Search

In the era of large-scale pretrained models, artificial neural networks (ANNs) have excelled in natural language understanding (NLU) tasks. However, their success often necessitates substantial computational resources and energy consumption. To address this, we explore the potential of spiking neural networks (SNNs) in NLU—a promising avenue with demonstrated advantages, including reduced power consumption and improved efficiency due to their event-driven characteristics. We propose the SpikingMiniLM, a novel spiking Transformer model tailored for natural language understanding. We first introduce a multistep encoding method to convert text embeddings into spike trains. Subsequently, we redesign the attention mechanism and residual connections to make our model operate on the pure spike-based paradigm without any normalization technique. To facilitate stable and fast convergence, we propose a general parameter initialization method grounded in the stable firing rate principle. Furthermore, we apply an ANN-to-SNN knowledge distillation to overcome the challenges of pretraining SNNs. Our approach achieves a macro-average score of 75.5 on the dev sets of the GLUE benchmark, retaining 98% of the performance exhibited by the teacher model MiniLMv2. Our smaller model also achieves similar performance to BERTMINI with fewer parameters and much lower energy consumption, underscoring its competitiveness and resource efficiency in NLU tasks. [ABSTRACT FROM AUTHOR]

Background: The diagnosis and treatment of attention deficit hyperactivity disorder (ADHD) comorbid with epilepsy have been insufficiently addressed in China. We conducted a study in China to investigate the current status, diagnosis, and treatment of ADHD in children to further our understanding of ADHD comorbid with epilepsy, strengthen its management, and improve patients' quality of life. Methods: We carried out a multicenter cross-sectional survey of children with epilepsy across China between March 2022 and August 2022. We screened all patients for ADHD and compared various demographic and clinical factors between children with and without ADHD, including gender, age, age at epilepsy onset, duration of epilepsy, seizure types, seizure frequency, presence of epileptiform discharges, and treatment status. Our objective was to explore any possible associations between these characteristics and the prevalence of ADHD. Results: Overall, 395 epilepsy patients aged 6–18 years were enrolled. The age at seizure onset and duration of epilepsy ranged from 0.1–18 to 0.5–15 years, respectively. Focal onset seizures were observed in 212 (53.6%) patients, while 293 (76.3%) patients had epileptiform interictal electroencephalogram (EEG) abnormalities. Among the 370 patients treated with anti-seizure medications, 200 (54.1%) had monotherapy. Although 189 (47.8%) patients had ADHD, only 31 received treatment for it, with the inattentive subtype being the most common. ADHD was more common in children undergoing polytherapy compared to those on monotherapy. Additionally, poor seizure control and the presence of epileptiform interictal EEG abnormalities may be associated with a higher prevalence of ADHD. Conclusions: While the prevalence of ADHD was higher in children with epilepsy than in normal children, the treatment rate was notably low. This highlights the need to give more importance to the diagnosis and treatment of ADHD in children with epilepsy. [ABSTRACT FROM AUTHOR]

The potential therapeutic benefits of human dental pulp stem cells (HDPSCs) in dental regenerative medicine have been demonstrated. However, little is known about the molecular mechanisms regulating the biological characteristics of HDPSCs. The experiment aims to explore whether VEGF activates signaling pathways such as FAK, PI3K, Akt, and p38 in HDPSCs, and to investigate the molecular mechanisms by which VEGF influences proliferation and migration of HDPSCs. Normal and inflamed human dental pulp (HDP) samples were collected, and the levels of VEGF in HDP were assessed. HDPSCs were cultured and purified. HDPSCs were stimulated with lipopolysaccharide (LPS) at gradient concentrations, and real-time quantitative polymerase chain reaction (qPCR) was used to assess changes in VEGF mRNA. Gradient concentrations of VEGF were used to stimulate HDPSCs, and cell migration ability was evaluated through scratch assays and Transwell chamber experiments. Phosphorylation levels of FAK, AKT, and P38 were assessed using Western blotting. Inhibitors of VEGFR2, FAK, AKT, P38, and VEGF were separately applied to HDPSCs, and cell migration ability and phosphorylation levels of FAK, AKT, and P38 were determined. The results indicated significant differences in VEGF levels between normal and inflamed HDP tissues, with levels in the inflamed state reaching 435% of normal levels (normal: 87.91 ng/mL, inflamed: 382.76 ng/mL, P < 0.05). LPS stimulation of HDPSCs showed a significant increase in VEGF mRNA expression with increasing LPS concentrations (LPS concentrations of 0.01, 0.1, 1, and 10 μg/mL resulted in VEGF mRNA expressions of 181.2%, 274.2%, 345.8%, and 460.9%, respectively, P < 0.05). VEGF treatment significantly enhanced the migration ability of HDPSCs in Transwell chamber experiments, with migration rates increasing with VEGF concentrations (VEGF concentrations of 0, 1, 10, 20, 50, and 100 ng/mL resulted in migration rates of 8.41%, 9.34%, 21.33%, 28.41%, 42.87%, and 63.15%, respectively, P < 0.05). Inhibitors of VEGFR2, FAK, AKT, P38, and combined VEGF stimulation demonstrated significant migration inhibition, with migration rates decreasing to 8.31%, 12.64%, 13.43%, 18.32%, and 74.17%, respectively. The migration rate with combined VEGF stimulation showed a significant difference (P < 0.05). The analysis of phosphorylation levels revealed that VEGF stimulation significantly activated phosphorylation of FAK, AKT, and P38, with phosphorylation levels increasing with VEGF concentrations (P < 0.05). The VEGF/VEGFR2 signaling axis regulated the migration ability of HDPSCs through the FAK/PI3K/AKT and P38MAPK pathways. This finding highlighted not only the crucial role of VEGF in injury repair of HDPSCs but also provided important clues for a comprehensive understanding of the potential applications of this signaling axis in dental regenerative medicine. [ABSTRACT FROM AUTHOR]

