Overview
FLORA (Federated Learning Ovulation tRacking App with Reward System) is a privacy-first mobile app that predicts ovulation dates and offers personalised health insights – without users ever handing over their raw data. Menstrual and fertility information is among the most sensitive personal data there is, so FLORA keeps it on the user’s phone and trains its AI models with Federated Learning.
Privacy is protected at every step. Fully Homomorphic Encryption secures model aggregation, local differential privacy guards against attacks on the shared models, and a blockchain-based consent system verifies, in a tamper-resistant way, what each user has agreed to share. A blockchain repository keeps track of every model version produced during training, and a token-based reward system compensates users who contribute models or data – turning privacy-preserving AI into a fair deal for its users.
FLORA was funded by the NGI TrustChain Open Call 3 and led by Prof. Pavlos Efraimidis at the ATHENA Research Center. Its results are published in “Federated Learning for Privacy-Friendly Health Apps: A Case Study on Ovulation Tracking” (Journal of Sensor and Actuator Networks, 2025) and in our 2026 study on tokenized reward systems for Federated Learning.