Abstract:With the breakthroughs in deep learning and contactless sensors, the recent years have witnessed a rise of ambient intelligence applications and services, spanning from healthcare delivery to intelligent home. Federated edge learning (FEEL), as a privacy-enhancing paradigm of collaborative learning at the network edge, is expected to be the core engine to achieve ambient intelligence. Sensing, computation, and communication (SC$^{2}$) are highly coupled processes in FEEL and need to be jointly designed in a task-oriented manner to achieve the best FEEL performance under stringent resource constraints at edge devices. However, this remains an open problem as there is a lack of theoretical understanding on how the SC$^{2}$ resources jointly affect the FEEL performance. In this paper, we address the problem of joint SC$^{2}$ resource allocation for FEEL via a concrete case study of human motion recognition based on wireless sensing in ambient intelligence. First, by analyzing the wireless sensing process in human motion recognition, we find that there exists a thresholding value for the sensing transmit power, exceeding which yields sensing data samples with approximately the same satisfactory quality. Then, the joint SC$^{2}$ resource allocation problem is cast to maximize the convergence speed of FEEL, under the constraints on training time, energy supply, and sensing quality of each edge device. Solving this problem entails solving two subproblems in order: the first one reduces to determine the joint sensing and communication resource allocation that maximizes the total number of samples that can be sensed during the entire training process; the second one concerns the partition of the attained total number of sensed samples over all the communication rounds to determine the batch size at each round for convergence speed maximization. The first subproblem on joint sensing and communication resource allocation is converted to a single-variable optimization problem by exploiting the derived relation between different control variables (resources), which thus allows an efficient solution via one-dimensional grid search. For the second subproblem, it is found that the number of samples to be sensed (or batch size) at each round is a decreasing function of the loss function value attained at the round. Based on this relationship, the approximate optimal batch size at each communication round is derived in closed-form as a function of the round index. Finally, extensive simulation results are provided to validate the superiority of the proposed joint SC$^{2}$ resource allocation scheme over baseline schemes in terms of FEEL performance.

Enabling Lightweight Device-Free Wireless Sensing with Network Pruning and Quantization

WiFine: Real-Time Gesture Recognition Using Wi-Fi with Edge Intelligence

Understanding Sensor Data Using Deep Learning Methods on Resource-Constrained Edge Devices.

Lightweight Device-Free Wireless Sensing Using Information-for-Complexity Strategy

Learning to Sense: Deep Learning for Wireless Sensing with Less Training Efforts

Improving Device-Edge Cooperative Inference of Deep Learning via 2-Step Pruning

Joint Model Pruning and Device Selection for Communication-Efficient Federated Edge Learning

Towards Efficient Compact Network Training on Edge-Devices

Implementing Ultra-Lightweight Co-Inference Model in Ubiquitous Edge Device for Atrial Fibrillation Detection

A Practical Fast Model Inference System Over Tiny Wireless Device

Edge-Enabled Modulation Classification in Internet of Underwater Things Based on Network Pruning and Ensemble Learning.

Radio Frequency Fingerprinting on the Edge

Keyword Spotting System and Evaluation of Pruning and Quantization Methods on Low-power Edge Microcontrollers

Toward Ambient Intelligence: Federated Edge Learning With Task-Oriented Sensing, Computation, and Communication Integration

Enabling Deep Learning on Edge Devices through Filter Pruning and Knowledge Transfer

Multistage Pruning of CNN Based ECG Classifiers for Edge Devices

An efficient pruning scheme of deep neural networks for Internet of Things applications

Three-Stage Global Channel Pruning for Resources-Limited Platform

A Fine-Grained End-to-End Latency Optimization Framework for Wireless Collaborative Inference

Cloud–Edge Collaborative Inference with Network Pruning

Energy-Efficient Processing and Robust Wireless Cooperative Transmission for Edge Inference