Coded/uncoded content placement in Mobile Edge Caching (MEC) has evolved as an efficient solution to meet the significant growth of global mobile data traffic by boosting the content diversity in the storage of caching nodes. To meet the dynamic nature of the historical request pattern of multimedia contents, the main focus of recent researches has been shifted to develop data-driven and real-time caching schemes. In this regard and with the assumption that users’ preferences remain unchanged over a short horizon, the Top-K popular contents. These contents refer to the most requested content in the upcoming period. Most existing data-driven popularity prediction models, however, are not suitable for the coded/uncoded content placement frameworks. On the one hand, in coded/uncoded content placement, in addition to classifying contents into two groups, i.e., popular and non-popular, the probability of content request is required to identify which content should be stored partially/completely, where this information is not provided by existing data-driven popularity prediction models. On the other hand, the assumption that users’ preferences remain unchanged over a short horizon only works for content with a smooth request pattern. To tackle these challenges, we develop a Multiple-model (hybrid) Transformer-based Edge Caching (MTEC) framework with higher generalization ability, suitable for various types of content with different time-varying behavior, that can be adapted with coded/uncoded content placement frameworks. In this work, we consider Top-K content as the output of the 1st Stage of the proposed MTEC framework, which includes both popular and mediocre content. Simulation results corroborate the effectiveness of the proposed MTEC caching framework in comparison to its counterparts in terms of the cache-hit ratio, classification accuracy, and the transferred byte volume.

Development of advance surface Electromyogram (sEMG)-based Human-Machine Interface (HMI) systems is of paramount importance to pave the way towards emergence of futuristic Cyber-Physical-Human (CPH) worlds. In this context, the main focus of recent literature was on development of different Deep Neural Network (DNN)-based architectures that perform Hand Gesture Recognition (HGR) at a macroscopic level (i.e., directly from sEMG signals). At the same time, advancements in acquisition of High-Density sEMG signals (HD-sEMG) have resulted in a surge of significant interest on sEMG decomposition techniques to extract microscopic neural drive information. However, due to complexities of sEMG decomposition and added computational overhead, HGR at microscopic level is less explored than its aforementioned macroscopic-level, DNN-based counterparts. In this regard, we propose the HYDRA-HGR framework, which is a hybrid model for HGR that simultaneously extracts a set of temporal and spatial features through its two independent Vision Transformer (ViT)-based parallel architectures (the so called Macro and Micro paths). The Macro Path is trained directly on the pre-processed HD-sEMG signals, while the Micro path is fed with the p-to-p values of the extracted Motor Unit Action Potentials (MUAPs) of each source. Extracted features at macroscopic and microscopic levels are then coupled via a Fully Connected (FC) fusion layer for final gesture classification. We evaluate the proposed hybrid HYDRA-HGR framework through a recently released HD-sEMG dataset, and show that it significantly outperforms its stand-alone counterparts. The proposed HYDRA-HGR framework achieves average accuracy of 94.86% for the 250 ms window size, which is 5.52 % and 8.22 % higher than that of the Macro and Micro paths, respectively.

Advancements in Biological Signal Processing (BSP) and Machine-Learning (ML) models have paved the path for development of novel immersive Human-Machine Interfaces (HMI). In this context, there has been a surge of significant interest in Hand Gesture Recognition (HGR) utilizing Surface-Electromyogram (sEMG) signals. This is due to its unique potential for decoding wearable data to interpret human intent for immersion in Mixed Reality (MR) environments. To achieve the highest possible accuracy, complicated and heavy-weighted Deep Neural Networks (DNNs) are typically developed, which restricts their practical application in low-power and resource-constrained wearable systems. In this work, we propose a light-weight hybrid architecture (HDCAM) based on Convolutional Neural Network (CNN) and attention mechanism to effectively extract local and global representations of the input. The proposed HDCAM model with 58, 441 parameters reached a new state-of-the-art (SOTA) performance with 82.91% and 81.28% accuracy on window sizes of 300 ms and 200 ms for classifying 17 hand gestures. The number of parameters to train the proposed HDCAM architecture is 18.87× less than its previous SOTA counterpart.

Advances in biosignal signal processing and machine learning, in particular Deep Neural Networks (DNNs), have paved the way for the development of innovative Human-Machine Interfaces for decoding the human intent and controlling artificial limbs. DNN models have shown promising results with respect to other algorithms for decoding muscle electrical activity, especially for recognition of hand gestures. Such data-driven models, however, have been challenged by their need for a large number of trainable parameters and their structural complexity. Here we propose the novel Temporal Convolutions-based Hand Gesture Recognition architecture (TC-HGR) to reduce this computational burden. With this approach, we classified 17 hand gestures via surface Electromyogram (sEMG) signals by the adoption of attention mechanisms and temporal convolutions. The proposed method led to 81.65% and 80.72% classification accuracy for window sizes of 300 ms and 200 ms, respectively. The number of parameters to train the proposed TC-HGR architecture is 11.9 times less than that of its state-of-the-art counterpart.

