Search

This paper considers a WP-MEC system consisting of multiple base stations (BSs) and mobile devices (MDs), where the MDs offload tasks to the BSs for computational resources and the BSs charge the MDs using wireless power transfer (WPT). In practice, each BS and MD are equipped with a task buffer with limited size and a battery with limited capacity. First, we develop a WP-MEC system with task and energy queuing dynamics to study long-term system performance under time-varying fading channels and stochastic task and energy arrivals. Second, we propose a dynamic throughput maximum (DTM) algorithm based on perturbed Lyapunov optimization to maximize the system throughput under task and energy queue stability constraints, by optimizing the allocation of communication, computation, and energy resources. For DTM, we characterize a throughput-backlog trade-off of [O}(1/V), O(V)] to indicate that the system throughput goes up as the queue backlog increases. However, as V goes large, the system throughput can be pushed arbitrarily close to the optimum at the cost of linearly increasing queue backlog. To reduce the cost, we further develop an improved dynamic throughput maximum (IDTM) algorithm, and verify that the IDTM algorithm can achieve a trade-off of ${[O(1/V), O((\log(V))^2)]}$ between the system throughput and the queue backlog.

In recent years, convolutional regression trackers have shown increasing attention for visual object tracking due to their favorable performance and easy implementation. However, most of them are restricted to features from a certain layer and hardly benefit from temporal spatial information, which limits the potential to significant appearance changes. In this work, we go beyond the traditional deep regression trackers and build a novel twofold tracking network, which exploits rich hierarchical features and incorporates both temporal and spatial information to boost the tracking performance. The proposed network is composed of two streams, i.e., an appearance stream and a semantic stream, each stream is independently learned from different convolutional layers. Specially, we propose temporal and spatial mechanism for robust target representation by considering historical information in previous frames as well as spatial information. By design, the proposed twofold convolutional regression tracking network with spatial and temporal mechanism can better tolerate the target appearance changes and improve the tracking accuracy. Extensive experimental results on the benchmarks OTB-2015, Temple-Color, UAV123, and VOT-2018 demonstrate the effectiveness of our method, as compared with a number of state-of-the-art trackers.

Over the last decade, simultaneous wireless information and power transfer (SWIPT) has become a practical and promising solution for connecting and recharging battery-limited devices, thanks to significant advances in low-power electronics technology and wireless communications techniques. To realize the promised potentials, advanced resource allocation design plays a decisive role in revealing, understanding, and exploiting the intrinsic rate-energy tradeoff capitalizing on the dual use of radio frequency (RF) signals for wireless charging and communication. In this paper, we provide a comprehensive tutorial overview of SWIPT from the perspective of resource allocation design. The fundamental concepts, system architectures, and RF energy harvesting (EH) models are introduced. In particular, three commonly adopted EH models, namely the linear EH model, the nonlinear saturation EH model, and the nonlinear circuit-based EH model are characterized and discussed. Then, for a typical wireless system setup, we establish a generalized resource allocation design framework which subsumes conventional resource allocation design problems as special cases. Subsequently, we elaborate on relevant tools from optimization theory and exploit them for solving representative resource allocation design problems for SWIPT systems with and without perfect channel state information (CSI) available at the transmitter, respectively. The associated technical challenges and insights are also highlighted. Furthermore, we discuss several promising and exciting future research directions for resource allocation design for SWIPT systems intertwined with cutting-edge communication technologies, such as intelligent reflecting surfaces, unmanned aerial vehicles, mobile edge computing, federated learning, and machine learning., Comment: 21 pages, 5 figures, To appear in Proceedings of the IEEE

We focus on the one-example person re-identification (Re-ID) task, where each identity has only one labeled example along with many unlabeled examples. Since each identity has only one labeled example, the number of initialized label examples is small, and the body parts of person are not aligned due to changes in person pose and camera angle under the camera. Therefore, the distinguishing information of learning labeled and unlabeled examples is challenging. To overcome these problems, we propose an end-to-end multi-task training network for semi-supervised Re-ID. First, we impose a part segmentation (PS) constraint on feature maps, forcing a module to predict part labels from the feature maps and enhance alignment. Second, we carefully design the network named Multiple Branch Network (MBN). MBN is a multi-branch deep network architecture, which consisting of one branch for global feature representation and two branches for local feature representation, local feature representation that including horizontal stripes representation and PS representation, respectively. Finally, loss function fusion is designed to learn discriminative features for semi-supervised Re-ID. Specifically, the MBN model is optimized by mining the object classification loss, exclusive loss and PS loss simultaneously. We validate the effectiveness of our approach by demonstrating its superiority over the state-of-the-art methods on the standard benchmark datasets, including Market-1501, DukeMTMC-reID. Notably, the rank-1 accuracy of our method outperforms the state-of-the-art method by 15.9 points (absolute, i.e., 71.7% vs. 55.8%) on Market-1501 and 8.9 points on DukeMTMC-reID.

In this paper, we investigate the detection problem for a massive machine-type communication (mMTC) system that has correlated user activities. Two deep learning assisted algorithms are proposed to exploit the user activity correlation to facilitate channel estimation and user identification. Due to the dependency among user activities, conventional element-wise minimum mean square error (MMSE) denoiser used in the orthogonal approximate message passing (OAMP) algorithm cannot achieve satisfying performance during the two-step iterative process. Therefore, we propose a deep learning modified OAMP (DL-mOAMP) algorithm, which iteratively modifies the user activity ratio via exploiting the user activity correlation in the MMSE denoiser based on the estimated sequence during each OAMP iteration. Moreover, given a specific false alarm probability, a constant threshold employed in the conventional user identification is not optimal in the presence of user activity correlation. Thus, we propose a neural network framework that is dedicated to the user identification (DL-mOAMP-UI algorithm), which minimizes the missed detection probability under a pre-determined false alarm probability. Numerical results show that the proposed DL-mOAMP algorithm provides a substantial mean squared error performance gain compared to the conventional OAMP algorithm and the DL-mOAMP-UI algorithm can further improve the user identification accuracy of an mMTC system.