SUMMARY: Low phytic acid (lpa) crop is considered as an effective strategy to improve crop nutritional quality, but a substantial decrease in phytic acid (PA) usually has negative effect on agronomic performance and its response to environment adversities. Myo‐inositol‐3‐phosphate synthase (MIPS) is the rate‐limiting enzyme in PA biosynthesis pathway, and regarded as the prime target for engineering lpa crop. In this paper, the rice MIPS gene (RINO2) knockout mutants and its wild type were performed to investigate the genotype‐dependent alteration in the heat injury‐induced spikelet fertility and its underlying mechanism for rice plants being imposed to heat stress at anthesis. Results indicated that RINO2 knockout significantly enhanced the susceptibility of rice spikelet fertility to heat injury, due to the severely exacerbated obstacles in pollen germination and pollen tube growth in pistil for RINO2 knockout under high temperature (HT) at anthesis. The loss of RINO2 function caused a marked reduction in inositol and phosphatidylinositol derivative concentrations in the HT‐stressed pollen grains, which resulted in the strikingly lower content of phosphatidylinositol 4,5‐diphosphate (PI (4,5) P2) in germinating pollen grain and pollen tube. The insufficient supply of PI (4,5) P2 in the HT‐stressed pollen grains disrupted normal Ca2+ gradient in the apical region of pollen tubes and actin filament cytoskeleton in growing pollen tubes. The severely repressed biosynthesis of PI (4,5) P2 was among the regulatory switch steps leading to the impaired pollen germination and deformed pollen tube growth for the HT‐stressed pollens of RINO2 knockout mutants. Significance Statement: For engineering lpa rice, the knockout of RINO2 significantly enhanced susceptibility of spikelet fertility to heat injury at anthesis, which was mostly ascribed to the heat injury‐induced obstacle in pollen germination and pollen tube growth. High temperature significantly reduced the concentrations of inositol and PI (4,5) P2 concentration in pollen grains, this occurrence was strongly responsible for the disorder Ca2+ gradient and actin filament cytoskeleton in growing pollen tubes. These findings provide new understanding for the regulatory mechanism by which the loss of RINO2 function negatively affects pollen germination and pollen tube growth in response to HT exposure at anthesis. [ABSTRACT FROM AUTHOR]

Sea Surface Temperature (SST) is of great importance to study several major phenomena due to ocean interactions with other earth systems. Previous studies on SST based on statistical inference methods were less accurate for longer prediction lengths. A considerable number of studies in recent years involve machine learning for SST modeling. These models were able to mitigate this problem to some length by modeling SST patterns and trends. Sequence analysis by decomposition is used for SST forecasting in several studies. Ensemble Empirical Mode Decomposition (EEMD) has been proven in previous studies as a useful method for this. The application of EEMD in spatiotemporal modeling has been introduced as Multidimensional EEMD (MEEMD). The aim of this study is to employ fast MEEMD methods to decompose the SST spatiotemporal dataset and apply a Convolutional Long Short-Term Memory (ConvLSTM)-based model to model and forecast SST. The results show that the fast MEEMD method is capable of enhancing spatiotemporal SST modeling compared to the Linear Inverse Model (LIM) and ConvLSTM model without decomposition. The model was further validated by making predictions from April to May 2023 and comparing them to original SST values. There was a high consistency between predicted and real SST values. [ABSTRACT FROM AUTHOR]

The identification of prototypical waveforms, such as sleep spindles and epileptic spikes, is crucial for the diagnosis of neurological disorders. These prototypical waveforms are usually recurrently presented in certain brain states, serving as potential biomarkers for clinical evaluations. Convolutional sparse coding (CSC) approaches have demonstrated strength in identifying recurrent patterns in time-series. However, existing CSC approaches do not explicitly explore state-specific patterns, making it difficult to identify state-related biomarkers. To address this problem, we propose state-sensitive CSC to learn state-specific prototypical waveforms. Specifically, we model signals of a certain state with specific waveforms that only appear frequently in this state and background waveforms that are independent of states. Based on this, state-sensitive CSC separates state-specific waveforms from background ones explicitly by incorporating incoherence constraints into optimizations. Experiments with epilepsy brain signals demonstrate that our approach can effectively identify prototypical waveforms in pre-ictal states, providing potential biomarkers for seizure prediction. Our approach provides a promising tool for automatic biomarker candidate identification. [ABSTRACT FROM AUTHOR]

Massive multiplayer online role-playing games (MMORPG) have been becoming one of the most popular and exciting online games. In recent years, a cheating phenomenon called real money trading (RMT) has arisen and damaged the fantasy world in many ways. RMT is the sale of in-game items, currency, or even characters to earn real money, breaking the balance of the game economy ecosystem and damaging the game experience. Therefore, some studies have emerged to address the problem of RMT detection. However, they cannot well handle the label uncertainty problem in practice, where there are only labeled RMT samples (positive samples) and unlabeled samples, which could either be RMT samples or normal transactions (negative samples). Meanwhile, the trading relationship between RMTers is modeled in a simple way, leading to some normal transactions being falsely classified as RMT. In this article, we propose PU-Detector, a novel framework based on PU learning (learning from positive and unlabeled data) for RMT detection, considering the fact that there are only labeled RMT samples and other unlabeled transactions. We first automatically estimate the likelihood of one transaction being RMT by developing an improved PU learning method and proposing an assessment rule. Sequentially, we use the estimated likelihood as edge weight to construct a trading graph to learn trader representation. Then, with the trader representations and basic trading features, we detect RMT samples by the improved PU learning method. PU-Detector is evaluated on a large-scale real world dataset consisting of 33,809,956 transaction logs generated by 43,217 unique players. Compared with other approaches, it achieves the state-of-the-art performance and demonstrates its advantages in detecting underlying RMT samples. [ABSTRACT FROM AUTHOR]

Spiking neural networks (SNNs) are gaining increasing attention for their biological plausibility and potential for improved computational efficiency. To match the high spatial-temporal dynamics in SNNs, neuromorphic chips are highly desired to execute SNNs in hardware-based neuron and synapse circuits directly. This paper presents a large-scale neuromorphic chip named Darwin3 with a novel instruction set architecture, which comprises 10 primary instructions and a few extended instructions. It supports flexible neuron model programming and local learning rule designs. The Darwin3 chip architecture is designed in a mesh of computing nodes with an innovative routing algorithm. We used a compression mechanism to represent synaptic connections, significantly reducing memory usage. The Darwin3 chip supports up to 2.35 million neurons, making it the largest of its kind on the neuron scale. The experimental results showed that the code density was improved by up to 28.3× in Darwin3, and that the neuron core fan-in and fan-out were improved by up to 4096× and 3072× by connection compression compared to the physical memory depth. Our Darwin3 chip also provided memory saving between 6.8× and 200.8× when mapping convolutional spiking neural networks onto the chip, demonstrating state-of-the-art performance in accuracy and latency compared to other neuromorphic chips. [ABSTRACT FROM AUTHOR]

