Low-Latency Broadband Analog Aggregation for Federated Edge Learning

By Guangxu Zhu, Yong Wang, Kaibin Huang. 2018

46 citations recorded by the source corpus.

How researchers see this paper

The researchers whose trusted networks rank it highest, strongest advocate first. Every researcher gets a different order.

Citations in the corpus

Scheduling Policies for Federated Learning in Wireless Networks (2019, methodology): “When spectrum resources become the communication bottleneck, there are new methods exploiting the compute-over-air mechanism and arrive at a jointly decode-and-average scheme at the edge computing unit [17], [18].”
Hierarchical Federated Learning ACROSS Heterogeneous Cellular Networks (2019, background): “ng problem have not been fully addressed. To this end, there are some very recent works that focus on the distributed machine learning framework with a particular focus on wireless communications [1]–[3], [7]–[13]. Most of these works propose new communication-efﬁcient learning strategies, speciﬁc t”
Over-the-Air Machine Learning at the Wireless Edge (2019, methodology): “We first present a generalization of the analog computation approach introduced in [10], referred to as the entry-wise scheduled analog DSGD (ESA-DSGD) scheme, and then propose our scheme, built upon our previous work [12], referred to as the compressed worker-wise scheduled analog DSGD (CWSA-DSGD) ”
Federated Learning in Mobile Edge Networks: A Comprehensive Survey (2019, background): “studies in [127], [129], and [130] consider multi-access schemes and over-the-air computation, it remains to be seen if such approaches are scalable, i.”
Wireless Federated Distillation for Distributed Edge Learning with Heterogeneous Data (2019, background): “Recently, the idea of over-the-air computing has been leveraged to improve the efficiency of distributed gradient computation for parallelized machine learning [9], [10].”
On Analog Gradient Descent Learning Over Multiple Access Fading Channels (2019, background): “In [12], [13], the authors considered transmissions over fading MAC, where each entry of the gradient vector is scheduled for transmission depending on the corresponding channel condition.”
Communication-Efficient Edge AI: Algorithms and Systems (2020, background): “• Fast aggregation via over-the-air computation [21], [84], [85], [86] • Aggregation frequency control with limited bandwidth and computation resources [87], [88], [89] • Data reshufﬂing via index coding and pliable index coding for improving training performance [90], [91], [92] • Straggler…”
Machine Learning at the Wireless Edge: Distributed Stochastic Gradient Descent Over-the-Air (2019, background): “While our work can be easily extended to the fading model, the distinctive contributions of our work with respect to the concurrent works of [40] and [41] are i) the introduction of a purely digital separate gradient compression and communication scheme; ii) the consideration of a bandwidth-constrai”
Energy Efficient Federated Learning Over Wireless Communication Networks (2019, methodology): “To minimize latency, a broadband analog aggregation multi-access scheme was designed in [28] for FL by exploiting the waveform-superposition property of a multi-access channel.”
Edge Intelligence: The Confluence of Edge Computing and Artificial Intelligence (2019, background): “[44] suggested that the simultaneously transmitted model updates in Federated learning should be analog aggregated by exploiting”
Federated Learning via Over-the-Air Computation (2018, background): “Note that the tradeoff of learning performance and aggregation error is also considered in the recent parallel work [24], which quatifies the device population of the truncation-based approach for excluding the devices with deep fading channel.”
Edge Intelligence: Paving the Last Mile of Artificial Intelligence With Edge Computing (2019, unknown): “Amiri and Gunduz [57] implement DSGD at the wireless edge with the help of a remote parameter server and further develop DSGD in digital and analog schemes, respectively.”
Federated Learning Over Wireless Fading Channels (2019, methodology): “ESA-DSGD With the ESA-DSGD scheme studied in [26], each device sends its gradient estimate entirely after applying power allocation, which is to satisfy the average power constraint.”
Machine Learning in the Air (2019, background): “This approach can also be extended to the scenario with fading [78], [79], in which case power control can be employed at the workers to align their transmissions at the same received power level.”