The El Niño-Southern Oscillation (ENSO) is one of the main drivers of the interannual climate variability of Earth and can cause a wide range of climate anomalies, so multi year ENSO forecasts are a paramount scientific issue. However, most existing works rely on the conventional iterative mechanism and, thus, fail to provide reliable long-term predictions due to error accumulation. Although methods based on deep learning (DL) apply the parallel modeling scheme for different lead times instead of a single iteration model, they leverage the same DL model for prediction, which can not fully mine the variability of different lead times, resulting in a decrease of prediction accuracy. To solve this problem, we propose a novel parallel deep convolutional neural network (CNN) with a heterogeneous architecture. In this study, by adaptively selecting network architectures for different lead times, we realize variability modeling of different tasks (lead times) and thereby improve the reliability of long-term predictions. Furthermore, we propose a relationship between different prediction lead times and neural network architecture from a unique perspective, namely, the receptive field originally proposed in computer vision. According to the spatio-temporal correlated area and sampling scale of lead times, the size of the convolution kernel and the mesh size of sampling are adjusted as the lead time increases. The Coupled Model Intercomparison Project phase 5 (CMIP5) from 1861 to 2004 and the Simple Ocean Data Assimilation (SODA) from 1871 to 1973 were used for model training, and the GODAS from 1982 to 2017 were used for testing the forecast skill of the model. Experimental results demonstrate that the proposed method outperforms the other well-known methods, especially for long-term predictions.

Numerous studies have shown deep hashing can facilitate large-scale image retrieval since it employs neural networks to learn feature representations and binary codes simultaneously. Despite supervised deep hashing has made great achievements under the guidance of label information, it is hardly applicable to a real-world image retrieval application because of its reliance on extensive human-annotated data. Furthermore, the pair-wise or triplet-wise unsupervised hashing can hardly achieve satisfactory performance due to the absence of local similarity of image pairs. To solve those problems, we propose a novel unsupervised deep hashing framework to learn compact binary codes, which takes the quadruplet forms as input units, called Unsupervised Deep Quadruplet Hashing with Isometric Quantization (UDQH-IQ). Specifically, by introducing the rotation invariance of images, the novel quadruplet-based loss is designed to explore the underlying semantic similarity of image pairs, which could preserve local similarity with its neighbors in Hamming space. To decrease the quantization errors, Hamming-isometric quantization is exploited to maximize the consistency of semantic similarity between binary-like embedding and corresponding binary codes. To alleviate redundancy in different bits, an orthogonality constraint is developed to decorrelate different bits in binary codes. Experimental results on three benchmark datasets indicate that our UDQH-IQ achieves promising performance.

Three-dimensional (3D) shape recognition is a popular topic and has potential application value in the field of computer vision. With the recent proliferation of deep learning, various deep learning models have achieved state-of-the-art performance. Among them, multiview-based 3D shape representation has received increased attention in recent years, and related approaches have shown significant improvement in 3D shape recognition. However, these methods focus on feature learning based on the design of the network and ignore the correlation among views. In this article, we propose a novel progressive feature guide learning network (PGNet) that focuses on the correlation among multiple views and integrates multiple modalities for 3D shape recognition. In particular, we propose two information fusion schemes from visual and feature aspects. The visual fusion scheme focuses on the view level and employs the soft-attention model to define the weights of views for visual information fusion. The feature fusion scheme focuses on the feature dimension information and employs the quantified feature as the mask to further optimize the feature. These two schemes jointly construct a PGNet for 3D shape representation. The classic ModelNet40 and ShapeNetCore55 datasets are applied to demonstrate the performance of our approach. The corresponding experiment also demonstrates the superiority of our approach.

State-of-the-art Re-ID methods based on unsupervised domain adaption transferred the domain knowledge by pre-training model on labeled source domain and fine-tuned the pre-trained model with pseudo labels generated by clustering samples on unlabeled target domain. Unfortunately, compared with supervised Re-ID methods, the performance of exiting clustering-based UDA Re-ID methods dropped sharply, since these methods utilized a not good enough baseline and seldom handled the noisy pseudo labels generated by clustering algorithm. In this paper, we try to address the UDA problem by designing a Strong Clustering-based Unsupervised Re-ID (SCURID) baseline for further research at first. Then, we integrate the hard labels and soft multilabels to learn more discriminative features on a unified angular regularized metric space. Specifically, we design the angular margin losses consisting of Hard Angular Margin Identification (HAMI) loss and Soft Angular Margin Identification (SAMI) loss. The HAMI loss can learn generalizable and discriminative features through hard pseudo labels generated by clustering on unlabeled target domain. The SAMI loss is proposed to refine the hard noisy pseudo labels through the soft multilabels obtained from peer network. Benefited from SCURID baseline and two angular margin losses, it enables the clustering-based UDA Re-ID model to alleviate the negative effect of noisy labels and toward more discriminability. Comprehensive and extensive experiments on three public available Re-ID datasets, e.g., Matket-1501, DukeMTMC-reID and MSMT17, demonstrate that our proposed method can outperform the state-of-the-art results on UDA Re-ID task.

With the advance of Convolutional Neural Network, deep hashing methods have shown the great promising performance in large-scale image retrieval. Without depending on extensive human-annotated data, unsupervised hashing is more applicable to image retrieval tasks compared to supervised methods. However, due to the lack of fine-grained supervised signals and multi-similarity constraints, most state-of-the-art unsupervised deep hashing algorithms cannot ensure the correct fine-grained similarity ranking for image pairs. In this paper, we propose a novel unsupervised deep multi-similarity hashing framework to learn compact binary codes by jointly exploiting global-aware and spatial-aware representations, called Unsupervised Deep Multi-Similarity Hashing with Semantic Structure (UDMSH). Specifically, to obtain distinguishing characteristics, we develop a sub-network by jointly learning global semantic structures from Convolutional Neural Network (CNN) and inherent spatial structures from Fully Convolutional Network (FCN). By computing the cosine distance for deep features from image pairs, we construct a similarity matrix with semantic structure, then utilize this matrix to guide hash code learning process. Based on it, we carefully design a multi-level pairwise loss to preserve the correct fine-grained similarity ranking. Furthermore, we introduce Hamming-isometric mapping into unsupervised hashing framework to decrease the quantization errors. Extensive experiments on three widely used benchmarks prove that our proposed UDMSH outperforms several state-of-the-art unsupervised hashing with respect to different evaluation metrics.

