DREAM CS - TUTORIAL PRICAI 2024

This repository contains code and materials for a tutorial on Retrieval-Augmented Generation (RAG), Agent, and Multi-Agent Systems, presented at PRICAI 2024. The focus is on developing a chatbot designed to enhance attendee experience by answering questions about the PRICAI 2024 conference, using data scraped from the PRICAI website.

Project Overview

This project demonstrates how to create a customer support chatbot using a multi-layered approach with RAG, Agents, and Multi-Agent systems. The chatbot can answer questions about PRICAI 2024, such as session schedules, speaker information, and more. By leveraging a Retrieval-Augmented Generation model enhanced with Agent and Multi-Agent capabilities, the chatbot provides accurate responses and handles complex queries through task specialization and interaction between multiple agents.

Key Components:

• Retrieval-Augmented Generation (RAG): Efficiently retrieves relevant information for user queries.
• Agent and Multi-Agent Systems: Enhances the chatbot's capabilities by distributing tasks among specialized agents, improving response accuracy and efficiency.

Setting Up the Environment

Ensure you have Python 3.8 or higher installed on your system.

Installation

1. Clone the repository:

   git clone https://github.com/KalbeDigitalLab/DREAM_CS-TUTORIAL-PRICAI-2024.git
   cd DREAM_CS-TUTORIAL-PRICAI-2024

2. Install required dependencies: Using requirements.txt:

   pip install -r requirements.txt

Configuration

1. Create an OpenAI API Key

   To use OpenAI's GPT-4, you need an API key. Create your key by signing up at the OpenAI website.

2. Set Up the .env File

   Create a .env file in the root directory of the project and add your OpenAI API key and any other necessary configuration variables:

   OPENAI_API_KEY=your_openai_api_key_here

   Replace your_openai_api_key_here with the API key you obtained from OpenAI.

Project Structure

The repository has the following structure:

├── .gitignore
├── README.md                                     # Project README file
├── 1. RAG Customer Support.ipynb                 # Jupyter notebook for RAG setup
├── 2. Agent Customer Support.ipynb               # Jupyter notebook with Agent setup examples
├── 3. Example Multi Agent Customer Support.ipynb # Example for Multi-Agent configuration
├── multi_agent_cs_pricai2024.py                  # Main Python script for multi-agent customer support
├── requirements.txt                              # General requirements
├── requirements.yaml                             # Conda environment file
├── requirements_ubuntu.txt                       # Requirements for Ubuntu
└── utils.py                                      # Utility functions

Usage and Running the Project

To run the project:

1. Run the Notebooks:

   • Open RAG Customer Support.ipynb to initialize and explore the RAG setup.
   • Use Agent Customer Support.ipynb for Agent setup and testing.
   • Open Example Multi Agent Customer Support.ipynb to understand and experiment with Multi-Agent configurations.

2. Run the Chatbot:

   • Execute the main script:

     python multi_agent_cs_pricai2024.py

3. Interact with the Chatbot:

   • Once running, the chatbot will be able to answer questions related to PRICAI 2024 conference content based on data from the website.

License

This project is licensed under the MIT License. See the LICENSE file for more details.