One of the main challenges in developing embedded radar-based gesture recognition systems is the requirement of energy efficiency. To facilitate this, we present an embedded gesture recognition system using a 60 GHz frequency modulated continuous wave radar using spiking neural networks (SNNs) applied directly to raw ADC data. The SNNs are sparse in time and space, and event-driven which makes them energy-efficient. In contrast to the previous state-of-the-art methods, the proposed system is only based on the raw ADC data of the target thus avoiding the overhead of performing the slow-time and fast-time Fourier transforms (FFTs). Furthermore, the pre-processing slow-time FFT is mimicked in the proposed SNN architecture, where the proposed model processing speed of 12 ms advances the state-of-the-art by a factor of 17.7. The experimental results demonstrate that despite the simplification the proposed implementation achieves recognition accuracy of 98.1 %, which is comparable to the conventional approaches.

Worldwide disasters like bushfires, earthquakes, floods, cyclones, heatwaves etc. have affected the lives of social media users in an unprecedented manner. They are constantly posting their level of negativity over the disaster situations at their location of interest. Understanding location-oriented sentiments about disaster situation is of prime importance for political leaders, and strategic decision-makers. To this end, we present a new fully automated algorithm based on artificial intelligence (AI) and Natural Language Processing (NLP), for extraction of location-oriented public sentiments on global disaster situation. We designed the proposed system to obtain exhaustive knowledge and insights on social media feeds related to disaster in 110 languages through AI and NLP based sentiment analysis, named entity recognition (NER), anomaly detection, regression, and Getis Ord Gi* algorithms. We deployed and tested this algorithm on live Twitter feeds from 28 September 2021 till 6 October 2021. Tweets with 67,515 entities in 39 different languages were processed during this period. Our novel algorithm extracted 9727 location entities with greater than 70% confidence from live twitter feed and displayed the locations of possible disasters with disaster intelligence. The rates of average precision, recall and F1-Score were measured to be 0.93, 0.88 and 0.90 respectively. Overall, the fully automated disaster monitoring solution demonstrated 97% accuracy. According to the best of our knowledge, this study is the first to report location intelligence with NER, sentiment analysis, regression and anomaly detection on social media messages related to disasters and has covered the largest set of languages.

In this paper, we propose a novel neural network for Compressed Sensing (CS) application: the Dictionary Learning based Compressed Sensing neural Network (DL-CSNet). It is fairly simple but highly effective, which consists of only three layers: 1) a DL layer for latent sparse features extraction; 2) a smoothing layer via Total Variation (TV) like constraint; and 3) a CS acquisition layer for neural network training. In particular, the TV-like smoothing layer is a perfect complement to the sparsity-oriented DL layer to achieve smooth images. The trained DL-CSNet can learn the optimal dictionary matrix so that images can be reconstructed in high quality. At last, extensive experiments have been carried out on binary images and compared to most classical CS algorithms, which shows the superior performance of the proposed DL-CSNet.

In recent years, Machine Learning (ML) algorithms have become widespread in all the fields of Remote Sensing (RS) and Earth Observation (EO). This has allowed the rapid development of new procedures to solve problems affecting these sectors. In this context, this work aims at presenting a novel method for filtering speckle noise from Sentinel-1 Ground Range Detected (GRD) data by applying Deep Learning (DL) algorithms, based on Convolutional Neural Networks (CNNs). The paper provides an easy yet very effective approach to extract the large amount of training data needed for DL approaches in this challenging case. The experimental results on simulated speckled images and an actual SAR dataset show a clear improvement with respect to the state of the art in terms of Peak Signal-to-Noise Ratio (PSNR), Structural Similarity Index (SSIM), Equivalent Number of Looks (ENL), proving the effectiveness of the proposed architecture.

WiFi-based human activity recognition (HAR) has been extensively studied due to its far-reaching applications in health domains, including elderly monitoring, exercise supervision, rehabilitation monitoring, etc. Although existing supervised deep learning techniques have achieved remarkable performances for these tasks, they are however data-hungry and hence are notoriously difficult due to the privacy and incomprehensibility of WiFi-based HAR data. Existing contrastive learning models, which are mainly designed for computer vision, cannot guarantee their performance on channel state information (CSI) data. To this end, we propose a new dual-stream contrastive learning model, that can process and learn the raw WiFi CSI data in a self-supervised manner. More specifically, our proposed method, coined as DualConFi, takes raw WiFI CSI data as input and consists of a channel stream and a temporal stream to learn highly-discriminative spatiotemporal features under mutual information constraint using unlabeled data. We demonstrate the efficacy of our model on three publicly available CSI data sets in various experiment settings, including linear evaluation, semi-supervised, and transfer learning. We show that DualConFi is able to perform favorably against challenging baselines in each setting. Moreover, by studying the effects of different transform functions on CSI data, we finally verify the effectiveness of highly-discriminative features.