Purpose: Grb2 associated binding protein 1 (Gab1) is an adaptor protein that is important for intracellular signal transduction which involved in several pathological process. However, the role of Gab1 in pressure overload-induced ventricular arrhythmias (VAs) remain poorly understood. In the current study, we aimed to test the role of Gab1 in VA susceptibility induced by pressure overload. Methods: We overexpressed Gab1 in the hearts using an adeno-associated virus 9 (AAV9) system through tail vein injection. Aortic banding (AB) surgery was performed in C57BL6/J mice to induce heart failure (HF). Four weeks following AB, histology, echocardiography, and biochemical analysis were conducted to investigate cardiac structural remodeling and electrophysiological studies were performed to check the electrical remodeling. Western blot analysis was used to explore the underlying mechanisms. Results: The mRNA and protein expression were downregulated in AB hearts compared to sham hearts. Gab1 overexpression significantly reversed AB-induced cardiac structural remodeling including ameliorated AB-induced cardiac dysfunction, cardiac fibrosis, and inflammatory response. Moreover, Gab1 overexpression also markedly alleviated AB-induced electrical remodeling including ion channel alterations and VA susceptibility. Mechanistically, we found that TLR4/MyD88/NF-κB contributes to the cardio protective effect of Gab1 overexpression on AB-induced VAs. Conclusions: Our study manifested that Gab1 may serve as a promising anti-arrhythmic target via inhibiting TLR4/MyD88/NF-κB signaling pathway induced by AB. [ABSTRACT FROM AUTHOR]

This paper aims to reveal the motion law and collision behaviors of shotcrete particle flow jets. A physical model of the jet flow field composed of a nozzle structure and jet area was constructed and meshes with various sizes were used to mesh the nozzle and jet area. With the basic contact parameters and contact model parameters of the particles set, the CFD-DEM-coupling simulation method was adopted to perform the numerical simulation of concrete-particle-flow-jet impingement. The variation laws of the continuous-phase velocity and pressure drop of the shotcrete, coarse-aggregate motion characteristics, and particle collision behavior under the interaction of the continuous and discrete phases were obtained. The results showed that the velocity field and pressure-drop field of the continuous phase had an ideal symmetry in the XY plane in the stable injection stage, the continuous-phase velocity gradually increased inside the nozzle and gradually decreased after entering the jet area, the continuous-phase pressure drop was the maximum at the nozzle inlet, and the pressure value at the nozzle outlet became atmospheric pressure. The central axis of the particle flow jet was displaced by 0.15 m in the negative direction of the Y-axis under the action of gravity, the diffusion angle of the small particles that exited the nozzle and entered the jet area was larger than that of the large particles, and the large-particle jets were more concentrated and easier to spray into the designated spraying areas. The particle flow reached a stable jet state about 0.3 s after the jet began, and the peak velocity of the 4 mm particles in the flow reached 25 m/s, while the peak velocity of the 12 mm particles was only 19 m/s. The acceleration time for particles of different sizes to reach the peak velocity also varied, and the large particles took longer to reach the maximum velocity: small particles reached their peak within 0.4 m–8 m of the jet area, and large particles reached their peak within 0.8 m–1.2 m of the jet area. The particle velocity peaked within 0.6 m–1 m of the jet area. Particle collision took three forms: particle collision with the inner wall of the nozzle, interparticle collision, and particle collision with the sprayed wall. The collision between the particles and the sprayed wall was the main form leading to the rebound of the wet shotcrete, and the rebound angle after particle collision was uncertain. [ABSTRACT FROM AUTHOR]

The assessment of consciousness states, especially distinguishing minimally conscious states (MCS) from unresponsive wakefulness states (UWS), constitutes a pivotal role in clinical therapies. Despite that numerous neural signatures of consciousness have been proposed, the effectiveness and reliability of such signatures for clinical consciousness assessment still remains an intense debate. Through a comprehensive review of the literature, inconsistent findings are observed about the effectiveness of diverse neural signatures. Notably, the majority of existing studies have evaluated neural signatures on a limited number of subjects (usually below 30), which may result in uncertain conclusions due to small data bias. This study presents a systematic evaluation of neural signatures with large‐scale clinical resting‐state electroencephalography (EEG) signals containing 99 UWS, 129 MCS, 36 emergence from the minimally conscious state, and 32 healthy subjects (296 total) collected over 3 years. A total of 380 EEG‐based metrics for consciousness detection, including spectrum features, nonlinear measures, functional connectivity, and graph‐based measures, are summarized and evaluated. To further mitigate the effect of data bias, the evaluation is performed with bootstrap sampling so that reliable measures can be obtained. The results of this study suggest that relative power in alpha and delta serve as dependable indicators of consciousness. With the MCS group, there is a notable increase in the phase lag index‐related connectivity measures and enhanced functional connectivity between brain regions in comparison to the UWS group. A combination of features enables the development of an automatic detector of conscious states. [ABSTRACT FROM AUTHOR]

Neuromorphic computing is an emerging research field that aims to develop new intelligent systems by integrating theories and technologies from multiple disciplines, such as neuroscience, deep learning and microelectronics. Various software frameworks have been developed for related fields, but an efficient framework dedicated to spike-based computing models and algorithms is lacking. In this work, we present a Python-based spiking neural network (SNN) simulation and training framework, named SPAIC, that aims to support brain-inspired model and algorithm research integrated with features from both deep learning and neuroscience. To integrate different methodologies from multiple disciplines and balance flexibility and efficiency, SPAIC is designed with a neuroscience-style frontend and a deep learning-based backend. Various types of examples are provided to demonstrate the wide usability of the framework, including neural circuit simulation, deep SNN learning and neuromorphic applications. As a user-friendly, flexible, and high-performance software tool, it will help accelerate the rapid growth and wide applicability of neuromorphic computing methodologies. [ABSTRACT FROM AUTHOR]

The comprehensive impacts of biotic and abiotic factors on the natural regeneration of Abies georgei var. smithii (Abies) forests in Tibet are not thoroughly understood. To address this gap, our study focused on the regeneration densities of Abies seedlings, saplings, and small trees across 21 plots (each 20 m × 20 m) along an elevation gradient from 3730 m to 4330 m in the Sygera Mountains of Southeastern Tibet. We meticulously measured a suite of 11 variables that describe stand structures and ecological conditions. Through analyses using Spearman's correlation analysis, hierarchical partitioning, and multiple linear regression, we identified key ecological drivers for successful Abies regeneration. Our results highlighted a peak in the abundance of seedlings, saplings, and small trees at an elevation of 3930 m. As the elevation rose from 3730 m to 4330 m, we observed an initial increase followed by a decrease in canopy cover (canopy), mother tree density (MotherT), 1000-seed weight (SeedW), litter thickness (LitterT), moss cover (MossC), moss thickness (MossT), soil moisture (SM), and soil bulk density, while mean annual temperature and soil depth to permafrost consistently decreased. The critical ecological drivers for Abies natural regeneration were identified as follows: MossT was pivotal for seedling density; canopy and MossC were influential for sapling density, and MotherT was the main factor affecting the density of small trees. This study suggests that a high density of mother trees and a thick and highly covered layer of moss are conducive to the natural regeneration of Abies in the Sygera Mountains. Understanding the current status of regeneration is vital for informing conservation and management strategies for Abies forests in Tibet. [ABSTRACT FROM AUTHOR]