One of the major challenges of learning to hash in large-scale image retrieval is the projective transformation from raw image to binary space with preserving semantic similarity. Recently, several deep hashing methods show many excellent properties compared with traditional hashing based on hand-designed representation. However, most of the existing hashing models only pay attention to the semantic similarity between image pairs, ignoring the ranking information of retrieval results, which limits its performance. In this paper, a novel deep hashing framework, named Deep Top Similarity Hashing with Class-wise loss (DTSH-CW), is proposed to preserve semantic similarity between top images of ranking list and query images. In this proposed framework, CNNs architecture with batch normalization module is adopted to extract deep semantic characteristics. With integrating the position information of images, a top similarity loss is carefully designed to ensure the similarities between top images of ranking list and query images. Unlike pair-wise or triplet-wise loss, by directly leveraging the class labels, a cubic constraint based on Gaussian distribution is introduced to optimize objective function so as to maintain semantic variations of different classes. Furthermore, in order to solve discrete optimization problem, Two-Stage strategy is developed to provide efficient model training. Quantities of comparison experiments on three multi-label benchmark datasets show that our proposed DTSH-CW achieves promising performance compared to several state-of-the-art hashing methods.

The key of fine-grained image categorization is to locate discriminative regions and feature extraction from these regions correspond to subtle visual traits. Some of the current methods use the attention mechanism to identify the discriminative region, but ignore that there is still a large amount of non-foreground noise information in these regions. In this work, we propose a Subtler Mixed Attention Network (SMA-Net), which contains two modules: 1) Discriminative region location module uses the channel attention mechanism to construct a feature pyramid network to locate the discriminative regions. And use the positive effect of classification to screen a group of the most discriminative regions and learn through rank to learn. 2) Mixed attention module (MAM) of feature extraction that can focus on subtler and differentiated regions. We divide the feature map into intervals according to regions, and learn attention features according to regional orientation. Then the attention maps are multiplied to the input feature map for adaptive features reinforce. At the same time, MAM is a lightweight module that can be easily integrated into advanced networks without increasing too much calculation. We validated our SMA-Net through substantial experiments on Caltech-UCSD Birds (CUB-200-2011), Stanford Cars, CIFAR-10, Fish4Knowledge and Flower17. In particular, the accuracy on two widely used fine-grained datasets, CUB-2011 and Stanford Cars, reached 87.71% and 94.37%, respectively.

Extractive distillation (ED) can be used to separate mixtures with low relative volatilities that are even close to unity. It is widely adopted for the separation of aromatics and nonaromatics in t...

Fine-grained recognition is a very challenging issue, since it is difficulty to mine discriminative and subtle feature for objects with similar visual appearance. Because massive manual annotations (e.g., bounding box for discriminative regions) are time-consuming and labor-consuming, existing methods designed single form of attention model outputted discriminative regions in a weakly supervised way. In this paper, we proposed a novel method named a Spatial-Channel Aware Attention Filters (SCAF) to address the weakly supervised fine-grained recognition problem. Compared with the previous attention models, SCAF can obtain attentions-aware features from two dimensions, i.e., the spatial location of image and the channel of feature maps. With the proposed SCAF, the model can enhance the discriminative regions on both spatial and channel dimensions simultaneously. In addition, the multi-channel network multi-level structure are designed to extract richer regional features. Moreover, focal loss is introduced to balance the samples’ distribution of fine-grained image dataset. Comprehensive and comparable experiments are conducted in publicly available datasets, and the experimental results show that our method can achieve the state-of-the-art performance on fine-grained recognition tasks. For instance, we achieve 99.370%, 80.749% accuracy on two underwater datasets respectively, i.e., Fish4Knowlege and Wild Fish.

Weather recognition is a significant technique for many potential computer vision applications in our daily lives. Generally, most existing works treat weather recognition as a single-label classification task, which cannot describe the weather conditions comprehensively due to the complex co-occurrence dependencies between different weather conditions. In this paper, we propose a novel Graph Convolution Networks with Attention (GCN-A) model for multi-label weather recognition. To our best knowledge, this is the first attempt to introduce GCN into weather recognition. Specifically, we employ GCN to capture weather co-occurrence dependencies via a directed graph. The graph is built over weather labels, where each node (weather label) is represented by word embeddings of a weather label. Furthermore, we design a re-weighted mechanism to build weather correlation matrix for information propagation among different nodes in GCN. In addition, we develop a channel-wise attention module to extract informative semantic features of weather for effective model training. Compared with the state-of-the-art methods, experiment results on two widely used benchmark datasets demonstrate that our proposed GCN-A model achieves promising performance.

Short QT syndrome (SQTS) is a genetic disease characterized by constantly short QT intervals and high risks of sudden death. SQTS6 is one of the identified SQTS genotype variants associated with the CACNA2D1 S755T mutation. However, the pathogenesis of SQTS induced arrhythmias remains unclear. To identify the underlying mechanisms of SQTS6 induced arrhythmias, a multi-scale human ventricle model comprising cell to organ levels was built. Cellular data was fitted at the cell level to reproduce the electrophysiological alterations reported in experiments. The influences were further explored at tissue and organ levels using idealized strand or tissue sheet models, and realistic ventricular slice and three-dimensional organ models. Simulation results suggested that, at the cellular level, the action potential duration (APD) and the effective refractory period (ERP) of myocytes were significantly abbreviated in the mutation condition. The unevenly changed APD and ERP led to transmural heterogeneity remodeling, and resulted in decreased temporal vulnerability. In addition, the S755T mutation shortened the critical length for initiating reentrant spiral waves, which enhanced the spatial vulnerability and provided substrates for reentry arrhythmias. Regarding the sustainability of arrhythmias, the evoked spiral waves or scroll waves persisted in the mutation condition but did not persist in the wild-type condition. The present study clearly suggested that the CACNA1DC S755T mutation can facilitate the initiation and maintenance of ventricular arrhythmias, and therefore contributes to higher risks of ventricular arrhythmias in SQTS6 patients.

