metafed-fl 1.0.0

Federated Learning for Metaverse Infrastructures with MARL, Privacy, and Sustainability

MetaFed-FL: Federated Learning for Metaverse Systems

---

🔬 Research Overview

MetaFed is a cutting-edge federated learning (FL) framework specifically designed for Metaverse infrastructures. This research addresses the critical challenges of privacy, performance, and sustainability in distributed AI systems.

Paper: MetaFed: Advancing Privacy, Performance, and Sustainability in Federated Metaverse Systems
Authors: Muhammet Anil Yagiz, Zeynep Sude Cengiz, Polat Goktas (2025)

🌐 Visit Project Website - Interactive demos, tutorials, and detailed documentation

---

🛠️ Installation

Prerequisites

• Python 3.9+
• PyTorch 2.2.2+
• CUDA (optional, for GPU acceleration)

Dependencies

# Core dependencies
pip install torch>=2.2.2 torchvision>=0.17.2
pip install numpy>=1.26.4 pandas>=2.2.2
pip install matplotlib>=3.9.2 timm>=1.0.8

# Development dependencies
pip install -r requirements-dev.txt

Package Installation

# Install as editable package
pip install -e .

# Or install from PyPI (when available)
pip install metafed-fl

---

🚀 Quick Start

# Clone the repository
git clone https://github.com/afrilab/MetaFed-FL.git
cd MetaFed-FL

# Install dependencies
pip install -r requirements.txt

# Run MNIST experiment
python -m experiments.mnist.run_experiment

# Run CIFAR-10 experiment
python -m experiments.cifar10.run_experiment

---

📊 Key Features

🤖 Multi-Agent Reinforcement Learning (MARL)

• Dynamic client orchestration and selection
• Adaptive resource allocation
• Intelligent scheduling algorithms

🔒 Privacy-Preserving Techniques

• Homomorphic encryption for secure aggregation
• Differential privacy for data protection
• Zero-knowledge proof mechanisms

🌱 Carbon-Aware Scheduling

• Real-time carbon intensity tracking
• Renewable energy-aligned orchestration
• Sustainable resource management

📈 Comprehensive Benchmarks

• MNIST with ResNet-18 architecture
• CIFAR-10 with ResNet-18 architecture
• Multiple federated learning algorithms
• Extensive performance metrics

---

📈 Performance Metrics

Dataset	Algorithm	Accuracy Improvement	CO2 Reduction	Efficiency Gain
MNIST	FedAvg+MARL	+15%	-35%	+20%
CIFAR-10	FedProx+MARL	+20%	-45%	+25%
CIFAR-10	SCAFFOLD+MARL	+18%	-40%	+22%

Compared to baseline federated learning algorithms

---

📂 Project Structure

MetaFed-FL/
├── src/metafed/              # Core package
│   ├── core/                 # FL core components
│   ├── algorithms/           # FL algorithms
│   ├── orchestration/        # MARL orchestration
│   ├── privacy/              # Privacy mechanisms
│   ├── green/                # Carbon-aware features
│   └── utils/                # Utilities
├── experiments/              # Experiment runners
│   ├── mnist/                # MNIST experiments
│   └── cifar10/              # CIFAR-10 experiments
├── tests/                    # Test suite
├── docs/                     # Documentation
└── scripts/                  # Utility scripts

---

🧪 Running Experiments

Command Line Interface

# MNIST with FedAvg + MARL
metafed-mnist --algorithm fedavg --orchestrator rl --rounds 100

# CIFAR-10 with privacy preservation
metafed-cifar10 --algorithm fedprox --privacy differential --epsilon 1.0

# Green-aware scheduling
metafed-mnist --green-aware --carbon-tracking

Configuration Files

# Using YAML configuration
python -m experiments.mnist.run_experiment --config configs/fedavg_privacy.yaml

Jupyter Notebooks (Legacy)

# For research and exploration
jupyter notebook experiments/notebooks/

---

🔧 Advanced Configuration

Algorithm Selection

• FedAvg: Standard federated averaging
• FedProx: Proximal federated optimization
• SCAFFOLD: Stochastic controlled averaging

Orchestration Methods

• Random: Random client selection
• RL-based: Multi-agent reinforcement learning
• Green-aware: Carbon-optimized selection

Privacy Settings

• Homomorphic Encryption: Secure aggregation
• Differential Privacy: ε-differential privacy
• Hybrid: Combined privacy mechanisms

---

🤝 Contributing

We welcome contributions! Please see our Contributing Guide for details.

Development Setup

# Clone and setup development environment
git clone https://github.com/afrilab/MetaFed-FL.git
cd MetaFed-FL
pip install -e ".[dev]"

# Run tests
pytest tests/

# Run linting
flake8 src/ tests/
black src/ tests/

Code Style

• Black for code formatting
• Flake8 for linting
• isort for import sorting
• Type hints for better code documentation

---

📄 License

This project is licensed under the MIT License - see the LICENSE file for details.

---

📝 Citation

If you use MetaFed-FL in your research, please cite:

@misc{yagiz2025metafedadvancingprivacyperformance,
  title={MetaFed: Advancing Privacy, Performance, and Sustainability in Federated Metaverse Systems},
  author={Muhammet Anil Yagiz and Zeynep Sude Cengiz and Polat Goktas},
  year={2025},
  eprint={2508.17341},
  archivePrefix={arXiv},
  primaryClass={cs.LG},
  url={https://arxiv.org/abs/2508.17341}
}

---

🌟 Acknowledgments

• AFRI Lab for supporting this research
• PyTorch Team for the excellent deep learning framework
• Federated Learning Community for inspiration and collaboration

---

🌐 Website

Built with ❤️ for the Federated Learning Community