FedRAS clusters item-embedding gradients into a few shared centroids, called actions, and transmits those instead of full embeddings, cutting federated recommender communication costs by up to 96.88%.
Recommender systems
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Gradients as an Action: Towards Communication-Efficient Federated Recommender Systems via Adaptive Action Sharing
FedRAS clusters item-embedding gradients into a few shared centroids, called actions, and transmits those instead of full embeddings, cutting federated recommender communication costs by up to 96.88%.