Dear Editor, In China, the ‘pull technique’ is increasingly used for catheter removal in peritoneal dialysis (PD) patients. When this technique is employed, the cuff is retained after PD catheter r...

Person re-identification (Re-ID) has attracted more attention in computer vision tasks recently and achieved high accuracy in some public available datasets in a supervised manner. The performance drops significantly when datasets are unlabeled, which limits the scalability of Re-ID algorithms in practical applications. Despite some unsupervised methods are proposed to address the scalability problem of Re-ID, it’s hard to learn discriminative feature representations due to the lack of pairwise labels in different camera views. To overcome this problem, we propose an end-to-end network named Adaptive Attention-Aware Network for unsupervised person re-identification. Specifically, we propose a novel adaptive attention-aware module that could be easily embedded into Re-ID architecture. The proposed module focuses on learning strong expressive relationship among channels of feature maps, and alleviating the key problems of Re-ID, e.g., occlusion and local deformation. In addition, we extract the camera-invariant features by adopting camera-style transfer feature learning since matching pairs in Re-ID suffers from appearance changes under different camera views. Besides, unsupervised hard negative mining is introduced to learn large intra-person appearance variance and discriminate high inter-person appearance similarity in an unlabeled target dataset with an auxiliary labeled dataset. Comprehensive experiments on three public available Re-ID datasets demonstrate that our method can achieve the state-of-the-art results of unsupervised Re-ID and is competitive with supervised learning.

For multi-label image retrieval based on deep hashing, the ultimate challenge is to map from the original image to binary space while preserving high-level semantic similarity. Recently, many supervised deep hashing approaches for multi-label image retrieval have been proposed to generate high-quality binary codes. However, most such methods are only introduced to learn simple similarity based on these image characteristics, therein ignoring complex multilevel semantic similarity with fine-grained features. In this paper, we propose a framework named deep hashing with fine-grained feature learning (DH-FFL) to preserve complex multilevel semantic similarity between multi-label image pairs. In this proposed model, compact bilinear pooling convolutional neural networks (CNNs) with normalization are adopted to extract fine-grained feature descriptors. In addition, a novel multilevel contrastive loss is designed to preserve multilevel semantic similarity by introducing a zero-loss parameter. Moreover, a multi-label classification loss is used to maintain the unique semantic structure of each image and maximize the distinguishing ability of binary codes. Comprehensive experiments on three benchmark datasets show that the proposed DH-FFL achieves promising performance compared with other state-of-the-art multi-label image retrieval applications.

In recent years, deep convolutional features have been applied to discriminative correlation filters-based methods, which have achieved impressive performance in tracking. Most of them utilize hierarchical features from a certain layer. However, this is not always sufficient to learn target appearance changes and to suppress the background interference in complicated interfering factors (e.g., deformation, fast motion, low resolution, and rotations). In this paper, we propose an adaptive multi-branch correlation filter tracking method, by constructing multi-branch models and using an adaptive selection strategy to improve the accuracy and robustness of visual tracking. Specially, the multi-branch models are introduced to tolerate temporal changes of the object, which can serve different circumstances. In addition, the adaptive selection strategy incorporates both foreground and background information to learn background suppression. To further improve the tracking performance, we propose a measurement method to handle tracking failures from unreliable samples. Extensive experiments on OTB-2013, OTB-2015, and VOT-2016 datasets show that the proposed tracker has comparable performance compared to state-of-the-art tracking methods. Especially, on the OTB-2015, our method significantly improves the baseline with a gain of 5.5% in overlap precision.

In this paper, we study the resource allocation and trajectory design for secure unmanned aerial vehicle (UAV)-enabled communication systems, where multiple multi-purpose UAV base stations are dispatched to provide secure communications to multiple legitimate ground users (GUs) in the existence of multiple eavesdroppers (Eves). Specifically, by leveraging orthogonal frequency division multiple access (OFDMA), active UAV base stations can communicate to their desired ground users via the assigned subcarriers while idle UAV base stations can serve as jammer simultaneously for communication security provisioning. To achieve fairness in secure communication, we maximize the average minimum secrecy rate per user by jointly optimizing the communication/jamming subcarrier allocation policy and the trajectory of UAVs, while taking into account the constraints on the minimum safety distance among multiple UAVs, the maximum cruising speed, the initial/final locations, and the existence of cylindrical no-fly zones (NFZs). The design is formulated as a mixed integer non-convex optimization problem which is generally intractable. Subsequently, a computationally-efficient iterative algorithm is proposed to obtain a suboptimal solution. Simulation results illustrate that the performance of the proposed iterative algorithm can significantly improve the average minimum secrecy rate compared to various baseline schemes., Accepted for publication, IEEE TCOM

With the development of multimedia technology, 3D model has been applied in many fields such as mechanical design, construction industry, entertainment industry, medical treatment and so on. The number of 3D model is becoming more and more in our lives. Therefore, effective automatic management and classification of 3D models become more and more important. In this paper, we propose a dual-meta-learner model based on LSTM to learn the exact optimization algorithm used to train another two learner neural network classifier in the few-shot regime. The parametrization of our model allows it to learn appropriate parameter updates specifically for the scenario where a set amount of updates will be made, while it can also achieve a general initialization of the learner (classifier) network that allows for quick convergence of training. Our method attains state-of-the-art performance by significant margins.

Person re-identification plays important roles in many practical applications. Due to various human poses, complex backgrounds and similarity of person clothes, person re-identification is still a challenging task. In this paper, we mainly focus on the robust and discriminative appearance feature representation and proposed a novel multi-appearance method for person re-identification. First, we proposed a deep feature fusion method and get the multi-appearance feature by combining two Convolutional Neural Networks. Then, in order to further enhance the representation of the appearance feature, the multi-part model was constructed by combining the whole body and the six body parts. Additionally, we optimized the feature extraction process by adding a pooling layer. Comprehensive and comparative experiments with the state-of-the-art methods over publicly available datasets demonstrated that the proposed method can get promising results.