Parkinson’s disease (PD) is a common and complex neurodegenerative disorder with 5 stages in the Hoehn and Yahr scaling. Given the heterogeneity of PD, it is challenging to classify early stages 1 and 2 and detect brain function alterations. Functional magnetic resonance imaging (fMRI) is a promising tool in revealing functional connectivity (FC) differences and developing biomarkers in PD. Some machine learning approaches like support vector machine and logistic regression have been successfully applied in the early diagnosis of PD using fMRI data, which outperform classifiers based on manually selected morphological features. However, the early-stage characterization in FC changes has not been fully investigated. Given the complexity and non-linearity of fMRI data, we propose the use of a long short-term memory (LSTM) network to characterize the early stages of PD. The study included 84 subjects (56 in stage 2 and 28 in stage 1) from the Parkinson’s Progression Markers Initiative (PPMI), the largest-available public PD dataset. Under a repeated 10-fold stratified cross-validation, the LSTM model reached an accuracy of 71.63%, 13.52% higher than the best traditional machine learning method, indicating significantly better robustness and accuracy compared with other machine learning classifiers. We used the learned LSTM model weights to select the top brain regions that contributed to model prediction and performed FC analyses to characterize functional changes with disease stage and motor impairment to gain better insight into the brain mechanisms of PD.

This paper provides an in-depth review about opinions and applications in the new area of Human Centered Artificial Intelligence

In this paper, we compare various deep neural models used to identify optical noise.

Click-through rate (CTR) prediction is task to estimate the possibility of user clicking on a recommended item. The ground truth labels of this task are the click history of users. As there are many reasons why noise data or low quality data may be generated, the training data has a certain amount of noise data or low quality data. In this work, We propose a simple but effective method to find the noise data. Our method could improve the offline AUC from 0.60 to 0.75 on our real-world test dataset of the un-drop part. We will verify the AUC on all the test data (droped and un-drop).

With the rapid development of Internet of Things (IoT) technologies, traditional disease diagnosis carried out in medical institutions can now be performed remotely at home or even ambient environments, yielding the concept of Internet of Health Things (IoHT). Among the diverse IoHT applications, inertial measurement units (IMUs) based systems play a significant role in the detection of diseases in many fields, such as neurological, musculoskeletal, and mental. However, traditional numerical interpretation methods have proven to be challenging to provide satisfying detection accuracies owing to the low quality of raw data especially under strong electromagnetic interference (EMI). To address this issue, in recent years, machine learning (ML) based techniques have been proposed to smartly map IMU captured data on disease detection and progress. After a decade of development, the combination of IMUs and ML algorithms for assistive disease diagnosis has become a hot topic, with an increasing number of studies reported yearly. To help readers comprehensively understand the fundamentals and state-of-the-art techniques, this article systematically reviews this field, by introducing and explaining relevant application scenarios, discussing challenges, and predicting foreseeable future trends.

Power distribution networks are experiencing large deployment of behind-the-meter distributed generation and storage along with electrified transport and heat, requiring fundamental changes in planning and operation. Synthetic representative networks may not be sufficient to fully address the needs of researchers, operators and decision-makers. There are increasing efforts to directly model large-scale real networks and conduct detailed analysis beyond conventional practices. This paper presents a large-scale medium voltage (MV)-low voltage (LV) integrated network model building, scenario determination and analysis approach for distribution networks. Data with different formats from several databases were used in model building. The model is validated using benchmark scenarios based on historical network measurements from the field. In collaboration with the national distribution system operator in Ireland (ESB Networks), a pilot study was conducted for a rural network in the Southwest of Ireland, highlighting possible challenges and offering suitable solutions for country-scale implementation.

This manuscript was submitted to IEEE Power Engineering Letters. We propose an extension of the Dissipating Energy Flow (DEF) method for locating sources of forced oscillations (FOs), based on an energy function constructed by the complex integral approach, which we call Complex DEF (CDEF). To trace sources of FO, we study the scalar projection of CDEF in a specific direction on the complex plane. We show by simulated cases that this alternative has a good performance in cases where the original DEF fails.

Smart contract is an important building block of blockchain. Automated market makers are working without an order book, and they determine the price of assets automatically. It is reported that he automated market makers have the impermanent loss, which causes financial damage to liquidity providers. Impermanent loss makes the liquidity providers hesitant to deposit assets in the liquidity pool. Therefore, their participation incentive from liquidity provision should be anticipated by automatic market makers inherently. However, the existence of impermanent gain has never been reported. Impermanent gain is important to attract liquidity providers without giving compensation incentives. This study shows that for some automated market makers, impermanent gain coexists with impermanent loss. Examples showing the coexistence and conditions are provided.

