Scheduling Policies for Federated Learning in Wireless Networks

Abstract

Motivated by the increasing computational capacity of wireless user equipments (UEs), e.g., smart phones, tablets, or vehicles, as well as the increasing concerns about sharing private data, a new machine learning model has emerged, namely federated learning (FL), that allows a decoupling of data acquisition and computation at the central unit. Unlike centralized learning taking place in a data center, FL usually operates in a wireless edge network where the communication medium is resource-constrained and unreliable. Due to limited bandwidth, only a portion of UEs can be scheduled for updates at each iteration. Due to the shared nature of the wireless medium, transmissions are subjected to interference and are not guaranteed. The performance of FL system in such a setting is not well understood. In this paper, an analytical model is developed to characterize the performance of FL in wireless networks. Particularly, tractable expressions are derived for the convergence rate of FL in a wireless setting, accounting for effects from both scheduling schemes and inter-cell interference. Using the developed analysis, the effectiveness of three different scheduling policies, i.e., random scheduling (RS), round robin (RR), and proportional fair (PF), are compared in terms of FL convergence rate. It is shown that running FL with PF outperforms RS and RR if the network is operating under a high signal-to-interference-plus-noise ratio (SINR) threshold, while RR is more preferable when the SINR threshold is low. Moreover, the FL convergence rate decreases rapidly as the SINR threshold increases, thus confirming the importance of compression and quantization of the update parameters. The analysis also reveals a trade-off between the number of scheduled UEs and subchannel bandwidth under a fixed amount of available spectrum.

Robust Federated Learning With Noisy Communication

“…Then, we give a brief definition of centralized learning. Definition 3 (Centralized learning problem under expectation-based model): Given the proposed local loss function in (13), the global loss function can be written as…”

“…A series of research concentrating on reducing the overhead in the updating step was to transmit the compressed gradient vector via exploiting the quantization scheme [11], [12]. Another research focused on scheduling the edge devices to save the transmission bandwidth [13]- [17]. Specifically, some novel updating rules were worked out, which only allowed the edge devices with significant training improvement [14], or the fast responding devices [15], to transmit their gradient vectors in each uploading round.…”

A Classification of the Enabling Techniques for Low Latency and Reliable Communications in 5G and Beyond: AI-Enabled Edge Caching

“…The work in [23], [29] promoted the use of network slicing technique to effectively allocate network resources to provide performance guarantees for URLLC services. In [9], [24], [25], [27], [28], [30] it was discussed that 5G networks would experience a shift from the conventional cloud computing setting to edge computing systems. Edge computing systems deploy computational power to the network edges to meet the requirements of low latency, high reliability, as well as supporting resource- constrained nodes which are reachable only over unreliable network connections.…”

Ultra-dense low data rate (UDLD) communication in the THz

“…In Line 7, the agent interacts with the environment and AP to obtain the next sate information and the rewards respectively and keeps on updating its local policy θ i θ i θ i . In Line 10, the agents exchange their θ i θ i θ i values so that other agents can learn an optimal policy and the new agents can learn faster based on experience gathered from other agents [23,24], as shown in Fig. 5.…”