Computer-aided drug design uses high-performance computers to simulate the tasks in drug design, which is a promising research area. Drug-target affinity (DTA) prediction is the most important step of computer-aided drug design, which could speed up drug development and reduce resource consumption. With the development of deep learning, the introduction of deep learning to DTA prediction and improving the accuracy have become a focus of research. In this paper, utilizing the structural information of molecules and proteins, two graphs of drug molecules and proteins are built up respectively. Graph neural networks are introduced to obtain their representations, and a method called DGraphDTA is proposed for DTA prediction. Specifically, the protein graph is constructed based on the contact map output from the prediction method, which could predict the structural characteristics of the protein according to its sequence. It can be seen from the test of various metrics on benchmark datasets that the method proposed in this paper has strong robustness and generalizability.

Molecular property prediction is important to drug design. With the development of artificial intelligence, deep learning methods are effective for extracting molecular features. In this paper, we propose a multichannel substructure-graph gated recurrent unit (GRU) architecture, which is a novel GRU-based neural network with attention mechanisms applied to molecular substructures to learn and predict properties. In the architecture, molecular features are extracted at the node level and molecule level for capturing fine-grained and coarse-grained information. In addition, three bidirectional GRUs are adopted to extract the features on three channels to generate the molecular representations. Different attention weights are assigned to the entities in the molecule to evaluate their contributions. Experiments are implemented to compare our model with benchmark models in molecular property prediction for both regression and classification tasks, and the results show that our model has strong robustness and generalizability.

ECG is a non-invasive tool used to detect cardiac arrhythmias. Many arrhythmias classification solutions with various ECG features have been reported in literature. In this work, a new method combined with a novel morphological feature is proposed for accurate recognition and classification of arrhythmias. First, the events of the ECG signals are detected. Then, parametric features of ECG morphology, i.e., amplitude, interval and duration, are extracted from selected ECG regions. Next, a novel feature for analyzing QRS complex morphology changes as visual patterns as well as a new clustering-based feature extraction algorithm is proposed. Finally, the feature vectors are applied to three well-known classifiers (neural network, SVM, and KNN) for automatic diagnosis. The proposed method was assessed with all fifteen types of heartbeats as recommended by the Association for Advancement of Medical Instrumentation from the MIT-BIH arrhythmia database and achieved the best overall accuracy of 97.70% based on KNN, using the combined parametric and visual pattern features of ECG Morphology. The accuracies for the six main types - normal (N), left bundled branch blocks (L), right bundled branch blocks (R), premature ventricular contractions (V), atrial premature beats (A) and paced beats (P) are 97.79%, 99.50%, 99.59%, 97.69%, 89.70%, and 99.92%, respectively. Comparisons with peer works prove a marginal progress in automatic heart arrhythmia classification performance.

Cardiovascular disease is the leading cause of death worldwide and kills over 17 million people per year. In the recent decade, growing epidemiological evidence links air pollution and cardiac arrhythmias, suggesting a detrimental influence of air pollution on cardiac electrophysiological functionality. However, the proarrhythmic mechanisms underlying the air pollution-induced cardiac arrhythmias are not fully understood. The purpose of this work is to provide recent advances in air pollution-induced arrhythmias with a comprehensive review of the literature on the common air pollutants and arrhythmias. Six common air pollutants of widespread concern are discussed, namely particulate matter, carbon monoxide, hydrogen sulfide, sulfur dioxide, nitrogen dioxide, and ozone. The epidemiological and clinical reports in recent years are reviewed by pollutant type, and the recently identified mechanisms including both the general pathways and the direct influences of air pollutants on the cellular electrophysiology are summarized. Particularly, this review focuses on the impaired ion channel functionality underlying the air pollution-induced arrhythmias. Alterations of ionic currents directly by the air pollutants, as well as the alterations mediated by intracellular signaling or other more general pathways are reviewed in this work. Finally, areas for future research are suggested to address several remaining scientific questions.

Ultra-large-scale molecular docking can improve the accuracy of lead compounds in drug discovery. In this study, we developed a molecular docking piece of software, Vina@QNLM, which can use more than 4,80,000 parallel processes to search for potential lead compounds from hundreds of millions of compounds. We proposed a task scheduling mechanism for large-scale parallelism based on Vinardo and Sunway supercomputer architecture. Then, we readopted the core docking algorithm to incorporate the full advantage of the heterogeneous multicore processor architecture in intensive computing. We successfully expanded it to 10, 465, 065 cores (1,61,001 management process elements and 0, 465, 065 computing process elements), with a strong scalability of 55.92%. To the best of our knowledge, this is the first time that 10 million cores are used for molecular docking on Sunway. The introduction of the heterogeneous multicore processor architecture achieved the best speedup, which is 11x more than that of the management process element of Sunway. The performance of Vina@QNLM was comprehensively evaluated using the CASF-2013 and CASF-2016 protein–ligand benchmarks, and the screening power was the highest out of the 27 pieces of software tested in the CASF-2013 benchmark. In some existing applications, we used Vina@QNLM to dock more than 10 million molecules to nine rigid proteins related to SARS-CoV-2 within 8.5 h on 10 million cores. We also developed a platform for the general public to use the software.

Many engineered approaches have been proposed over the years for solving the hard problem of performing indoor localization. However, specialising solutions for the edge cases remains challenging. Here we propose to build the solution with zero hand-engineered features, but having everything learned directly from data. We use a modality specific neural architecture for extracting preliminary features, which are then integrated with cross-modality neural network structures. We show that each modality-specific neural architecture branch is capable of estimating the location with good accuracy independently. But for better accuracy a cross-modality neural network fusing the features of those early modality-specific representations is a better proposition. Our multimodal neural network, MM-Loc, is effective because it allows the uniform flow of gradients during training across modalities. Because it is a data driven approach, complex features representations are learned rather than relying heavily on hand-engineered features.

