Search

Synthetic aperture radar (SAR) is an important tool for observing the oceanic internal wave phenomenon. However, owing to the unstable imaging quality of SAR on oceanic internal waves, the texture details of internal wave images are usually unclear, which is not conducive to the subsequent applications of the images. To cope with this problem, a texture enhancement method for oceanic internal wave SAR images is proposed in this paper, which is based on non-local mean (NLM) filtering and texture layer enhancement (TLE). Since the strong speckle noise commonly present in internal wave images is simultaneously enhanced during texture enhancement, resulting in degraded image quality, NLM filtering is first performed to suppress speckle noise. Then, the denoised image is decomposed into the structure layer and the texture layer, and a texture layer enhancement method oriented to the texture characteristics of oceanic internal waves is proposed and applied. Finally, the enhanced texture layer and the structure layer are combined to reconstruct the final enhanced image. Experiments are conducted based on the Gaofen-3 real SAR data, and the results demonstrate that the proposed method performs well in suppressing speckle noise, maintaining overall image brightness, and enhancing internal wave texture details. [ABSTRACT FROM AUTHOR]

While action recognition (AR) has gained large improvements with the introduction of large-scale video datasets and the development of deep neural networks, AR models robust to challenging environments in real-world scenarios are still under-explored. We focus on the task of action recognition in dark environments, which can be applied to fields such as surveillance and autonomous driving at night. Intuitively, current deep networks along with visual enhancement techniques should be able to handle AR in dark environments, however, it is observed that this is not always the case in practice. To dive deeper into exploring solutions for AR in dark environments, we launched the UG 2 + Challenge Track 2 (UG2-2) in IEEE CVPR 2021, with a goal of evaluating and advancing the robustness of AR models in dark environments. The challenge builds and expands on top of a novel ARID dataset, the first dataset for the task of dark video AR, and guides models to tackle such a task in both fully and semi-supervised manners. Baseline results utilizing current AR models and enhancement methods are reported, justifying the challenging nature of this task with substantial room for improvements. Thanks to the active participation from the research community, notable advances have been made in participants' solutions, while analysis of these solutions helped better identify possible directions to tackle the challenge of AR in dark environments. [ABSTRACT FROM AUTHOR]

In recent years, laser scanners integrated with Unmanned Aerial Vehicles (UAVs) have exhibited great potential in conducting power line inspections in harsh environments. The point clouds collected for power line inspections have numerous advantages over remote image data. However, point cloud-based individual power line extraction, which is a crucial technology required for power line inspections, still poses several challenges such as massive 3D points, imbalanced category points, etc. Moreover, in various power line scenarios, previous studies often require manual setup and careful adjustment of different thresholds to separate different power lines, which is inefficient for practical applications. To handle these challenges, in this paper, we propose a multi-branch network to automatically extract an arbitrary number of individual power lines from point clouds collected by UAV-based laser scanners. Specifically, to handle the massive 3D point clouds in complex outdoor scenarios, we propose to leverage deep neural network for efficient and rapid feature extraction in large-scale point clouds. To mitigate imbalanced data quantities across different categories, we propose to design a weighted cross-entropy loss function to measure the varying importance of each category. To achieve the effective extraction of an arbitrary number of power lines, we propose leveraging a loss function to learn the discriminative features that can differentiate the points belonging to different power lines. Once the discriminative features are learned, the Mean Shift method can distinguish the individual power lines by clustering without supervision. The evaluations are executed on two datasets, which are acquired at different locations with UAV-mounted laser scanners. The proposed method has been thoroughly tested and evaluated, and the results and discussions confirm its outstanding ability to extract an arbitrary number of individual power lines in point clouds. [ABSTRACT FROM AUTHOR]

The past few years have witnessed a remarkable advance in deep learning for EEG-based sleep stage classification (SSC). However, the success of these models is attributed to possessing a massive amount of labeled data for training, limiting their applicability in real-world scenarios. In such scenarios, sleep labs can generate a massive amount of data, but labeling can be expensive and time-consuming. Recently, the self-supervised learning (SSL) paradigm has emerged as one of the most successful techniques to overcome labels’ scarcity. In this paper, we evaluate the efficacy of SSL to boost the performance of existing SSC models in the few-labels regime. We conduct a thorough study on three SSC datasets, and we find that fine-tuning the pretrained SSC models with only 5% of labeled data can achieve competitive performance to the supervised training with full labels. Moreover, self-supervised pretraining helps SSC models to be more robust to data imbalance and domain shift problems. [ABSTRACT FROM AUTHOR]

Copyright of Frontiers of Information Technology & Electronic Engineering is the property of Springer Nature and its content may not be copied or emailed to multiple sites or posted to a listserv without the copyright holder's express written permission. However, users may print, download, or email articles for individual use. This abstract may be abridged. No warranty is given about the accuracy of the copy. Users should refer to the original published version of the material for the full abstract. (Copyright applies to all Abstracts.)

