A LinUCB contextual bandit selects federated aggregation rules online based on update variance, cosine similarity, and mean norm, claiming superior accuracy and tunable risk posture.
Manipulating the Byzantine: Optimizing Model Poisoning Attacks and Defenses for Federated Learning
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
REJECT 1representative citing papers
citing papers explorer
-
FedStrategist: A Meta-Learning Framework for Adaptive and Robust Aggregation in Federated Learning
A LinUCB contextual bandit selects federated aggregation rules online based on update variance, cosine similarity, and mean norm, claiming superior accuracy and tunable risk posture.