Significant progress has been made in image inpainting methods in recent years. However, they are incapable of producing inpainting results with reasonable structures, rich detail, and sharpness at the same time. In this paper, we propose the Pyramid-VAE-GAN network for image inpainting to address this limitation. Our network is built on a variational autoencoder (VAE) backbone that encodes high-level latent variables to represent complicated high-dimensional prior distributions of images. The prior assists in reconstructing reasonable structures when inpainting. We also adopt a pyramid structure in our model to maintain rich detail in low-level latent variables. To avoid the usual incompatibility of requiring both reasonable structures and rich detail, we propose a novel cross-layer latent variable transfer module. This transfers information about long-range structures contained in high-level latent variables to low-level latent variables representing more detailed information. We further use adversarial training to select the most reasonable results and to improve the sharpness of the images. Extensive experimental results on multiple datasets demonstrate the superiority of our method. Our code is available at https://github.com/thy960112/Pyramid-VAE-GAN. [ABSTRACT FROM AUTHOR]

Emerging memories have been developed as new physical infrastructures for hosting neural networks owing to their low-power analog computing characteristics. However, accurately and efficiently programming devices in an analog-valued array is still largely limited by the intrinsic physical non-idealities of the devices, thus hampering their applications in in-situ training of neural networks. Here, we demonstrate a passive electrochemical memory (ECRAM) array with many important characteristics necessary for accurate analog programming. Different image patterns can be open-loop and serially programmed into our ECRAM array, achieving high programming accuracies without any feedback adjustments. The excellent open-loop analog programmability has led us to in-situ train a bilayer neural network and reached software-like classification accuracy of 99.4% to detect poisonous mushrooms. The training capability is further studied in simulation for large-scale neural networks such as VGG-8. Our results present a new solution for implementing learning functions in an artificial intelligence hardware using emerging memories. Memory devices with open-loop analog programmability are highly desired in training tasks. Here, the authors developed an electrochemical memory array that can be accurately programmed without any feedback, offering unique capabilities for training. [ABSTRACT FROM AUTHOR]

This paper developed a comprehensive magnetic resonance imaging (MRI)-compatible electrophysiological (EP) acquisition system, which can acquire various physiological electrical signals, including electrocardiography (ECG), electromyography (EMG), electroencephalography (EEG) and electrocorticogram (ECoG), and EP recording combined with multimodal stimulation. The system is designed to be compatible with the 7-Tesla (7T) ultra-high field MRI environment, providing convenience for neuroscience and physiological research. To achieve MRI compatibility, the device uses magnetically compatible materials and shielding measures on the hardware and algorithm processing on the software side. Different filtering algorithms are adopted for different signals to suppress all kinds of interference in the MRI environment. The system can allow input signals up to ±0.225 V and channels up to 256. The equipment has been tested and proven to be able to collect a variety of physiological electrical signals effectively. When scanned under the condition of a 7T high-intensity magnetic field, the system does not generate obvious heating and can meet the safety requirements of MRI and EEG acquisition requirements. Moreover, an algorithm is designed and improved to efficiently and automatically remove the gradient artifact (GA) noise generated by MRI, which is a thousand-fold gradient artifact. Overall, this work proposes a complete, portable, MRI-compatible system that can collect a variety of physiological electrical signals and integrate more efficient GA removal algorithms. [ABSTRACT FROM AUTHOR]

Biometric features, e.g., fingerprints, the iris, and the face, have been widely used to authenticate individuals. However, most biometrics are not cancellable, i.e., once these biometric features are cloned or stolen, they cannot be replaced easily. Unlike traditional biometrics, brain biometrics are extremely difficult to clone or forge due to the natural randomness across different individuals, which makes them an ideal option for identity authentication. Most existing brain biometrics are based on electroencephalogram (EEG), which is usually demonstrated unstable performance due to the low signal-to-noise ratio (SNR). For the first time, we propose the use of intracortical brain signals, which have higher resolution and SNR, to realize the construction of the high-performance brain biometrics. Specifically, we put forward a novel brain-based key generation approach called multidimensional Gaussian fitted bit allocation (MGFBA). The proposed MGFBA method extracts keys from the local field potential of ten rats with high reliability and high entropy. We found that with the proposed MGFBA, the average effective key length of the brain biometrics was 938 bits, while achieving high authentication accuracy of 88.1% at a false acceptance rate of 1.9%, which is significantly improved compared to conventional EEG-based approaches. In addition, the proposed MGFBA-based keys can be conveniently revoked using different motor behaviors with high entropy. Experimental results demonstrate the potential of using intracortical brain signals for reliable authentication and other security applications. [ABSTRACT FROM AUTHOR]

The theory of viscoelastic collision mechanics is used to explore the rebound and adhesion mechanisms of single aggregates in shotcrete at the meso-level in this study. Abrasive balls with different radiuses were used, fully considering the characteristics of the aggregates and the sprayed wall, to build multifactor physical models for the shotcrete rebound of the aggregate. The movement trajectories, velocities, and energy dissipations of abrasive balls are analyzed, and the shotcrete-rebound mechanical model of a single aggregate based on wet-mix shotcrete is proposed. The results of the study show that there are obvious differences in the rebound velocity of the aggregate in different collision forms, and the viscoelastic properties of aggregates affect the jet stability of wet-mix shotcrete. According to the quantitative and qualitative analyses, the energy loss during the collision-rebound of the aggregate is inversely proportional to the rebound velocity and proportional to the rebound resistance. The functional relationship between the resistance F and the particle radius R during the rebound and adhesion of aggregate can be expressed as F=CR1/5 , where C is a constant. [ABSTRACT FROM AUTHOR]