Artificial Intelligence (AI) is a fast-growing research and development (R&D) discipline which is attracting increasing attention because it promises to bring vast benefits for consumers and businesses, with considerable benefits promised in productivity growth and innovation. To date, significant accomplishments have been reported in many areas that have been deemed challenging for machines, ranging from computer vision, natural language processing, audio analysis to smart sensing and many others. The technology trend in realizing success has developed towards increasingly complex and large-size AI models to solve more complex problems at superior performance and robustness. This rapid progress, however, has taken place at the expense of substantial environmental costs and resources. In addition, debates on the societal impacts of AI, such as fairness, safety, and privacy, have continued to grow in intensity. These issues have reflected major concerns pertaining to the sustainable development of AI. In this work, major trends in machine learning approaches that can address the sustainability problem of AI have been reviewed. Specifically, the emerging AI methodologies and algorithms are examined for addressing the sustainability issue of AI in two major aspects, i.e., environmental sustainability and social sustainability of AI. Then, the major limitations of the existing studies are highlighted, and potential research challenges and directions are proposed for the development of the next generation of sustainable AI techniques. It is believed that this technical review can help promote a sustainable development of AI R&D activities for the research community. [ABSTRACT FROM AUTHOR]

Today's manufacturing systems are becoming increasingly complex, dynamic and connected hence continual prediction of manufactured product quality is a key to look for patterns that can eventually lead to improved accuracy and productivity. Recent developments in artificial intelligence, especially machine learning have shown great potential to transform the manufacturing domain through analytics for processing vast amounts of manufacturing data generated (Esmaeilian et al. in J Manuf Syst 39:79–100, 2016). Although prediction models have been built to predict product quality with good accuracy, they assume that same distribution applies on training data and testing data hence fail to produce satisfying results when machines work under different conditions with varying data distribution. Naïve re-collection and re-annotation of data for each new working condition can be very expensive thus is not a feasible solution. To cope with this problem, we adopt transfer learning approach called domain adaptation to transfer the knowledge learned from one labelled operating condition (source domain) to another operating condition (target domain) without labels. Particularly, we propose an end-to-end framework for cross-machine product quality prediction, which is able to alleviate domain shift problem. To facilitate the cross-machine prediction performance, a systematic feature selection approach is designed and integrated to generate most suitable feature set to characterize the collected data. Comprehensive experiments have been conducted using actual manufacturing data and the results demonstrate significant improvement on cross-machine product quality prediction as compared to conventional techniques. [ABSTRACT FROM AUTHOR]

Extracting channel state information (CSI) from WiFi signals is of proved high-effectiveness in locating human locations in a device-free manner. However, existing localization/positioning systems are mainly trained and deployed in a fixed environment, and thus they are likely to suffer from substantial performance declines when immigrating to new environments. In this paper, we address the fundamental problem of WiFi-based cross-environment indoor localization using a semi-supervised approach, in which we only have access to the annotations of the source environment while the data in the target environments are un-annotated. This problem is of high practical values in enabling a well-trained system to be scalable to new environments without tedious human annotations. To this end, a deep neural forest is introduced which unifies the ensemble learning with the representation learning functionalities from deep neural networks in an end-to-end trainable fashion. On top of that, an adversarial training strategy is further employed to learn environment-invariant feature representations for facilitating more robust localization. Extensive experiments on real-world datasets demonstrate the superiority of the proposed methods over state-of-the-art baselines. Compared with the best-performing baseline, our model excels with an average 12.7% relative improvement on all six evaluation settings. [ABSTRACT FROM AUTHOR]

Shear stress exerted by the blood stream modulates endothelial functions through altering gene expression. KLF2 and KLF4, the mechanosensitive transcription factors, are promoted by laminar flow to maintain endothelial homeostasis. However, how the expression of KLF2/4 is regulated by shear stress is poorly understood. Here, we showed that the activation of PIEZO1 upregulates the expression of KLF2/4 in endothelial cells. Mice with endothelial-specific deletion of Piezo1 exhibit reduced KLF2/4 expression in thoracic aorta and pulmonary vascular endothelial cells. Mechanistically, shear stress activates PIEZO1, which results in a calcium influx and subsequently activation of CaMKII. CaMKII interacts with and activates MEKK3 to promote MEKK3/MEK5/ERK5 signaling and ultimately induce the transcription of KLF2/4. Our data provide the molecular insight into how endothelial cells sense and convert mechanical stimuli into a biological response to promote KLF2/4 expression for the maintenance of endothelial function and homeostasis. [ABSTRACT FROM AUTHOR]

