Implementation of asynchronous federated learning in flower.

Federated Learning framework extending the nnUNet

A federated learning-based intrusion detection system designed for securing edge IoT networks through decentralized anomaly detection and privacy-preserving intelligence sharing.

This project introduces a system that utilizes the Flower framework, along with the ESP32 microcontroller and the TinyDB database, for stress classification. The system collects and processes real-time biomarker data, enabling local model training on edge devices.

Secure Federated Learning system with Byzantine attack detection, trust scoring, and real-time SOC dashboard. Built with Flower (flwr), PyTorch, FastAPI, and Next.js. Final Year Project — Bahria University 2026.

Federated Learning in Satellite Constellations using Flower

Asynchronous Byzantine-robust Federated Learning system for pathology classification with differential privacy and defense filters against poisoning attacks.

Federated Learning simulation using Flower with decentralized client training, secure aggregation concepts, and SHA-256 audit logging.

Sovereign Map is a production-grade, Byzantine-tolerant Federated Learning framework. Utilizing the Mohawk Protocol for streaming aggregation, it achieves a 224x memory reduction, enabling secure orchestration of 100M+ nodes via TPM 2.0 hardware-rooted trust. Features full-stack observability with Prometheus & Grafana, built-in tokenomics telemetry

Federated Learning-based Intrusion Detection System for Smart Energy IoT environments with adversarial defense using PyTorch and Flower.

Federated Learning with 1D-CNN for Web Attack Detection on Edge-IIoTset using the Flower Framework. This project explores both IID and Non-IID data partitions to evaluate federated performance in decentralized IoT environments.

federated learning framework built with Flower and PyTorch to evaluate the robustness of FL systems under data poisoning attacks.

Federated Learning IDS — Privacy Attack Analysis

Privacy-preserving healthcare AI for global oncology research. Features policy-gated federated learning, HIPAA/GDPR compliance evidence, and a comprehensive research dashboard.

Privacy-preserving phishing email detection using Federated Learning (BiLSTM + GloVe) with Byzantine-tolerant defense against label flipping attacks.

This repository provides a comprehensive solution and codebase for the migration from centralized to federated learning. It demonstrates centralized training, its drawbacks, and how federated learning addresses these issues. It also serves as a tutorial to guide users through the transition process.

Federated learning sentiment analysis system (BiLSTM) trained across non-IID review data — IMDB, Sentiment140, Amazon. Implements AT-FedAvg: a novel Adaptive Trust-Aware aggregation strategy combining performance-based client weighting with cosine-similarity trust scoring to resist Byzantine clients. Built with PyTorch, Flower, FastAPI, Streamlit.

This repository contains the code for the the research project submitted to be able to graduate in masters of computing.

Prototype for Byzantine-robust Federated Learning using Flower framework

Privacy-preserving federated learning for NIH Chest X-ray classification. Demonstrates distributed AI that respects patient privacy by training models locally at hospitals and sharing only model updates.