Energy Efficient Federated Learning over Heterogeneous Mobile Devices via Joint Design of Weight Quantization and Wireless Transmission

Chen, Rui; Li, Liang; Xue, Kaiping; Zhang, Chi; Pan, Miao; Fang, Yuguang

doi:10.1109/tmc.2022.3213766

Cited by 24 publications

(8 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…• Swarm Intelligence-based Routing and Self-Healing Protocols: Configured with 20 to 100 agents or nodes. Evaluation Phase: The performance was measured against three existing methods, referenced as [4], [18], and [25], across various metrics such as spectral efficiency, latency, and energy consumption.…”

Section: Results Analysismentioning

confidence: 99%

“…The proposed model's integration of Dynamic Spectrum Sharing (DSS), Reinforcement Learning (RL)-based Resource Allocation, Mobile Edge Computing (MEC), Differential Privacy Techniques, and Self-Healing Network Protocols, among other innovative features, results in enhanced network efficiency, resilience, and sustainability. The advancements over the existing methods [4], [18], and [25] are evident in the model's ability to sustain high throughput, efficiently utilize available spectrum, adaptively respond to dynamic network conditions, and maintain low energy consumption and latency. This multifaceted improvement underscores the effectiveness of the proposed approach in addressing the complex demands of modern wireless communication networks.…”

Section: Overall Analysismentioning

confidence: 99%

“…The experimental results underscore the efficacy of the proposed model in enhancing network capabilities while addressing the challenges of scalability, privacy, and dynamic network conditions. The proposed model demonstrated significant improvements in handling traffic demands, with throughput reaching up to 950 Mbps compared to 800 Mbps, 850 Mbps, and 900 Mbps for methods [4], [18], and [25], respectively. This represents a substantial uplift in network capacity, particularly under high-demand scenarios.…”

Section: Conclusion and Future Scopesmentioning

confidence: 99%

See 2 more Smart Citations

Design of an Iterative Method for Hybrid Device-to-Device and Network Integration via Dynamic Spectrum Sharing and Intelligent Resource Management

Ashish Bajpai

2024

jes

View full text Add to dashboard Cite

In the evolving landscape of wireless communication, the integration of Device-to-Device (D2D) communication within traditional network infrastructures emerges as a pivotal solution to meet the burgeoning demand for efficient, scalable, and reliable communication. Despite the promise of D2D to augment network capabilities, existing frameworks often falter in optimizing resource utilization, ensuring privacy and security, and maintaining high-quality service amidst dynamic network conditions and escalating data traffic. Addressing these limitations, this work introduces a comprehensive hybrid model that synergizes D2D communication with network integration through the innovative application of Dynamic Spectrum Sharing (DSS), Reinforcement Learning (RL)-based Resource Allocation, Mobile Edge Computing (MEC), Differential Privacy Techniques, Swarm Intelligence-based Routing, Cognitive Radio Networks, Self-Healing Network Protocols, and Low-Power Communication Protocols. DSS is employed to dynamically allocate spectrum, enhancing communication flexibility between D2D and network modes. RL algorithms adaptively manage resources, optimizing network performance in real-time scenarios. MEC decentralizes data processing, slashing latency and bandwidth demands. Differential privacy fortifies data security within D2D interactions in different communication scenarios. Swarm intelligence and cognitive radios bolster network resilience and spectral efficiency through self-organized routing and adaptive spectrum access. Self-healing protocols and energy-efficient designs ensure robust, sustainable network operations. The proposed model markedly elevates network efficiency, slashes latency, and augments scalability by intelligently offloading traffic, distributing computing tasks, and preserving user privacy. Its implementation fosters a resilient, self-organizing network fabric capable of meeting contemporary demands. The impacts of this work are profound, offering a blueprint for next-generation wireless networks that are more adaptive, secure, and sustainable, thereby significantly enhancing user experience and trust in D2D communication systems.

show abstract

Section: Results Analysismentioning

confidence: 99%

Section: Overall Analysismentioning

confidence: 99%

Section: Conclusion and Future Scopesmentioning

confidence: 99%

See 1 more Smart Citation

Design of an Iterative Method for Hybrid Device-to-Device and Network Integration via Dynamic Spectrum Sharing and Intelligent Resource Management

Ashish Bajpai

2024

jes

View full text Add to dashboard Cite

show abstract

“…To mitigate this issue, there are two research directions to be explored: One is to design judicious resource scheduling and transmission schemes, including client selection [20,21], resource allocation [22][23][24][25], and hierarchical model aggregation [26,27]. The other is to integrate model compression frameworks [8,9]. However, the latter introduces inevitable approximation errors.…”

Section: Related Workmentioning

confidence: 99%

“…For instance, the popular image recognition model VGG16 [7] comprises 138 million parameters (528MB in 32bit float) and requires 15.5G MACs (multiply-add computation) for forward propagation (FP) process, essentially making exclusive on-device training impractical for resourceconstrained mobile or IoT devices. One promising solution is to compress the ML model to lower both communication and computing workload [8][9][10]. However, model compression/quantization inevitably induces more inference/learning errors due to the low-precision calculations.…”

Section: Introductionmentioning

confidence: 99%

Efficient Parallel Split Learning over Resource-constrained Wireless Edge Networks

Zheng¹,

Zhu²,

Deng³

et al. 2023

Preprint

View full text Add to dashboard Cite

The increasingly deeper neural networks hinder the democratization of privacy-enhancing distributed learning, such as federated learning (FL), to resource-constrained devices. To overcome this challenge, in this paper, we advocate the integration of edge computing paradigm and parallel split learning (PSL), allowing multiple client devices to offload substantial training workloads to an edge server via layer-wise model split. By observing that existing PSL schemes incur excessive training latency and large volume of data transmissions, we propose an innovative PSL framework, namely, efficient parallel split learning (EPSL), to accelerate model training. To be specific, EPSL parallelizes client-side model training and reduces the dimension of activations' gradients for back propagation (BP) via last-layer gradient aggregation, leading to a significant reduction in server-side training and communication latency. Moreover, by considering the heterogeneous channel conditions and computing capabilities at client devices, we jointly optimize subchannel allocation, power control, and cut layer selection to minimize the per-round latency. Simulation results show that the proposed EPSL framework significantly decreases the training latency needed to achieve a target accuracy compared with the stateof-the-art benchmarks, and the tailored resource management and layer split strategy can considerably reduce latency than the counterpart without optimization.

show abstract

A comprehensive survey on client selection strategies in federated learning

Li,

Chen,

Teng

2024

Computer Networks

View full text Add to dashboard Cite

Energy Efficient Federated Learning over Heterogeneous Mobile Devices via Joint Design of Weight Quantization and Wireless Transmission

Cited by 24 publications

References 20 publications

Design of an Iterative Method for Hybrid Device-to-Device and Network Integration via Dynamic Spectrum Sharing and Intelligent Resource Management

Design of an Iterative Method for Hybrid Device-to-Device and Network Integration via Dynamic Spectrum Sharing and Intelligent Resource Management

Efficient Parallel Split Learning over Resource-constrained Wireless Edge Networks

A comprehensive survey on client selection strategies in federated learning

Contact Info

Product

Resources

About