Adipose-derived stem cells (ASCs) improve the self-renewal and survival of fat grafts in breast reconstruction after oncology surgery. However, ASCs have also been found to enhance breast cancer growth, and its role in tumor proliferation remains largely elusive. Here, we explored a novel mechanism that mediates hTERT reactivation during ASC-induced tumor growth in breast cancer cells. In this study, we found the proliferative ability of breast cancer cells markedly increased with ASC coculture. To explore the molecular mechanism, we treated cells with anibody/inhibitor and found that the activation of MEK-ERK pathway was triggered in breast cancer cells by SCF secreted from ASCs, leading to the nuclear recruitment of CBP. As a coactivator of hTERT, CBP subsequently coordinated with RFPL-3 upregulated hTERT transcription and telomerase activity. The inhibition of CBP and RFPL-3 abrogated the activation of hTERT transcription and the promotion of proliferation in breast cancer cells with cocultured ASCs in vitro and in vivo. Collectively, our study findings indicated that CBP coordination with RFPL-3 promotes ASC-induced breast cancer cell proliferation by anchoring to the hTERT promoter and upregulating telomerase activity, which is activated by the MAPK/ERK pathway. [ABSTRACT FROM AUTHOR]

Machine remaining useful life (RUL) prediction is vital in improving the reliability of industrial systems and reducing maintenance cost. Recently, long short-term memory (LSTM) based algorithms have achieved state-of-the-art performance for RUL prediction due to their strong capability of modeling sequential sensory data. In many cases, the RUL prediction algorithms are required to be deployed on edge devices to support real-time decision making, reduce the data communication cost, and preserve the data privacy. However, the powerful LSTM-based methods which have high complexity cannot be deployed to edge devices with limited computational power and memory. To solve this problem, we propose a knowledge distillation framework, entitled KDnet-RUL, to compress a complex LSTM-based method for RUL prediction. Specifically, it includes a generative adversarial network based knowledge distillation (GAN-KD) for disparate architecture knowledge transfer, a learning-during-teaching based knowledge distillation (LDT-KD) for identical architecture knowledge transfer, and a sequential distillation upon LDT-KD for complicated datasets. We leverage simple and complicated datasets to verify the effectiveness of the proposed KDnet-RUL. The results demonstrate that the proposed method significantly outperforms state-of-the-art KD methods. The compressed model with 12.8 times less weights and 46.2 times less total float point operations even achieves a comparable performance with the complex LSTM model for RUL prediction. [ABSTRACT FROM AUTHOR]

An efficient synthesis of unsymmetrical bisindolylmethanes has been accomplished using triarylmethyl cations to catalyze the reaction of N‐arylimines with two different indoles. Optimization of the organocatalyst by tuning cation stability allows for excellent single addition selectivity when coupled with p‐nitrophenyl imines. The optimal catalyst is commercially available, and the reaction minimizes waste and environmental impact by employing a one‐to‐one ratio of starting materials. The intermediates can be isolated or used in situ in a one‐pot two‐step reaction to generate unsymmetrical bisindolylmethanes in high yields. The reaction tolerates a broad range of imines with the highest yields observed for electron‐poor and neutral imines. A wide range of indole nucleophiles are also successfully employed allowing for the creation of a large variety of unsymmetrical bisindolylmethanes. [ABSTRACT FROM AUTHOR]

In this article, we propose an attention based convolutional neural network long short-term memory (CNN-LSTM) approach for sleep-wake detection with heterogeneous sensor data, i.e., acceleration and heart rate variability (HRV). Since the three-dimensional acceleration data was sampled with a high frequency, we firstly design a CNN-LSTM structure to effectively learn latent features from the acceleration. Meanwhile, considering the unique format of the HRV data, some effective features are extracted based on domain knowledge. Next, we design a unified architecture to efficiently merge the features learned by CNN-LSTM approach from the acceleration and the extracted features from the HRV, which enables us to make full use of all the available information from these two heterogeneous sources. Taking into consideration that these two heterogeneous sources may have distinct contributions for the sleep and wake states, we propose an attention network to dynamically adjust the importance of features from the two sources. Real-world experiments have been conducted to verify the effectiveness of the proposed approach for sleep-wake detection. The results demonstrate that the proposed method outperforms all existing approaches for sleep-wake classification. In the evaluation of leave-one-subject-out (LOSO) cross-validation which is more challenging and practical, the proposed method achieves remarkable improvements ranging from 5% to 46% over the benchmark approaches. [ABSTRACT FROM AUTHOR]

Enabling precise forecasting of the remaining useful life (RUL) for machines can reduce maintenance cost, increase availability, and prevent catastrophic consequences. Data-driven RUL prediction methods have already achieved acclaimed performance. However, they usually assume that the training and testing data are collected from the same condition (same distribution or domain), which is generally not valid in real industry. Conventional approaches to address domain shift problems attempt to derive domain-invariant features, but fail to consider target-specific information, leading to limited performance. To tackle this issue, in this article, we propose a contrastive adversarial domain adaptation (CADA) method for cross-domain RUL prediction. The proposed CADA approach is built upon an adversarial domain adaptation architecture with a contrastive loss, such that it is able to take target-specific information into consideration when learning domain-invariant features. To validate the superiority of the proposed approach, comprehensive experiments have been conducted to predict the RULs of aeroengines across 12 cross-domain scenarios. The experimental results show that the proposed method significantly outperforms state-of-the-arts with over 21% and 38% improvements in terms of two different evaluation metrics. [ABSTRACT FROM AUTHOR]