Biometrics, e.g., fingerprints, the iris, and the face, have been widely used to authenticate individuals. However, most biometrics are not cancellable, i.e., once these traditional biometrics are cloned or stolen, they cannot be replaced easily. Unlike traditional biometrics, brain biometrics are extremely difficult to clone or forge due to the natural randomness across different individuals, which makes them an ideal option for identity authentication. Most existing brain biometrics are based on an electroencephalogram (EEG), which typically demonstrates unstable performance due to the low signal-to-noise ratio (SNR). Thus, in this paper, we propose the use of intracortical brain signals, which have higher resolution and SNR, to realize the construction of a high-performance brain biometric. Significantly, this is the first study to investigate the features of intracortical brain signals for identification. Specifically, several features based on local field potential are computed for identification, and their performance is compared with different machine learning algorithms. The results show that frequency domain features and time-frequency domain features are excellent for intra-day and inter-day identification. Furthermore, the energy features perform best among all features with 98% intra-day and 93% inter-day identification accuracy, which demonstrates the great potential of intracraial brain signals to be biometrics. This paper may serve as a guidance for future intracranial brain researches and the development of more reliable and high-performance brain biometrics. [ABSTRACT FROM AUTHOR]

Crime analysis is important for social security management. With the advance of crowd sensing techniques, abundant multisource crowd sensed data could be used for crime analysis. The occurrence of crimes usually has some patterns in terms of temporal and spatial aspects. Investigating the spatio-temporal correlation of crimes could provide more useful cues for crime analysis and help discover underlying crime patterns. In this paper, we conduct a spatio-temporal study to understand urban crimes leveraging multisource crowd sensed data, including crime data, meteorological data, POI distribution, and taxi trips. Specifically, we first present monthly temporal trends and spatial distribution of crimes. We then investigate the spatio-temporal correlation using meteorological data (e.g., weather conditions and air temperature) and POI distribution and taxi trips. It is found that taxi trips and air temperature have a strong correlation with the crime, and some POI categories have a valuable correlation with the crime, e.g., College & University. We also find that Overcast days would witness more crime than other weather conditions. [ABSTRACT FROM AUTHOR]

High temperatures (HT) cause pollen abortion and poor floret fertility in rice, which is closely associated with excessive accumulation of reactive oxygen species (ROS) in the developing anthers. However, the relationships between accumulation of abscisic acid (ABA) and ROS, and their effects on tapetum‐specific programmed cell death (PCD) in HT‐stressed anthers are poorly characterised. Here, we determined the spatiotemporal changes in ABA and ROS levels, and their relationships with tapetal PCD under HT exposure. Mutants lacking ABA‐activated protein kinase 2 (SAPK2) functions and exogenous ABA treatments were used to explore the effects of ABA signalling on the induction of PCD and ROS accumulation during pollen development. HT‐induced pollen abortion was tightly associated with ABA accumulation and oxidative stress. The higher ABA level in HT‐stressed anthers resulted in the earlier initiation of PCD induction and subsequently abnormal tapetum degeneration by activating ROS accumulation in developing anthers. Interactions between SAPK2 and DEAD‐box ATP‐dependent RNA helicase elF4A‐1 (RH4) were required for ABA‐induced ROS generation in developing anthers. The OsSAPK2 knockout mutants showed the impaired PCD responses in the absence of HT. However, the deficiency of SAPK2 functions did not suppress the ABA‐mediated ROS generation in HT‐stressed anthers. Summary statement: High temperatures (HT) disturbed the cellular distributions of ABA and ROS contents in rice anthers. The higher ABA in HT‐stressed anthers contributed to HT‐induced pollen abortion by activating ROS generation and oxidation in developing anthers. Interactions between SAPK2 and DEAD‐box ATP‐dependent RNA helicase elF4A‐1 (RH4) were required for ABA‐induced ROS generation during pollen development and its response to HT exposure. [ABSTRACT FROM AUTHOR]

In sewer pipes, haze caused by the humid environment seriously impairs the quality of closed‐circuit television (CCTV) images, which leads to poor performance of subsequent pipe defects detection. Meanwhile, the complexity of sewer images, such as steep depth change and extensive textureless regions, brings great challenges to the performance or application of general dehazing algorithms. Therefore, this study estimates sewer depth maps first with the help of the water–pipewall borderlines to produce the paired dehazing dataset. Then a structure‐aware nonlocal network (SANL‐Net) is proposed with the detected borderlines and the dehazing result as two supervisory signals. SANL‐Net shows its superiority over other state‐of‐the‐art approaches with 147 in mean square error (MSE), 27.28 in peak signal to noise ratio (PSNR), 0.8963 in structural similarity index measure (SSIM), and 15.47M in parameters. Also, the outstanding performance in real image dehazing implies the accuracy of depth estimation. Experimental results indicate that SANL‐Net significantly improves the performance of defects detection tasks, such as an increase of 23.16% in mean intersection over union (mIoU) for semantic segmentation. [ABSTRACT FROM AUTHOR]

Copyright of Rare Metals 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.)

The method of atmospheric corrosion protection of metals has always been a research hot spot at home and abroad. In this paper, superhydrophobic coating is prepared on the surface of galvanized steel by chemical etching using 1H,1H,2H,2H-perfluorodecyltriethoxysilane (PFDTES), graphene oxide (GO), anhydrous ethanol and water-based varnish as the main raw materials. The constitution of the superhydrophobic coating surface and the corrosion resistance of the steel are studied by electrochemical testing, X-ray diffraction (XRD), scanning electron microscopy (SEM), X-ray energy dispersive spectroscopy (EDS) and atomic force microscopy (AFM). Meanwhile, COMSOL software is used for the simulation of experiment. The results show that the surface of the superhydrophobic coating is composed of micro–nano sized papillary bulges, which play an important role in the improvement of metal corrosion resistance. The superhydrophobic coating effectively improves the alternating current impedance of the electrode and reduces the corrosion rate of the substrate. In addition, the results calculated by COMSOL software consist of the experimental results. [ABSTRACT FROM AUTHOR]

Active electrocorticogram (ECoG) electrodes can amplify weak electrophysiological signals and improve anti‐interference ability; however, traditional active electrodes are opaque and cannot realize photoelectric collaborative observation. In this study, an active and fully transparent ECoG array based on zinc oxide thin‐film transistors (ZnO TFTs) is developed as a local neural signal amplifier for electrophysiological monitoring. The transparency of the proposed ECoG array is up to 85%, which is superior to that of the previously reported active electrode arrays. Various electrical characterizations have demonstrated its ability to record electrophysiological signals with a higher signal‐to‐noise ratio of 19.9 dB compared to the Au grid (13.2 dB). The high transparency of the ZnO‐TFT electrode array allows the concurrent collection of high‐quality electrophysiological signals (32.2 dB) under direct optical stimulation of the optogenetic mice brain. The ECoG array can also work under 7‐Tesla magnetic resonance imaging to record local brain signals without affecting brain tissue imaging. As the most transparent active ECoG array to date, it provides a powerful multimodal tool for brain observation, including recording brain activity under synchronized optical modulation and 7‐Tesla magnetic resonance imaging. [ABSTRACT FROM AUTHOR]

