector similarity search for relevant document retrieval
• Hybrid search combining structured and unstructured data
• Real-time Q&A through web interface
• High accuracy on domain-specific queries

Quick Start

1. Install dependencies:

pip install langchain langchain-community langchain-openai neo4j streamlit

2. Configure environment:

export OPENAI_API_KEY="your-key"
export NEO4J_URI="your-uri" 
export NEO4J_USERNAME="username"
export NEO4J_PASSWORD="password"

3. Run application:

streamlit run app.py

Authors

• Ajay Bandi
• Jameer Babu
• Ruida Zeng
• Sai Ram Muthyala