The remaining useful life (RUL) prediction plays a pivotal role in the predictive maintenance of industrial manufacturing systems. However, one major problem with the existing RUL estimation algorithms is the assumption of a single health degradation trend for different machine health stages. To improve the RUL prediction accuracy with various degradation trends, this article proposes an algorithm dubbed degradation-aware long short-term memory (LSTM) autoencoder (AE) (DELTA). First, the Hilbert transform is adopted to evaluate the degradation stage and factor with the real-time sensory signal. Second, we adopt LSTM AE to predict RUL based on multisensor time-series data and the degradation factor. Distinct from the existing studies, the proposed framework is able to dynamically model the degradation factor and explore latent variables to improve RUL prediction accuracy. The performance of DELTA is evaluated with the open-source FEMTO bearing data set. Compared with the existing algorithms, DELTA achieves appreciable improvements in the RUL prediction accuracy. [ABSTRACT FROM AUTHOR]

Automatic sleep stage mymargin classification is of great importance to measure sleep quality. In this paper, we propose a novel attention-based deep learning architecture called AttnSleep to classify sleep stages using single channel EEG signals. This architecture starts with the feature extraction module based on multi-resolution convolutional neural network (MRCNN) and adaptive feature recalibration (AFR). The MRCNN can extract low and high frequency features and the AFR is able to improve the quality of the extracted features by modeling the inter-dependencies between the features. The second module is the temporal context encoder (TCE) that leverages a multi-head attention mechanism to capture the temporal dependencies among the extracted features. Particularly, the multi-head attention deploys causal convolutions to model the temporal relations in the input features. We evaluate the performance of our proposed AttnSleep model using three public datasets. The results show that our AttnSleep outperforms state-of-the-art techniques in terms of different evaluation metrics. Our source codes, experimental data, and supplementary materials are available at https://github.com/emadeldeen24/AttnSleep. [ABSTRACT FROM AUTHOR]

The seasonal and interannual variabilities of sea surface wind (SSW) in the South China Sea (SCS), especially in coral reef regions such as Nansha Islands, Xisha Islands, Zhongsha Islands and Dongsha Islands were investigated in detail using the Blended Sea Winds dataset (1988–2017). Annual and monthly variations of SSW and sea surface temperature (SST) in the four zones were investigated. Empirical Orthogonal Function (EOF) analysis of wind field was performed to aid in better understanding the different spatial patterns. The results indicate that, as observed in the spatial distribution of the first mode of monthly mean wind speed anomaly, the magnitudes in the four island zones are all negative and are similar to each other, showing that the variations of SSW in the four island zones are consistent. In the second mode, the magnitudes in Nansha Islands are opposite to those in the other three zones. The spatial distribution of the third mode reflects regional differences. The maximum annual SSW appears in Dongsha Islands, and the minimum appears in Nansha Islands. The interannual variations of SSW in all island zones are basically concurrent. The island zones with high SSW mostly have low SST, and vice versa. There may be an inverse relationship between SSW and SST in coral reef regions in the SCS. The multi-year monthly variations of SSW in the island zones present a 'W'-shaped structural variation. Each island undergoes two months of minimum SSW every year, one during March–May (MAM) and the other during September–November (SON). Both months are in monsoon transition periods. During the months with low SSW, high SST appears. The SST peaks almost correspond to the SSW troughs. This further indicates that SSW and SST may have opposite changes in coral reef regions. Coral bleaching events often correspond to years of high SST and low SSW. [ABSTRACT FROM AUTHOR]

For prognostics and health management of mechanical systems, a core task is to predict the machine remaining useful life (RUL). Currently, deep structures with automatic feature learning, such as long short-term memory (LSTM), have achieved great performances for the RUL prediction. However, the conventional LSTM network only uses the learned features at last time step for regression or classification, which is not efficient. Besides, some handcrafted features with domain knowledge may convey additional information for the prediction of RUL. It is thus highly motivated to integrate both those handcrafted features and automatically learned features for the RUL prediction. In this article, we propose an attention-based deep learning framework for machine's RUL prediction. The LSTM network is employed to learn sequential features from raw sensory data. Meanwhile, the proposed attention mechanism is able to learn the importance of features and time steps, and assign larger weights to more important ones. Moreover, a feature fusion framework is developed to combine the handcrafted features with automatically learned features to boost the performance of the RUL prediction. Extensive experiments have been conducted on two real datasets and experimental results demonstrate that our proposed approach outperforms the state-of-the-arts. [ABSTRACT FROM AUTHOR]

