Federated Learning with RAG

A comprehensive implementation combining Federated Learning using the Flower Framework with Retrieval Augmented Generation (RAG) for privacy-preserving, knowledge-enhanced AI systems.

🎯 Project Overview

This project demonstrates how to:

• Train machine learning models across distributed clients using Flower Framework
• Implement a RAG system for enhanced information retrieval
• Combine federated learning with RAG for privacy-preserving knowledge systems

🏗️ Architecture

federated-learning-cosys/
├── fl_layer/              # Federated Learning components
│   ├── server.py          # FL server implementation
│   ├── client.py          # FL client implementation
│   ├── strategy.py        # Training strategy
│   └── model.py           # Model definition
├── rag_layer/             # RAG components
│   ├── document_loader.py # Document processing
│   ├── embeddings.py      # Embedding generation
│   ├── vector_store.py    # Vector database
│   └── retriever.py       # Retrieval mechanism
├── integration/           # Integration layer
│   ├── fl_rag_system.py   # Combined system
│   └── config.py          # Configuration
├── examples/              # Usage examples
│   ├── run_fl_training.py
│   └── run_rag_query.py
├── data/                  # Sample data
└── requirements.txt

🚀 Quick Start

Installation

# Create virtual environment
python -m venv venv
source venv/bin/activate  # On Windows: venv\Scripts\activate

# Install dependencies
pip install -r requirements.txt

Running Federated Learning

# Start FL server
python examples/run_fl_training.py --mode server

# Start FL clients (in separate terminals)
python examples/run_fl_training.py --mode client --client-id 1
python examples/run_fl_training.py --mode client --client-id 2

Using RAG System

# Run RAG query example
python examples/run_rag_query.py

📚 Features

Federated Learning Layer

• ✅ Flower Framework integration
• ✅ Customizable training strategies
• ✅ Support for multiple clients
• ✅ Model aggregation and evaluation

RAG Layer

• ✅ Document loading and chunking
• ✅ Sentence embeddings with transformers
• ✅ Vector store with ChromaDB/FAISS
• ✅ Semantic search and retrieval
• ✅ Context-aware query answering

🔧 Configuration

Edit integration/config.py to customize:

• Number of federated rounds
• Client selection strategy
• Embedding model
• Vector store settings
• Document chunk size

📖 Documentation

For detailed documentation, see:

• Flower Framework Docs
• LangChain RAG Guide