Mitigating data poisoning in Federated Learning
Instruction: Discuss strategies to detect and mitigate data poisoning attacks in Federated Learning systems.
Context: Candidates must showcase their knowledge in securing Federated Learning systems against data poisoning, ensuring the integrity of the learning process.
Updated
Official answer available
Read the opening below, then unlock the full answer and practical guidance.
Data poisoning in federated learning happens when clients train on intentionally corrupted or adversarially manipulated local data so that the resulting updates push the...