English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

This paper considers improving wireless communication and computationefficiency in federated learning (FL) via model quantization. In the proposedbitwidth FL scheme, edge devices train and transmit quantized versions of theirlocal FL model parameters to a coordinating server, which aggregates them intoa quantized global model and synchronizes the devices. The goal is to jointlydetermine the bitwidths employed for local FL model quantization and the set ofdevices participating in FL training at each iteration. We pose this as anoptimization problem that aims to minimize the training loss of quantized FLunder a per-iteration device sampling budget and delay requirement. However,the formulated problem is difficult to solve without (i) a concreteunderstanding of how quantization impacts global ML performance and (ii) theability of the server to construct estimates of this process efficiently. Toaddress the first challenge, we analytically characterize how limited wirelessresources and induced quantization errors affect the performance of theproposed FL method. Our results quantify how the improvement of FL trainingloss between two consecutive iterations depends on the device selection andquantization scheme as well as on several parameters inherent to the modelbeing learned. Then, we show that the FL training process can be described as aMarkov decision process and propose a model-based reinforcement learning (RL)method to optimize action selection over iterations. Compared to model-free RL,this model-based RL approach leverages the derived mathematicalcharacterization of the FL training process to discover an effective deviceselection and quantization scheme without imposing additional devicecommunication overhead. Simulation results show that the proposed FL algorithmcan reduce the convergence time.

Performance Optimization for Variable Bitwidth Federated Learning in Wireless Networks