Neurological and orthopedic chronic diseases are proven to numerous complications and related dysfunction for patients, and are leading factors of death and disability. Conventional methods of diagnosis and rehabilitation are conducted at hospitals. However, due to the aging society and controversial privacy protection, traditional methods implemented in medical institutions have faced numerous challenges. Hence, internet of healthcare things (IoHT) based techniques have recently begun to be implemented. In order to provide flexible utilization and secure privacy for patients, wearable devices, like whole body devices based on electromyography (EMG) and radio frequency identification (RFID) insole systems, based on gait-analysis were applied. Among these techniques, insole systems for medical analysis are more suitable, because the root cause of neurologic and orthopedic chronic diseases are directly related to several gait characteristics. Neurologic and orthopedic chronic diseases are caused by lesions of the central nervous system (CNS), peripheral nerves (PN) and orthopedic limbs, similarly gait features are comprehensive results generated by the coordination of CNS, PN and lower limbs However, owing to lack of standards for instrumentation and the knowledge gap between engineering and pathogenic mechanism, the medical application of insole systems and the development of IoHT are hindered. This article is presented to address these issues. In this context, five plantar sensing mechanisms and their related products will be overviewed. Next, pathogenic mechanisms and pathological gait features of six representative chronic neurologic diseases will be displayed. Likewise, the development of corresponding insole systems for supervision and recovery for these diseases will be reviewed as well. Finally, challenges and future trends for current work will be discussed. This article will not only provide guidelines for the design of insole systems targeted for specific diseases but will also inspire the innovation of novel ideas.

The sixth-generation (6G) wireless communication system is expected to integrate communication, intelligence, sensing, positioning, control, and calculation to adapt to time-critical, ultra-reliable, and energy-saving data delivery, as well as accurate positioning of personnel and equipment, serving the Internet of Things (IoT). On one hand, reconfigurable intelligent surface (RIS) can intelligently manipulate radio waves and is considered to be one of the candidate technologies for the 6G wireless communication. Hence, there are more and more surveys on RIS-assisted communications. On the other hand, the potential of RIS in positioning has attracted growing attention, and articles on RIS-assisted positioning have been blown out. Therefore, it is time to review this literature to understand the potential of RIS positioning, research status, and point out the direction for future research. This paper first explains the working principle and channel model of RIS and summarizes some characteristics of RIS suitable for positioning. Then we give a concise review and classification of existing RIS positioning research. Finally, we put forward our views on the future research challenges and attractive directions for RIS-aided wireless positioning technology.

Inferring music time structures has a broad range of applications in music production, processing and analysis. Scholars have proposed various methods to analyze different aspects of time structures, such as beat, downbeat, tempo and meter. Many state-of-the-art (SOFA) methods, however, are computationally expensive. This makes them inapplicable in real-world industrial settings where the scale of the music collections can be millions. This paper proposes a new state space and a semi-Markov model for music time structure analysis. The proposed approach turns the commonly used 2D state spaces into a 1D model through a jump-back reward strategy. It reduces the state spaces size drastically. We then utilize the proposed method for causal, joint beat, downbeat, tempo, and meter tracking, and compare it against several previous methods. The proposed method delivers similar performance with the SOFA joint causal models with a much smaller state space and a more than 30 times speedup.

Online beat tracking (OBT) has always been a challenging task. Due to the inaccessibility of future data and the need to make inference in real-time. We propose Don’t Look back! (DLB), a novel approach optimized for efficiency when performing OBT. DLB feeds the activations of a unidirectional RNN into an enhanced Monte-Carlo localization model to infer beat positions. Most preexisting OBT methods either apply some offline approaches to a moving window containing past data to make predictions about future beat positions or must be primed with past data at startup to initialize. Meanwhile, our proposed method only uses activation of the current time frame to infer beat positions. As such, without waiting at the beginning to receive a chunk, it provides an immediate beat tracking response, which is critical for many OBT applications. DLB significantly improves beat tracking accuracy over state-of-the-art OBT methods, yielding a similar performance to offline methods.

The orientation-dependence of the diffraction patterns of the 1D-grating experiments have been systematically studied by utilizing the 3-axis-rotation apparatus. It is shown that the spacings between adjacent diffraction orders are non-uniform. Namely, from “one end” of the pattern to “other end”, the spacings between adjacent diffraction orders gradually increase. In this article, we systematically study the orientation-dependence of the diffraction patterns of the 2D-cross-grating. The experiments show that the diffraction patterns are non-uniform as that of the 1D-grating experiments. This extraordinary phenomenon challenges the existing theoretical description of the grating experiments. The above observed phenomena provide the comprehensive phenomena to theoretical study of the grating experiments. We suggest that the complete mathematical model should contain three rotation angles as parameters and should be able to describe the phenomena of the non-uniform-spacing, curved, expanded, inclined and mirror-symmetric diffraction patterns consistently.