The growing demand for iterative computation over large-scale graphs has attracted a lot of enthusiasm. Distributed-disk systems can accommodate the high-level scalability requirement as graphs grow in size, but the computation is greatly expensive due to a large number of communications and a high frequency of random data-accesses. Alleviating the two limiting factors pose great challenges for graph partitioning, disk-oriented data management and the iterative mechanism. This paper derives insights from the natural locality of raw graphs and then proposes a lightweight partitioning algorithm GPNL with the goal of balancing load and accelerating communication. Accordingly, a hybrid index RC-Index is proposed to improve the I/O-efficiency by reducing disk-accesses for graph data and message data. We also introduce an across-iteration mechanism (AIM) based on the extended BSP model, and then design two policies AIMP and AIMC to prune the message scale and accelerate the message-spreading respectively. Comprehensive experiments versus the state-of-the-art solutions demonstrate significant performance gains over a broad spectrum of real-world and synthetic graphs with up to billion edges.

The prediction of drug–target affinity (DTA) is a crucial step for drug screening and discovery. In this study, a new graph-based prediction model named SAG-DTA (self-attention graph drug–target affinity) was implemented. Unlike previous graph-based methods, the proposed model utilized self-attention mechanisms on the drug molecular graph to obtain effective representations of drugs for DTA prediction. Features of each atom node in the molecular graph were weighted using an attention score before being aggregated as molecule representation. Various self-attention scoring methods were compared in this study. In addition, two pooing architectures, namely, global and hierarchical architectures, were presented and evaluated on benchmark datasets. Results of comparative experiments on both regression and binary classification tasks showed that SAG-DTA was superior to previous sequence-based or other graph-based methods and exhibited good generalization ability.

The recently proposed orthogonal time frequency space (OTFS) modulation has demonstrated superior performances over the conventional orthogonal frequency-division multiplexing (OFDM) modulation. In this paper, we provide a tutorial to the OTFS modulation for future wireless communications. We first revisit the properties of the wireless channels and then introduce the OTFS system models. In particular, we also discuss the characteristics of OTFS signals over different domains. Based on the OTFS characteristics, we briefly discuss the performances of OTFS modulation, including the performances of diversity and coding gain, channel estimation, and detection. Finally, we provide a discussion regarding the potential applications of OTFS modulation for wireless communications, where the research challenges are also highlighted.

Background: Multiple sclerosis (MS) is an immune-mediated demyelinating disease of the central nervous system. Pathogenesis is closely related to environment, genetic and immune system, but the underlying interactions are still not clearly elucidated. Our study aims to uncover the genetic basis and immune landscape of multiple sclerosis pathogenesis with bioinformatics.Methods: In our study, gene matrix were retrieceed from gene expression database NCBI GEO. We then used bioinformatics to standardize the samples and obtain differential gene expressions (DEGs). We constructed a protein-protein interaction network with DEGs on the STRING website, and used Cytohubba plug-in and MCODE plug-in to to mine hub genes. Then we had a functional enrichment hub genes and DEGs. Meanwhile, we use CIBERSORTX algorithm to explore the characteristics of immune cells infiltration in brain tissues of multiple sclerosis, and did a Spearman correlation analysis between genes and immune cells. We also analyzed the correlation between genes and types of brain tissues with WGCNA method.Results: We included a total of 90 samples from 2 datasets in the study, extracted 882 differential genes and 10 hub genes closely related to multiple sclerosis. Functional enrichment analysis suggested roles of immune response in multiple sclerosis. Besides, with CIBERSORTX algorithm we found brain tissues of MS contain a variety of infiltrating immune cells. Correlation analysis suggested that hub genes are highly relevant to chronic active white matter lesions and certain hub genes in our study may play a role in the activation of immune cells such as macrophages and natural killer cells.Conclusions: Our study provides guidance for the study of genetic basis and immune infiltration mechanism of multiple sclerosis.

The integrated access and backhaul (IAB) has demonstrated tremendous potential in the fifth-generation (5G) network, acting as an economical and sustainable alternative to conventional fiber-like architectures. Although promising, how to efficiently and effectively allocate bandwidth and power resource to the backhaul and access links is a challenging issue for IAB. In this paper, we propose a novel scheme to jointly allocate the spectrum and transmitting power of the IAB donor and IAB nodes, so as to serve their respective access links and backhaul links. With the objective of maximizing the downlink data rate, we formulate the spectrum allocation and power management as a mix-integer and non-linear programming problem. Then, we propose to use an advanced deep reinforcement learning (DRL), i.e., double deep Q-learning, to achieve an efficient policy for joint spectrum allocation and power management, namely SAPM-DDQN. We perform the proposed SAPM-DDQN in a decentralized manner, which can significantly reduce the transmission overhead, compared to the centralized one. To the best of the authors’ knowledge, this is the first attempt on the joint spectrum allocation and power management in a decentralized manner for IAB, leveraging the property of DRL techniques. The proposed SAPM-DDQN does not require any prior information from other units for optimization, which is suitable for practical deployment. The extensive simulation results are presented to verify the effectiveness of the proposed scheme for joint spectrum allocation and power management.

Recently, deep neural networks have received intense interests in polarimetric synthetic aperture radar (PolSAR) image classification. However, its success is subject to the availability of large amounts of annotated data which require great efforts of experienced human annotators. Aiming at improving the classification performance with greatly reduced annotation cost, this paper presents an active deep learning approach for minimally supervised PolSAR image classification, which integrates active learning and fine-tuned convolutional neural network (CNN) into a principled framework. Starting from a CNN trained using a very limited number of labeled pixels, we iteratively and actively select the most informative candidates for annotation, and incrementally fine-tune the CNN by incorporating the newly annotated pixels. Moreover, to boost the performance and robustness of the proposed method, we employ Markov random field (MRF) to enforce class label smoothness, and data augmentation technique to enlarge the training set. We conducted extensive experiments on four real benchmark PolSAR images, and experiments demonstrated that our approach achieved state-of-the-art classification results with significantly reduced annotation cost.