Data-driven fault classification methods are receiving great attention as they can be applied to many real-world applications. However, they work under the assumption that training data and testing data are drawn from the same distribution. Practical scenarios have varying operating conditions, which results in a domain-shift problem that significantly deteriorates the diagnosis performance. Recently, domain adaptation (DA) has been explored to address the domain-shift problem by transferring the knowledge from labeled source domain (e.g., source working condition) to unlabeled target domain (e.g., target working condition). Yet, all the existing methods are working under single-source single-target (1S1T) settings. Hence, a new model needs to be trained for each new target domain. This shows limited scalability in handling multiple working conditions since different models should be trained for different target working conditions, which is clearly not a viable solution in practice. To address this problem, we propose a novel adversarial multiple-target DA (AMDA) method for single-source multiple-target (1SmT) scenario, where the model can generalize to multiple-target domains concurrently. Adversarial adaptation is applied to transform the multiple-target domain features to be invariant from the single-source-domain features. This leads to a scalable model with a novel capability of generalizing to multiple-target domains. Extensive experiments on two public datasets and one self-collected dataset have demonstrated that the proposed method outperforms state-of-the-art methods consistently. Our source codes and data are available at https://github.com/mohamedr002/AMDA. [ABSTRACT FROM AUTHOR]

Nowadays, deep learning-based models have been widely developed for atrial fibrillation (AF) detection in electrocardiogram (ECG) signals. However, owing to the inevitable over-fitting problem, classification accuracy of the developed models severely differed when applying on the independent test datasets. This situation is more significant for AF detection from dynamic ECGs. In this study, we explored two potential training strategies to address the over-fitting problem in AF detection. The first one is to use the Fast Fourier transform (FFT) and Hanning-window-based filter to suppress the influence from individual difference. Another is to train the model on the wearable ECG data to improve the robustness of model. Wearable ECG data from 29 patients with arrhythmia were collected for at least 24 h. To verify the effectiveness of the training strategies, a Long Short-Term Memory (LSTM) and Convolution Neural Network (CNN)-based model was proposed and tested. We tested the model on the independent wearable ECG data set, as well as the MIT-BIH Atrial Fibrillation database and PhysioNet/Computing in Cardiology Challenge 2017 database. The model achieved 96.23%, 95.44%, and 95.28% accuracy rates on the three databases, respectively. Pertaining to the comparison of the accuracy rates on each training set, the accuracy of the model trained in conjunction with the proposed training strategies only reduced by 2%, while the accuracy of the model trained without the training strategies decreased by approximately 15%. Therefore, the proposed training strategies serve as effective mechanisms for devising a robust AF detector and significantly enhanced the detection accuracy rates of the resulting deep networks. [ABSTRACT FROM AUTHOR]

On the basis of ecological footprint theory and tourism ecological footprint theory, the sustainable development indexes such as ecological footprint, ecological carrying capacity, ecological deficit, and ecological surplus of the research area were calculated and the long-term change pattern of each index was analyzed. This paper shows that the ecological footprint of the research area increases year by year, but the ecological footprint is always smaller than the ecological carrying capacity, indicating that the area is still in the state of sustainable development. However, the per capita ecological surplus shows a decreasing trend year by year, indicating that the sustainable development of the region is getting worse. This paper proposes a reordering method of tourist attractions based on heterogeneous information fusion, and realizes the retrieval and reordering of tourist attractions based on user query and fusion of heterogeneous information, so as to help users make travel decisions. In view of the shortage of tourism commercial websites to passively provide scenic spot information, this paper puts forward a scenic spot retrieval method based on query words to enable users to obtain scenic spot information according to their needs, and constructs a tourist consumer data analysis system. The preprocessing methods and methods adopted by the data preprocessing module are analyzed in detail, and the algorithms used in the travel route analysis and consumer spending ability analysis are described in detail. The data of tourism consumers are analyzed by this system, and the results are evaluated. [ABSTRACT FROM AUTHOR]

Human activity recognition (HAR) using smartphone sensors has attracted great attention due to its wide range of applications. A standard solution for HAR is to first generate some features defined based on domain knowledge (handcrafted features) and then to train an activity classification model based on these features. Very recently, deep learning with automatic feature learning from raw sensory data has also achieved great performance for HAR task. We believe that both the handcrafted features and the learned features may convey some unique information that can complement each other for HAR. In this article, we first propose a feature fusion framework to combine handcrafted features with automatically learned features by a deep algorithm for HAR. Then, taking the regular dynamics of human behavior into consideration, we develop a maximum full a posteriori algorithm to further enhance the performance of HAR. Our extensive experimental results show the proposed approach can achieve superior performance comparing with the state-of-the-art methodologies across both a public data set and a self-collected data set. [ABSTRACT FROM AUTHOR]

