Learning and Unlearning Your Data in Federated Settings

Explore federated learning and machine unlearning in this 15-minute conference talk from PEPR '24. Delve into the challenges of data privacy and regulatory compliance in distributed machine learning environments. Examine how federated learning allows multiple data owners to collaboratively train models without revealing their datasets. Investigate the emerging field of machine unlearning, focusing on minimizing or removing data inputs from training sets and trained models in response to regulations like GDPR. Analyze various federated unlearning approaches, evaluating their effectiveness and performance. Gain practical guidance on implementing federated unlearning methods to address privacy concerns and regulatory requirements in collaborative machine learning scenarios.