To optimize the flight performance under multi-waypoint required time of arrival constraints, characteristics of the vertical flight trajectory during the en route descent process are studied. A multi-constraint segment sequence model including flight distance, flight altitude and arrival time is constructed. To ensure the rationality of constraints, the rationality detection model for multi-constraint is established. The control variables and their variation ranges are determined by analyzing the optimization process. Mathematical models of optimization objectives are studied from three aspects: fuel economy, greenhouse effect and variation of flight speed. Based on the multi-objective genetic algorithm, the optimization solution model is established. Effects of speed on the objectives are analyzed. The results indicate that the optimization model can effectively optimize the flight parameters with multiple RTA constraints. The impact of aviation on the environment can be effectively reduced. And the optimization method has a good trade-off between fuel consumption and temperature rise by changing weighted factors. Optimization results of the flight parameters are mainly affected by the maximum RTA time constraint. The proposed optimization method provides a reference for the optimization of flight parameters and trajectory under multi-RTA constraints in four dimensions.

With the continuous growth of air traffic flow, some airports in China or other countries begin to construct Closely Spaced Parallel Runways (CSPR) to meet the capacity requirement of civil aviation. In this paper, the simulation and calculation method of flight trajectory is developed, as well as the collision risk calculation method and wake vortex encounter risk calculation method. New methods for departure and go-around procedures are proposed to achieve approximate segregated parallel operation in an attempt to obtain approximately the same results on closely spaced parallel runways as on widely spaced parallel runways. By comparing with the independent parallel departure and segregated parallel operation in ICAO Doc. 9643, it is found that the lateral separation between aircraft in this proposal is not less than the separations in the other two cases. Based on the simulation calculation of flight trajectories under different conditions, the probabilities of collision conflict and encountering wake vortex are lower than those in current operation plan. The proposed plan has no special requirements in pilot operation, control command procedures, airport facilities, or meteorological observation and prediction, so it is convenient to promote and implement at the airports with closely spaced parallel runways. Keywords: Air traffic control, Air transportation, Closely spaced parallel runways, Segregated parallel operation, Wake vortex encounter, Wake vortex separation standards

Person re-identification plays a critical part in many surveillance applications. Due to complicated illumination environments and various viewpoints, it is still a challenging problem to extract robust features. To solve this issue, we propose a novel deep feature ranking scheme. Our main contribution is to rank achieved deep features, which are obtained by classic deep learning model, and set the sort order number as our feature vector, named as ordinal deep features (ODFs). Person re-identification results are acquired by ranking person candidates by measuring distance based on ODFs. Since applying for rank orders rather than original feature values, our method achieves robust results, especially under the situation of viewpoints shift. Comprehensive experiments are carried out to demonstrate the significance of the proposed feature. Meanwhile, comparative experiments are applied over the publicly available dataset, our method achieves promising performance and outperforms the state of the art methods. Moreover, we applied the proposed feature in the scenario of image classification and discussed the effectiveness.

The fruit of Xanthium strumarium L. (Xanthii Fructus), a traditional Chinese medicine commonly used to treat rhinitis, need to be processed to reduce their toxicity before clinical use. The chemical components and colour features of Xanthii Fructus change during processing, and thus these changes could be used to evaluate the processing degree and quality of Xanthii Fructus. In the present study, high performance liquid chromatography equipped with a diode array detector (HPLC-DAD) was used to quantify the chemical components in Xanthii Fructus. A computer vision system (CVS) was also established to determine three kinds of colour spaces, including red-green-blue (RGB), L*a*b*, and HSV (hue, saturation and value), and the colour values were used to predict the chemical component content using machine learning. A support vector machine (SVM) and an artificial neural network (ANN) were employed to establish regression models between the colour values and the chemical composition contents. The different colour values were chosen as inputs of the model, and the contents of each chemical component were considered as outputs, respectively. The experimental results showed that a CVS combined with an SVM could be a cost-efficient, easy-to-build, and rapid detection system for on-line monitoring of Xanthii Fructus processing compared with those obtained using an ANN.

Non-orthogonal multiple access (NOMA) has been widely recognized as a promising multiple access scheme for realizing next generation wireless communications. Unlike existing NOMA schemes assuming perfectly time synchronized user’s signals received at the base station (BS), in this paper, we investigate the achievable rates of uplink NOMA with asynchronized transmission. By invoking Szego’s Theorem, we derive both the upper- and lower-bounds of the achievable rates of asynchronized NOMA (aNOMA) systems. In particular, we reveal that the derived lower-bound is essentially the achievable rate for conventional synchronized NOMA systems, which indicates that the asynchronization is not necessarily a foe. More specifically, we show that aNOMA systems are superior to conventional NOMA systems in terms of the achievable rates with non-sinc shaping pulses. Important insights are also unveiled based on the derived bounds. Simulation results confirm the validity of our derived analysis and demonstrate considerable achievable rates gains of aNOMA systems over conventional NOMA systems.

Aiming at the prominent problem of short durability life of concrete in saline soil area and the shortcomings of indoor accelerated test, an outdoor field exposure test was designed. The concrete specimens were semiburied in the Tianshui area with salinized soil characteristics, and nondestructive testing was conducted every 180d (days). The durability evaluation parameters and mechanical performance indexes were selected for macroscopic analysis, and the corrosion mechanism was analyzed by using the SEM image and the XRD phase. The Birnbaum-Saunders model based on physical failure and probability statistics was used for life prediction. The results show that there are rod-shaped and chip-shaped crystals growing from the surface of the gel and the internal holes in the exposed end and the embedded end of the concrete. However, the damage and deterioration of the buried end are more serious than those of the exposed end. The corrosion products mainly included ettringite, gypsum, calcium carbonate, sodium sulfate hydrate, carbosilite, and Friedel’s salt. The reliability life curve based on the Birnbaum-Saunders model can describe the whole process of exposed concrete from damage accumulation to failure. In addition, the dynamic modulus degradation index is more sensitive to concrete durability damage, and the life obtained by the Birnbaum-Saunders model is shorter than the quality degradation index. The life obtained by this degradation index is taken as the life of the concrete exposed in the saline soil site, and the concrete life of C30, C40, and C50 is about 3340d, 3930d, and 4360d, respectively.