Building occupancy estimation holds great promise for building control systems to save energy and provide a comfortable indoor environment. Existing solutions turn out to be lacking in practice due to their specific hardware requirements and/or poor performances. Recently, an light-emitting diode (LED) light sensor based occupancy estimation system, which is nonintrusive and does not require any additional hardware, has been proposed. However, the performance of the system is limited, especially in a complicated dynamic scenario. In this article, a Bayes filter with neural networks is proposed for the optimal estimation of occupancy based on light sensor data. Specifically, based on the formulation of Bayes filter, the posterior probability of the building occupancy can be decoupled into three components: The prior, likelihood, and evidence. The prior and likelihood are, respectively, estimated from a Markov model and an efficient single-hidden layer feedforward neural network (SLFN). Finally, the evidence can be obtained by the results of prior and likelihood. Real experiments have been conducted to verify the effectiveness of the proposed approach in two complicated scenarios, i.e., dynamic and regular. Results indicate that the proposed Bayes filter outperforms all the benchmark approaches. The impacts of the number of LED sensing units and the number of hidden layers for neural networks are also evaluated. The results manifest that the number of sensing units should be chosen based on the required performance and the SLFN is sufficient for this application. [ABSTRACT FROM AUTHOR]

Indoor localization has attracted more and more attention because of its importance in many applications. One of the most popular techniques for indoor localization is the received signal strength indicator (RSSI) based fingerprinting approach. Since RSSI values are very complicated and noisy, conventional machine learning algorithms often suffer from limited performance. Recently developed deep learning algorithms have been shown to be powerful for the analysis of complex data. In this paper, we propose a local feature-based deep long short-term memory (LF-DLSTM) approach for WiFi fingerprinting indoor localization. The local feature extractor attempts to reduce the noise effect and extract robust local features. The DLSTM network is able to encode temporal dependencies and learn high-level representations for the extracted sequential local features. Real experiments have been conducted in two different environments, i.e., a research lab and an office. We also compare the proposed approach with some state-of-the-art methods for indoor localization. The results show that the proposed approach achieves the best localization performance with mean localization errors of 1.48 and 1.75 m under the research lab and office environments, respectively. The improvements of our proposed approach over the state-of-the-art methods range from $\text{18.98}{\%}$ to $\text{53.46}{\%}$. [ABSTRACT FROM AUTHOR]

Reducing traffic delay is of crucial importance for the development of sustainable transportation systems, which is a challenging task in the studies of stochastic shortest path (SSP) problem. Existing methods based on the probability tail model to solve the SSP problem, seek for the path that minimizes the probability of delay occurrence, which is equal to maximizing the probability of reaching the destination before a deadline (i.e., arriving on time). However, they suffer from low accuracy or high computational cost. Therefore, we design a novel and practical Q-learning approach where the converged Q-values have the practical meaning as the actual probabilities of arriving on time so as to improve the accuracy of finding the real optimal path. By further adopting dynamic neural networks to learn the value function, our approach can scale well to large road networks with arbitrary deadlines. Moreover, our approach is flexible to implement in a time dependent manner, which further improves the performance of returned path. Experimental results on some road networks with real mobility data, such as Beijing, Munich and Singapore, demonstrate the significant advantages of the proposed approach over other methods. [ABSTRACT FROM AUTHOR]

Human activity recognition can benefit various applications including healthcare services and context awareness. Since human actions will influence WiFi signals, which can be captured by the channel state information (CSI) of WiFi, WiFi CSI based human activity recognition has gained more and more attention. Due to the complex relationship between human activities and WiFi CSI measurements, the accuracies of current recognition systems are far from satisfactory. In this paper, we propose a new deep learning based approach, i.e., attention based bi-directional long short-term memory (ABLSTM), for passive human activity recognition using WiFi CSI signals. The BLSTM is employed to learn representative features in two directions from raw sequential CSI measurements. Since the learned features may have different contributions for final activity recognition, we leverage on an attention mechanism to assign different weights for all the learned features. Real experiments have been carried out to evaluate the performance of the proposed ABLSTM for human activity recognition. The experimental results show that our proposed ABLSTM is able to achieve the best recognition performance for all activities when compared with some benchmark approaches. [ABSTRACT FROM AUTHOR]

Human activity recognition (HAR) based on inertial sensors has been investigated for many industrial informatics applications, such as healthcare and ubiquitous computing. Existing methods mainly rely on supervised learning schemes, which require large labeled training data. However, labeled data are sometimes difficult to acquire, while unlabeled data are readily available. Thus, we intend to make use of both labeled and unlabeled data with semisupervised learning for accurate HAR. In this paper, we propose a semisupervised deep learning approach, using temporal ensembling of deep long short-term memory, to recognize human activities with smartphone inertial sensors. With the deep neural network processing, features are extracted for local dependencies in the recurrent framework. Besides, with an ensemble approach based on both labeled and unlabeled data, we can combine together the supervised and unsupervised losses, so as to make good use of unlabeled data that the supervised learning method cannot leverage. Experimental results indicate the effectiveness of our proposed semisupervised learning scheme, when compared to several state-of-the-art semisupervised learning approaches. [ABSTRACT FROM AUTHOR]

