Decentralized federated meta‐learning framework for few‐shot multitask learning

Abstract

Federated learning is increasingly attractive, however as the number of training samples on a single device is too small and the training tasks of the devices are different, it faces the few-shot multitask learning problem. Moreover, federated learning frameworks are usually vulnerable to malicious attacks of the central server and diverse clients.To address these problems, we propose a decentralized federated meta-learning framework (DFMLF) for fewshot multitask learning. In DFMLF, the devices take the rapid adaptation as objective and learn the metaknowledge shared by tasks to deal with the few-shot multitask problem. In addition, DFMLF conducts crossvalidation and secure aggregation mechanism by a small number of committee nodes, which not only eliminates the central server to avoid the security risks brought by the malicious central server, but also avoids the attack of malicious devices. Moreover, to address the extra communication cost brought by the committee strategy, we propose a communication-efficient method to make the training and aggregation carried out in parallel.We conduct extensive experiments based on real-world data sets, and the experimental results demonstrate the effectiveness, robustness, and efficiency of our framework.

D2PDRF: A differential privacy recommendation algorithm based on federated learning

“…SUN [5] developed a mobile social app recommendation system using federated online learning to utilize user behavior and social connections without data sharing. LI [6] introduced a decentralized federated meta-learning for multi-task learning, avoiding central server security risks. These advancements show cloud federated learning can protect user privacy, reduce transmission and storage costs, and improve recommendation system performance and efficiency.…”