Electrical Engineering and Systems Science > Signal Processing
[Submitted on 19 Nov 2022 (v1), last revised 19 Mar 2024 (this version, v2)]
Title:Non-Coherent Over-the-Air Decentralized Gradient Descent
View PDF HTML (experimental)Abstract:Decentralized Gradient Descent (DGD) is a popular algorithm used to solve decentralized optimization problems in diverse domains such as remote sensing, distributed inference, multi-agent coordination, and federated learning. Yet, executing DGD over wireless systems affected by noise, fading and limited bandwidth presents challenges, requiring scheduling of transmissions to mitigate interference and the acquisition of topology and channel state information -- complex tasks in wireless decentralized systems. This paper proposes a DGD algorithm tailored to wireless systems. Unlike existing approaches, it operates without inter-agent coordination, topology information, or channel state information. Its core is a Non-Coherent Over-The-Air (NCOTA) consensus scheme, exploiting a noisy energy superposition property of wireless channels. With a randomized transmission strategy to accommodate half-duplex operation, transmitters map local optimization signals to energy levels across subcarriers in an OFDM frame, and transmit concurrently without coordination. It is shown that received energies form a noisy consensus signal, whose fluctuations are mitigated via a consensus stepsize. NCOTA-DGD leverages the channel pathloss for consensus formation, without explicit knowledge of the mixing weights. It is shown that, for the class of strongly-convex problems, the expected squared distance between the local and globally optimum models vanishes with rate $\mathcal O(1/\sqrt{k})$ after $k$ iterations, with a proper design of decreasing stepsizes. Extensions address a broad class of fading models and frequency-selective channels. Numerical results on an image classification task depict faster convergence vis-à-vis running time than state-of-the-art schemes, especially in densely deployed networks.
Submission history
From: Nicolò Michelusi [view email][v1] Sat, 19 Nov 2022 19:15:34 UTC (643 KB)
[v2] Tue, 19 Mar 2024 00:53:45 UTC (4,065 KB)
Current browse context:
cs.LG
References & Citations
Bibliographic and Citation Tools
Bibliographic Explorer (What is the Explorer?)
Litmaps (What is Litmaps?)
scite Smart Citations (What are Smart Citations?)
Code, Data and Media Associated with this Article
CatalyzeX Code Finder for Papers (What is CatalyzeX?)
DagsHub (What is DagsHub?)
Gotit.pub (What is GotitPub?)
Papers with Code (What is Papers with Code?)
ScienceCast (What is ScienceCast?)
Demos
Recommenders and Search Tools
Influence Flower (What are Influence Flowers?)
Connected Papers (What is Connected Papers?)
CORE Recommender (What is CORE?)
arXivLabs: experimental projects with community collaborators
arXivLabs is a framework that allows collaborators to develop and share new arXiv features directly on our website.
Both individuals and organizations that work with arXivLabs have embraced and accepted our values of openness, community, excellence, and user data privacy. arXiv is committed to these values and only works with partners that adhere to them.
Have an idea for a project that will add value for arXiv's community? Learn more about arXivLabs.