As a consequence of the COVID-19 pandemic, the demand for telecommunication for remote learning/working and telemedicine has significantly increased. Mobile Edge Caching (MEC) in the 6G networks has been evolved as an efficient solution to meet the phenomenal growth of the global mobile data traffic by bringing multimedia content closer to the users. Although massive connectivity enabled by MEC networks will significantly increase the quality of communications, there are several key challenges ahead. The limited storage of edge nodes, the large size of multimedia content, and the time-variant users’ preferences make it critical to efficiently and dynamically predict the popularity of content to store the most upcoming requested ones before being requested. Recent advancements in Deep Neural Networks (DNNs) have drawn much research attention to predict the content popularity in proactive caching schemes. Existing DNN models in this context, however, suffer from long-term dependencies, computational complexity, and unsuitability for parallel computing. To tackle these challenges, we propose an edge caching framework incorporated with the attention-based Vision Transformer (ViT) neural network, referred to as the Transformer-based Edge (TEDGE) caching, which to the best of our knowledge, is being studied for the first time. Moreover, the TEDGE caching framework requires no data pre-processing and additional contextual information. Simulation results corroborate the effectiveness of the proposed TEDGE caching framework in comparison to its counterparts.

This work is motivated by the recent advancements of Deep Neural Networks (DNNs) for myoelectric prosthesis control. In this regard, hand gesture recognition via surface Electromyogram (sEMG) signals has shown a high potential for improving the performance of myoelectric control prostheses. Although the recent researches in hand gesture recognition with DNNs have achieved promising results, they are still in their infancy. The recent literature uses traditional supervised learning methods that usually have poor performance if a small amount of data is available or requires adaptation to a changing task. Therefore, in this work, we develop a novel hand gesture recognition framework based on the formulation of FewShot Learning (FSL) to infer the required output given only one or a few numbers of training examples. Thus in this paper, we proposed a new architecture (named as FHGR which refers to "Few-shot Hand Gesture Recognition") that learns the mapping using a small number of data and quickly adapts to a new user/gesture by combing its prior experience. The proposed approach led to 83.99% classification accuracy on new repetitions with few-shot observations, 76.39% accuracy on new subjects with few-shot observations, and 72.19% accuracy on new gestures with few-shot observations.

Accelerating Magnetic Resonance Imaging (MRI) acquisition process is a critical and challenging medical imaging problem as basic reconstructions obtained from the undersampled k-space often exhibit blur or aliasing effects. Despite its significance and recent advancements in the field of deep neural networks (DNNs), development of deep learning-based MRI reconstruction algorithms is not yet flourished due to unavailability of public and large datasets. The recently introduced large-scale fastMRI dataset is posed to change this state of affairs, however, existing DNN solutions developed based on fastMRI require learning a large number of parameters rendering their practical application limited due to the strict low-latency requirements of real-time MRI acquisition. In this paper, we aim to address this drawback and target reducing the computational cost associated with single-coil reconstruction task. More specifically, the paper proposes a novel deep model referred to as the SepUnet architecture achieving significant reduction in the required number of parameters while maintaining high accuracy. Performance of the proposed SepUnet architecture is evaluated based on the official test dataset from fastMRI illustrating accuracy improvement in comparison to its published counterparts while requiring significantly reduced number of trainable parameters (i.e., the SepUnet architecture is much faster and lighter than its counterparts).

Capitalizing on the goal of addressing identified shortcomings of recent solutions developed for recognition tasks via sparse multichannel surface Electromyography (sEMG) signals, the paper proposes a novel deep learning model, referred to as the XceptionTime architecture. The proposed innovative XceptionTime architecture is designed by integration of depthwise separable convolutions, adaptive average pooling, and a novel no-linear normalization technique. At the hearth of the proposed architecture is several XceptionTime modules concatenated in series fashion designed to captures both temporal and spatial information-bearing contents of the sparse multichannel sEMG signals without the need for data augmentation and manual design of feature extraction. In addition to instruction of the new XceptionTime module, by integration of adaptive average pooling, instead of fully connected layers, and utilization of a novel non-linear normalization approach, the proposed architecture is less prone to overfitting, more robust to temporal translation of the input, and more importantly is independent from the input window size, i.e., there is no need to change/reconfigure the architecture by changing the size of the input sequence. Finally, by utilizing the depthwise separable convolutions, the XceptionTime network has far less parameters resulting in less complex network.