Unsupervised person re-identification (re-ID) remains a challenging task. While extensive research has focused on the framework design and loss function, this paper shows that sampling strategy plays an equally important role. We analyze the reasons for the performance differences between various sampling strategies under the same framework and loss function. We suggest that deteriorated over-fitting is an important factor causing poor performance, and enhancing statistical stability can rectify this problem. Inspired by that, a simple yet effective approach is proposed, termed group sampling, which gathers samples from the same class into groups. The model is thereby trained using normalized group samples, which helps alleviate the negative impact of individual samples. Group sampling updates the pipeline of pseudo-label generation by guaranteeing that samples are more efficiently classified into the correct classes. It regulates the representation learning process, enhancing statistical stability for feature representation in a progressive fashion. Extensive experiments on Market-1501, DukeMTMC-reID and MSMT17 show that group sampling achieves performance comparable to state-of-the-art methods and outperforms the current techniques under purely camera-agnostic settings. Code has been available at https://github.com/ucas-vg/GroupSampling. [ABSTRACT FROM AUTHOR]

Objective: Brain-machine interfaces (BMIs) aim to provide direct brain control of devices such as prostheses and computer cursors, which have demonstrated great potential for motor restoration. One major limitation of current BMIs lies in the unstable performance due to the variability of neural signals, especially in online control, which seriously hinders the clinical availability of BMIs. Method: We propose a dynamic ensemble Bayesian filter (DyEnsemble) to deal with the neural variability in online BMI control. Unlike most existing approaches using fixed models, DyEnsemble learns a pool of models that contains diverse abilities in describing the neural functions. In each time slot, it dynamically weights and assembles the models according to the neural signals in a Bayesian framework. In this way, DyEnsemble copes with variability in signals and improves the robustness of online control. Results: Online BMI experiments with a human participant demonstrate that, compared with the velocity Kalman filter, DyEnsemble significantly improves the control accuracy (increases the success rate by 13.9% in the random target pursuit task) and robustness (performs more stably over different experiment days). Conclusion: Experimental results demonstrate the superiority of DyEnsemble in online BMI control. Significance: DyEnsemble frames a novel and flexible dynamic decoding framework for robust BMIs, beneficial to various neural decoding applications. [ABSTRACT FROM AUTHOR]

This article proposes a facial expression removal method to recover a 3D neutral face from a single 3D expressional or non-neutral face. We treat a 3D non-neutral face as the sum of its neutral one and the residual. This can be satisfied if the correspondence between 3D vertices of expressional faces and those of neutral faces is established. We propose a non-rigid deformation method to establish the correspondence between 3D faces. Then, according to algebra inequality, the minimization of a neutral face model can be replaced by the minimization of its upper bound, i.e., the errors of an expressional face model and a residual model. Thus, we co-optimize the representation errors of the latter two models and build the relationship between the representation coefficients of the two models. Given an expressional face as the input, its corresponding neutral face can be inferred by the associative representation parameters in these two models. In the testing stage, we use an iterative joint fitting scheme to obtain a more accurate recovery. Extensive experiments are conducted to evaluate our method. The results show that our method obtains considerably better performance than existing methods in terms of average root mean square errors and recognition rates, and also better visual effects. [ABSTRACT FROM AUTHOR]

The surfaces of real objects can visually appear to be glossy, matte, or anywhere in between, but essentially, they display varying degrees of diffuse and specular reflectance. Diffuse and specular reflectance provides different clues for light estimation. However, few methods simultaneously consider the contributions of diffuse and specular reflectance for light estimation. To this end, we propose ReDDLE-Net, which performs Reflectance Decomposition for Directional Light Estimation. The primary idea is to take advantage of diffuse and specular clues and adaptively balance the contributions of estimated diffuse and specular components for light estimation. Our method achieves a superior performance advantage over state-of-the-art directional light estimation methods on the DiLiGenT benchmark. Meanwhile, the proposed ReDDLE-Net can be combined with existing calibrated photometric stereo methods to handle uncalibrated photometric stereo tasks and achieve state-of-the-art performance. [ABSTRACT FROM AUTHOR]

Key message: An alanine to valine mutation of glutamyl-tRNA reductase's 510th amino acid improves 5-aminolevulinic acid synthesis in rice. 5-aminolevulinic acid (ALA) is the common precursor of all tetrapyrroles and plays an important role in plant growth regulation. ALA is synthesized from glutamate, catalyzed by glutamyl-tRNA synthetase (GluRS), glutamyl-tRNA reductase (GluTR), and glutamate-1-semialdehyde aminotransferase (GSAT). In Arabidopsis, ALA synthesis is the rate-limiting step in tetrapyrrole production via GluTR post-translational regulations. In rice, mutations of GluTR and GSAT homologs are known to confer chlorophyll deficiency phenotypes; however, the enzymatic activity of rice GluRS, GluTR, and GSAT and the post-translational regulation of rice GluTR have not been investigated experimentally. We have demonstrated that a suppressor mutation in rice partially reverts the xantha trait. In the present study, we first determine that the suppressor mutation results from a G → A nucleotide substitution of OsGluTR (and an A → V change of its 510th amino acid). Protein homology modeling and molecular docking show that the OsGluTRA510V mutation increases its substrate binding. We then demonstrate that the OsGluTRA510V mutation increases ALA synthesis in Escherichia coli without affecting its interaction with OsFLU. We further explore homologous genes encoding GluTR across 193 plant species and find that the amino acid (A) is 100% conserved at the position, suggesting its critical role in GluTR. Thus, we demonstrate that the gain-of-function OsGluTRA510V mutation underlies suppression of the xantha trait, experimentally proves the enzymatic activity of rice GluRS, GluTR, and GSAT in ALA synthesis, and uncovers conservation of the alanine corresponding to the 510th amino acid of OsGluTR across plant species. [ABSTRACT FROM AUTHOR]