This paper discusses greedy methods for sensor placement in linear inverse problems. We comprehensively review the greedy methods in the sense of optimizing the mean squared error (MSE), the volume of the confidence ellipsoid, and the worst-case error variance. We show that the greedy method of optimizing an MSE related cost function can find a near-optimal solution. We then provide a new fast algorithm to optimize the MSE. In greedy methods, we select the sensing location one by one. In this way, the searching space is greatly reduced but many valid solutions are ignored. To further improve the current greedy methods, we propose a group-greedy strategy, which can be applied to optimize all the three criteria. In each step, we reserve a group of suboptimal sensor configurations which are used to generate the potential sensor configurations of the next step and the best one is used to check the terminal condition. Compared with the current greedy methods, the group-greedy strategy increases the searching space but greatly improve the solution performance. We find the necessary and sufficient conditions that the current greedy methods and the proposed group greedy method can obtain the optimal solution. The illustrative examples show that the group greedy method outperforms the corresponding greedy method. We also provide a practical way to find a proper group size with which the proposed group greedy method can find a solution that has almost the same performance as the optimal solution. [ABSTRACT FROM AUTHOR]

Human activity recognition is a core problem in intelligent automation systems due to its far-reaching applications including ubiquitous computing, health-care services, and smart living. Due to the nonintrusive property of smartphones, smartphone sensors are widely used for the identification of human activities. However, unlike applications in vision or data mining domain, feature embedding from deep neural networks performs much worse in terms of recognition accuracy than properly designed handcrafted features. In this paper, we posit that feature embedding from deep neural networks may convey complementary information and propose a novel knowledge distilling strategy to improve its performance. More specifically, an efficient shallow network, i.e., single-layer feedforward neural network (SLFN), with handcrafted features is utilized to assist a deep long short-term memory (LSTM) network. On the one hand, the deep LSTM network is able to learn features from raw sensory data to encode temporal dependencies. On the other hand, the deep LSTM network can also learn from SLFN to mimic how it generalizes. Experimental results demonstrate the superiority of the proposed method in terms of recognition accuracy against several state-of-the-art methods in the literature. [ABSTRACT FROM AUTHOR]

The present study aimed to assess the concordance of 24-h intraocular pressure (IOP) curves between glaucomatous and contralateral eyes for patients with untreated unilateral primary open-angle glaucoma (POAG). A total of 32 patients with unilateral POAG and 32 age-matched normal subjects were enrolled. The IOP measurements were performed every 2 h over a 24-h period. The concordance of the 24-h IOP curves was assessed via the correlation coefficient (r), intraclass correlation coefficient (ICC) and repeated-measures analysis of variance (ANOVA). No significant difference was identified between all IOPs, as well as the mean, peak and trough IOP or IOP fluctuations of the paired eyes in the two groups. The strength of association of all IOPs was moderate in the glaucoma group (r, 0.752-0.867) and the normal controls (r, 0.625-0.873). IOP readings at each time-point indicated a high agreement in the glaucoma group (ICC, 0.857-0.929) and the normal controls (ICC, 0.768-0.932). Repeated-measures ANOVA indicated that the 24-h IOP curves of the paired eyes had parallel profiles in the two study groups (P=0.837 and P=0.897, respectively). The glaucoma patients had significantly higher proportions of all IOPs displaying absolute differences of ≥2 and ≥3 mmHg (46.09 vs. 35.68%, P<0.001; 29.69 vs. 12.50%, P<0.001, respectively). In conclusion, the 24-h IOP curves of the paired eyes had parallel profiles in unilateral glaucoma patients and normal subjects. However, unilateral glaucoma patients had a significantly larger proportion of IOP differences of ≥2 and ≥3 mmHg. [ABSTRACT FROM AUTHOR]