Air pollution has been long recognized as a hazardous factor for the human cardiovascular system. Sulfur dioxide $(SO_{2})$ is a common ambient air pollutant that is able to cause detrimental effects on hearts. Though the cardiotoxicity effects by sulfur dioxide were well documented in epidemiological reports, however, the underlying mechanisms remain unclear owing to the technical limitations that exist in traditional experimental measures. In this article, we developed a multi-scale virtual ventricular tissue, which incorporated electrophysiological activities from subcellular to tissue levels and could provide comprehensive records and insightful mechanisms of the SO 2 induced ventricular arrhythmias. Based on the available cellular and molecular experimental data, our findings provide a rationale at tissue level in support of epidemiologic studies pointing to the deleterious effects of SO 2 pollution on cardiac function.

This paper investigates multiuser multi-input single-output (MISO) downlink communications assisted by a self-sustainable intelligent reflection surface (IRS), which can harvest power from the received signals. We study the joint design of the beamformer at an access point (AP) and the phase shifts and the power harvesting schedule at an IRS for maximizing the system sum-rate. The design is formulated as a non-convex optimization problem taking into account the capability of IRS elements to harvest wireless power for realizing self-sustainability. Subsequently, we propose a computationally-efficient alternating algorithm to obtain a suboptimal solution to the design problem. Our simulation results unveil that: 1) there is a non-trivial trade-off between the system sum-rate and self-sustainability in IRS-assisted systems; 2) the performance gain achieved by the proposed scheme is improved with an increasing number of IRS elements; 3) an IRS equipped with small bit-resolution discrete phase shifters is sufficient to achieve a considerable system sum-rate of an ideal case with continuous phase shifts., submitted to IEEE Global Commun. Conf. (GLOBECOM), Taiwan, Dec. 2020

In this paper, we propose the use of intelligent reflecting surface (IRS) in unmanned aerial vehicle (UAV)- based orthogonal frequency division multiple access (OFDMA) communication systems. The proposed scheme exploits both the rich beamforming gain brought by the IRS and the high mobility of UAV for improving the system sum-rate. The joint design of UAV’s trajectory, IRS scheduling, and communication resource allocation for the proposed system is formulated as a non-convex optimization problem to maximize the system sum-rate. The existence of an IRS introduces both frequencyselectivity and spatial-selectivity in the fading of the composite channel from the UAV to ground users. To facilitate the design, we first derive the expression of the composite channel gain and propose a parametric approximation approach to establish a lower bound for the formulated problem. An alternating optimization algorithm is devised to handle the lower bound optimization problem. Simulation results unveil the promising sum-rate gain achieved by the deployment of an IRS in UAVbased communication systems.

Unmanned aerial vehicle (UAV)-assisted communication becomes a promising technique to realize the beyond fifth generation (5G) wireless networks, due to the high mobility and maneuverability of UAVs which can adapt to heterogeneous requirements of different applications. However, the movement of UAVs impose challenge for accurate beam alignment between the UAV and the ground user equipment (UE). In this letter, we propose a deep learning-based location-aware predictive beamforming scheme to track the beam for UAV communications in a dynamic scenario. Specifically, a long short-term memory (LSTM)-based recurrent neural network (LRNet) is designed for UAV location prediction. Based on the predicted location, a predicted angle between the UAV and the UE can be determined for effective and fast beam alignment in the next time slot, which enables reliable communications between the UAV and the UE. Simulation results demonstrate that the proposed scheme can achieve a satisfactory UAV-to-UE communication rate, which is close to the upper bound of communication rate obtained by the perfect genie-aided alignment scheme., 4 pages, 4 figures, submitted to IEEE Wireless Communications Letters

Accurate estimation of food volume is critical in the medical field. However, estimating food volume is a challenging task due to the diverse nature of food, multi-scale and other characteristics. In this paper, we explore the relationship between the properties and volume of the object (food and reference) in the image. By combining Faster R-CNN, Grabcut, Median filtering, and CNN algorithm, we propose a framework for estimating food volume based on reference. The framework uses a front view which contains reference and food to estimate food volume and is applied to image datasets for 5 kinds of foods. The experimental results show the effective performance of this method for predicting volume, and the mean absolute error of each kind of food is less than 4.5%, which shows the model is robust to estimate volume for irregular food.

With traditional embedded memories, such as eFlash and SRAM, facing major challenge of scaling beyond 28 nm, STT-MRAM stands out from competing emerging NVM technologies as the preferred technology for both embedded flash and SRAM replacements. Having high data retention, reflow compatibility and high endurance, MRAM is ready for production to replace eFlash at 28 nm node and below. On the other end of the spectrum, by tuning magnetic properties of magnetic tunnel junction storage layers, MRAM can achieve high speed/endurance, and offer higher density compared with embedded SRAM at advance nodes.

Modern socio-economic development and climate prediction depend greatly on the application of ocean big data. With the accelerated development of ocean observation methods and the continuous improvement of the big data science, the challenges of multiple data sources and data diversity have emerged in the ocean field. As a result, the current data magnitude has reached the terabyte scale. Currently, the traditional theoretical foundation and technical methods have their inherent limitations and demerits that cannot satisfied the temporal and spatial attributes of the current ocean big data. Numerous scholars and countries were involved in ocean big data research. To explore the focus and current status, and determine the topics of research on bursts and acquisition of trend related to ocean big data, 400 articles between 1990 and 2019 were collected from the &ldquo, Web of Science.&rdquo, Combined with visualization software CiteSpace, bibliometrics method and literature combing technology, the pivotal literature related to ocean big data, including significant level countries, institutions, authors, journals and keywords were recognized. A synthetical analysis has revealed research hot spots and research frontiers. The purpose of this study is to provide researchers and practitioners in the field of ocean big data with the main research domains and research hotspots, and orientation for further research.