Nutrient enrichment of lakes from anthropogenic activities is a significant and increasing issue globally, impairing the health, biodiversity and service provisioning from lakes, with impacts on cultural, recreational, economic and aesthetic values. Internal nutrient loads from lakebed sediment releases are a primary cause of lake eutrophication and have necessitated geoengineering methods to mitigate releases and speed up recovery from eutrophication. Our objective in this review was to evaluate the use of oxygen nanobubbles as a geoengineering technology to remediate low oxygen conditions at the lake sediment/water interface, as a precursor to alleviating eutrophication linked to high internal nutrient loads. Oxygen nanobubbles (NBs) are bubbles < 1000 nm formed at the interface of solid surfaces and aqueous solutions. These bubbles have higher density than water, persist for longer and facilitate greater oxygen solubility than larger bubbles. Methods have been developed to enable NB formation at the surface of carrier materials, which are then used in conjunction with modified local soils (MLSs), to 'floc, lock and oxygenate' to strip nutrients from the water column, locking them in lakebed sediments and oxygenating the sediments to prevent re-release of nutrients. Most studies of NBs for lake restoration have thus far only demonstrated their potential for this purpose, using short-term, small-scale core incubations conducted mainly in laboratory settings. Work is required to (1) address scalability, including procurement and cost, (2) extend laboratory incubation studies to large outdoor enclosures and pond/lake trials, (3) examine longevity of the effects in the natural environment, including potential for MLSs to smother benthos and/or have toxic effects, and (4) extend to a range of lake environments and MLS types. Legal, cultural and social acceptance of the technology is another prerequisite of applications in the natural environment and requires individualised analysis. Until these issues are addressed in a systematic way that addresses scalability and recommends suitable carrier materials and MLSs, NBs may continue to remain largely untried as a geoengineering method to address lake eutrophication. [ABSTRACT FROM AUTHOR]

Push chain moist-mix concrete spraying machine is widely used in shotcrete construction processes, but shotcrete surplus has always been the main problem affecting work efficiency. To improve the efficiency, the orifice plate is employed in the discharge cylinder, considering three different distributions (equal, left concentrated and both side distribution) of the holes and two hole diameters (6 mm and 8 mm). In different orifice plate, the discharge cylinder flow field is studied employing the FLUENT software by the computational fluid dynamic method. Besides, dimensionless evaluation index of the discharge cylinder flow field is defined based on the force of particles. By comparing dimensionless index and velocity vector, the best orifice plate (both side distribution and 8 mm) is presented. Compared without orifice plate, the experiment showed the shotcrete surplus ability is reduced by 26 %–87 %. This study provides guidance for the design and optimization of pneumatic conveying equipment. [ABSTRACT FROM AUTHOR]

Macrophages originating from the yolk sac or bone marrow play essential roles in tissue homeostasis and disease. Bone marrow-derived monocytes differentiate into Ly6Chi and Ly6Clo macrophages according to the differential expression of the surface marker protein Ly6C. Ly6Chi and Ly6Clo cells possess diverse functions and transcriptional profiles and can accelerate the disease process or support tissue repair and reconstruction. In this review, we discuss the basic biology of Ly6Chi and Ly6Clo macrophages, including their origin, differentiation, and phenotypic switching, and the diverse functions of Ly6Chi and Ly6Clo macrophages in homeostasis and disease, including in injury, chronic inflammation, wound repair, autoimmune disease, and cancer. Furthermore, we clarify the differences between Ly6Chi and Ly6Clo macrophages and their connections with traditional M1 and M2 macrophages. We also summarize the limitations and perspectives for Ly6Chi and Ly6Clo macrophages. Overall, continued efforts to understand these cells may provide therapeutic approaches for disease treatment. [ABSTRACT FROM AUTHOR]

With the reduction of the semiconductor process size, the problems of "memory wall" and "power wall" in von Neumann architectures are becoming increasingly prominent. To solve these problems, brain-inspired computing unifies computing and storage. [ABSTRACT FROM AUTHOR]

Recent work on spiking neural networks (SNNs) has focused on achieving deep architectures. They commonly use backpropagation (BP) to train SNNs directly, which allows SNNs to go deeper and achieve higher performance. However, the BP training procedure is computing intensive and complicated by many trainable parameters. Inspired by global pooling in convolutional neural networks (CNNs), we present the spike probabilistic global pooling (SPGP) method based on a probability function for training deep convolutional SNNs. It aims to remove the difficulty of too many trainable parameters brought by multiple layers in the training process, which can reduce the risk of overfitting and get better performance for deep SNNs (DSNNs). We use the discrete leaky-integrate-fire model and the spatiotemporal BP algorithm for training DSNNs directly. As a result, our model trained with the SPGP method achieves competitive performance compared to the existing DSNNs on image and neuromorphic data sets while minimizing the number of trainable parameters. In addition, the proposed SPGP method shows its effectiveness in performance improvement, convergence, and generalization ability. [ABSTRACT FROM AUTHOR]

Ambient temperature and nitrogen (N) fertilizer are two of the most important factors that affect rice grain quality. However, less information has been available on the interactive effect of N fertilizer and ambient temperature on grain quality under stressful high temperature (HT). In this article, the effects of panicle N fertilizer, ambient temperature, and their interaction on starch composition, particle size distribution of starch granules, starch physicochemical properties, and storage protein accumulation in milled grains were investigated to clarify the potential role of panicle N fertilizer topdressing in regulating rice grain quality under stressful HT by using a two-factor experiment of three N levels in combination with two temperature regimes. Results showed that appropriate application of panicle N fertilizer could attenuate the adverse effect of HT during grain filling on milling quality and chalky occurrence to some extent, particularly for the effective alleviation of HT-induced decrease in milling quality. However, the topdressing of panicle N fertilizer tended to enhance starch gelatinization enthalpy (ΔH) and its setback viscosity in HT-ripening grains, with the simultaneous decrements in the number and surface area proportions of smaller starch granules under the higher N fertilizer in combination with HT exposure. The effects of higher nitrogen fertilizer and HT exposure on total protein content and gluten composition of grains were additively increased. Hence, the topdressing of panicle N fertilizer exacerbated HT-induced deterioration in cooking and eating quality, rather than alleviating the negative impact of HT exposure on the palatability of cooked rice. [ABSTRACT FROM AUTHOR]

