Create a rigorous, clear and visually strong PhD defence presentation for the thesis “Robust and Trustworthy Federated Learning in Distributed Systems”. The presentation must tell a coherent research story: why federated learning becomes unreliable in adversarial and heterogeneous distributed environments, what research gaps remain, what methods were developed to address them, how they were evaluated, and what the thesis contributes overall. The final deck should feel like an expert researcher’s defence, not a generic AI-generated presentation.
Create a rigorous, clear and visually strong PhD defence presentation for the thesis “Robust and Trustworthy Federated Learning in Distributed Systems”. The presentation must tell a coherent research story: why federated learning becomes unreliable in adversarial and heterogeneous distributed environments, what research gaps remain, what methods were developed to address them, how they were evaluated, and what the thesis contributes overall. The final deck should feel like an expert researcher’s defence, not a generic AI-generated presentation.
Created using ChatSlide
This research explores the complexities of robust Federated Learning (FL), emphasising the need for explicit trust decisions in adaptive aggregation processes. It highlights challenges such as non-IID objectives and client drift, while proposing solutions like DistFL for isolating unreliable contributors and WMA for enhancing update reliability. ClusFed is introduced to select coherent client populations, ensuring resilience against attacks. The findings advocate for a comprehensive trust...