Buildings consume quite a lot of energy; hence, the issue of building energy efficiency has attracted a great deal of attention in recent years. A key factor in achieving this objective is occupancy information that directly impacts on energy-related building control systems. In this paper, we leverage on environmental sensors that are nonintrusive and cost-effective for building occupancy estimation. Our result relies on feature engineering and learning. The conventional feature engineering requires one to manually extract relevant features without a clear guideline. This blind feature extraction is labor intensive and may miss some significant implicit features. To address this issue, we propose a convolutional deep bidirectional long short-term memory (CDBLSTM) approach that contains a convolutional network and a deep structure to automatically learn significant features from the sensory data without human intervention. Moreover, the long short-term memory networks are able to capture temporal dependencies in the data and the bidirectional structure can take the past and future contexts into consideration for the final identification of occupancy. We have conducted real experiments to evaluate the performance of our proposed CDBLSTM approach. Instead of estimating the exact number of occupants, we attempt to identify the range of occupants, i.e., zero, low, medium, and high, which is adequate for most of building control systems. The experimental results indicate the effectiveness of our proposed approach compared with the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Human activity recognition using either wearable devices or smartphones can benefit various applications including healthcare, fitness, smart home, etc. Instead of using wearable devices which are intrusive and require extra cost, we shall leverage on modern smartphones embedded with a variety of sensors. Due to the flexibility of using smartphones, the recognition accuracy will degrade with orientation, placement, and subject variations. In this paper, we propose a robust human activity recognition system in terms of orientation, placement, and subject variations based on coordinate transformation and principal component analysis (CT-PCA) and online support vector machine (OSVM). The proposed CT-PCA scheme is utilized to eliminate the effect of orientation variations. Experiments show that the proposed scheme significantly improves the activity recognition accuracy and outperforms the state-of-the-art methods on leave one orientation out experiments, which demonstrates the generalization ability of the proposed scheme on the data from unseen orientations. We also show the effectiveness of this scheme on placement and subject variations. However, the inherent difference of signal properties for different placement and subject dramatically reduces the recognition accuracy, especially for different placement. Thus, we present an efficient OSVM algorithm, that is, online-independent support vector machine (OISVM), which utilizes a small portion of data from the unseen placement or subject to online update the parameters of the SVM algorithm. The experimental results demonstrate the effectiveness of this OISVM algorithm on placement and subject variations. [ABSTRACT FROM PUBLISHER]

Occupancy estimation in buildings can benefit various applications such as heating, ventilation, and air-conditioning control, space monitoring, and emergency evacuation. Due to the consideration of temporal dependency in occupancy data, hidden Markov model (HMM) has been shown to be effective in occupancy estimation. However, the conventional HMM that assumes invariant temporal dependency of occupancy dynamics for different time instances is unrealistic. Moreover, the performance of the conventional HMM that utilizes mixture of Gaussian for emission probability in terms of continuous observations can be easily affected by the noise in sensory data. To address these problems, in this paper, we propose a new architecture, i.e., inhomogeneous hidden Markov model with multinomial logistic regression (IHMM-MLR), for building occupancy estimation using nonintrusive environmental sensors. Instead of using the time-invariant transition probability matrix, we apply a time-dependent (inhomogeneous) transition probability matrix which can capture the temporal dependency for different time instances. Meanwhile, we employ an efficient probabilistic model, i.e., MLR, for emission probability. Online and offline occupancy estimation schemes are presented for real-time and accurate long-term applications respectively. Real experiments have indicated the effectiveness of our proposed approach. [ABSTRACT FROM PUBLISHER]

Occupancy information can help us to achieve high energy-efficient buildings. Previous works mainly focus on predicting the presence and absence of occupants in homes or single person offices. We attempt to predict regular occupancy level in a commercial building deployment scenario. The occupancy prediction models can be divided into two categories of occupancy models and data mining approaches. For the occupancy models, we shall investigate the efficiencies of two widely used multi-occupant models, that is, inhomogeneous Markov chain and multivariate Gaussian. For the data mining approaches, we propose the application of autoregressive integrated moving average, artificial neural network and support vector regression. Experiments have been conducted using actual occupancy data under four different prediction horizons, that is, 15 min, 30 min, 1 and 2 h. The results demonstrated a guideline in how to choose a proper method for the prediction of occupancy in commercial buildings under different prediction horizons. [ABSTRACT FROM AUTHOR]

High-temperature superconducting (HTS) dc cables become viable for future dc electric power transmissions due to their low loss, light weight, high current density, and high transmission capacity. According to the structure, HTS dc cables are commonly divided into unipolar, bipolar coaxial and multicore types. The HTS critical current is magnetic field dependent, while different HTS dc cable structures have different field distribution features. The influence of different HTS dc cable structures on the critical current is analyzed and compared, in order to provide a basis for the design of a practical HTS dc cable. [ABSTRACT FROM AUTHOR]

The Global Positioning System (GPS) can be readily used for outdoor localization, but GPS signals are degraded in indoor environments. How to develop a robust and accurate indoor localization system is an emergent task. In this paper, we propose a smartphone inertial sensor-based indoor localization and tracking system with occasional iBeacon corrections. Some important issues in a smartphone-based pedestrian dead reckoning (PDR) approach, i.e., step detection, walking direction estimation, and initial point estimation, are studied. One problem of the PDR approach is the drift with walking distance. We apply a recent technology, iBeacon, to occasionally calibrate the drift of the PDR approach. By analyzing iBeacon measurements, we define an efficient calibration range where an extended Kalman filter is utilized. The proposed localization and tracking system can be implemented in resource-limited smartphones. To evaluate the performance of the proposed approach, real experiments under two different environments have been conducted. The experimental results demonstrated the effectiveness of the proposed approach. We also tested the localization accuracy with respect to the number of iBeacons. [ABSTRACT FROM PUBLISHER]