Background: Medical information has rapidly increased on the internet and has become one of the main targets of search engine use. However, medical information on the internet is subject to the problems of quality and accessibility, so ordinary users are unable to obtain answers to their medical questions conveniently. As a solution, researchers build medical question answering (QA) systems. However, research on medical QA in the Chinese language lags behind work on English-based systems. This lag is mainly due to the difficulty of constructing a high-quality knowledge base and the underutilization of medical corpora in the Chinese language. Results: This study developed an end-to-end solution to implement a medical QA system for the Chinese language with low cost and time. First, we created a high-quality medical knowledge graph from hospital data (electronic health/medical records) in a nearly automatic manner that trained a supervised model based on data labeled using bootstrapping techniques. Then, we designed a QA system based on a memory-based neural network and attention mechanism. Finally, we trained the system to generate answers from the knowledge base and a QA corpus on the internet. Conclusions: Bootstrapping and deep neural network techniques can construct a knowledge graph from electronic health/medical records with satisfactory precision and coverage. Our proposed context bridge mechanisms perform training with a variety of language features. Our QA system can achieve state-of-the-art quality in answering medical questions with constrained topics. As we evaluated, complex Chinese language processing techniques, such as segmentation and parsing, were not necessary for practice and complex architectures were not necessary to build the QA system. Lastly, we created an application using our method for internet QA usage. [ABSTRACT FROM AUTHOR]

Long noncoding RNAs (lncRNAs) play a significant role in cancer biology. This study aimed to determine the roles of lncRNAs in establishing the differences in clinical features between patients with papillary thyroid carcinoma (PTC) without Hashimoto's thyroiditis (HT) and patients with PTC and HT. In the present study, we detected the differentially expressed lncRNAs between tumor tissues of patients with PTC with or without HT through lncRNA microarrays. The data were verified and analyzed through qRT-PCR, cell viability, cell cycle and bioinformatics analyses. We found that 1031 lncRNAs and 1338 mRNAs were abnormally expressed in 5 tissue samples of PTC complicated with HT [PTC/HT (+)] compared with 5 samples of PTC without HT [PTC/HT (-)]. Gene Ontology and pathway analyses of the mRNAs suggested that several biological processes and pathways, particularly immune system processes, were induced in the PTC/HT (+) tissues. Twenty lncRNAs were verified in 31 PTC/HT (+) and 64 PTC/HT (-) specimens by qRT-PCR, and the results were consistent with the microarray data. Specifically, ENST00000452578, a downregulated lncRNA in PTC/HT(+), was negatively correlated with the tumor size. Cell viability assays revealed that ENST00000452578 could inhibit cell proliferation. Our results indicate that lncRNAs and mRNAs play an important role in establishing the different clinical characteristics between patients with PTC/HT(+) and patients with PTC/HT(-), and might provide new insights from the perspective of RNA for obtaining a further understanding of the clinical features related to PTC with HT. [ABSTRACT FROM AUTHOR]

Model-based force control for motion and force tracking faces significant challenges on real quadruped platforms due to the apparent model inaccuracies. In this paper, we present a multi-objective optimal torque control for hydraulic quadruped robots under significant model errors, such as non-modelable hydraulic components, linearization, disturbances, etc. More specifically, the centroidal dynamics are first modeled to project the dynamics of the floating-based whole-body behaviors to the centroidal frame. Model error compensation mechanisms are subsequently developed to track the reference motion of the CoM, torso, and foot-end trajectories, which are mapped into the joint space. Furthermore, a multi-objective optimal torque control scheme is formulated using quadratic programming (QP) to coordinate follow the reference motion and ground reaction forces simultaneously while satisfying all constraints. Finally, we present a series of simulations as well as experiments on a real hydraulic quadruped platform, EHbot. The results demonstrate that the proposed torque control scheme is robust to large model inaccuracies and improves the performance of the overall system. [ABSTRACT FROM AUTHOR]

This study aimed to evaluate the fermentation quality, bacterial community, and nitrate content of sorghum-sudangrass silage with two ensiling densities [550 kg fresh weight (FW)/m3 (low density, LD) and 650 kg FW/m3 (high density, HD)] stored at two temperatures [10°C (low temperature, LT) and 25°C (normal temperature, NT)] for 60 days. The fermentation parameters, microbial counts, bacterial community, nutritional composition, and nitrate and nitrite levels were assessed. The pH and ammonia nitrogen (N) in all silages were below 4.0 and 80 g/kg total N, respectively. Compared with LT treatments, NT treatments had lower pH and lactic acid (LA) bacteria and yeasts counts and contained higher LA and LA/acetic acid (LA/AA) (p < 0.05). The LT-LD contained more ammonia–N than LT-HD (p < 0.05) and had higher nitrate and lower nitrate degradation than other treatments (p < 0.05). Lactobacillus was the most dominant genus with all treatments (57.2–66.9%). The LA, LA/AA, and abundances of Pantoea , Pseudomonas , and Enterobacter in the silage negatively correlated with nitrate concentration and positively correlated with nitrate degradation (p < 0.05). Moreover, pH and ammonia–N were positively correlated with nitrate concentration and negatively correlated with nitrate degradation (p < 0.05). Overall, all silage had satisfactory fermentation quality, and the silage with HD and NT had better fermentation quality and higher nitrate degradation. The bacterial communities in all silages were dominated by Lactobacillus. The nitrate degradation during the fermentation process might be related to the fermentation quality and the activity of Pantoea , Pseudomonas , and Enterobacter in silage. [ABSTRACT FROM AUTHOR]

Crime risk prediction is helpful for urban safety and citizens’ life quality. However, existing crime studies focused on coarse-grained prediction, and usually failed to capture the dynamics of urban crimes. The key challenge is data sparsity, since that 1) not all crimes have been recorded, and 2) crimes usually occur with low frequency. In this paper, we propose an effective framework to predict fine-grained and dynamic crime risks in each road using heterogeneous urban data. First, to address the issue of unreported crimes, we propose a cross-aggregation soft-impute (CASI) method to deal with possible unreported crimes. Then, we use a novel crime risk measurement to capture the crime dynamics from the perspective of influence propagation, taking into consideration of both time-varying and location-varying risk propagation. Based on the dynamically calculated crime risks, we design contextual features (i.e., POI distributions, taxi mobility, demographic features) from various urban data sources, and propose a zero-inflated negative binomial regression (ZINBR) model to predict future crime risks in roads. The experiments using the real-world data from New York City show that our framework can accurately predict road crime risks, and outperform other baseline methods. [ABSTRACT FROM AUTHOR]

Aftershocks can cause disasters again after mainshocks, which result in threat to life and economic loss. In order to avoid secondary disasters, it is necessary to predict whether aftershocks would happen in a given region. There have been studies using different features and methods to predict aftershocks spatial distribution. However, it is still unclear which are more important for aftershock prediction, input features or models; which type of features is more predictive for the prediction task. In this paper, we predict aftershock spatial distribution by combining different types of features and applying different machine learning methods. We introduce five different types of features and combine them together for prediction: the stress change sensors, their logarithmic values, the physical quantities, the magnitude of mainshocks, and the distance between the grid cell and the epicenter of mainshocks. We train different classifiers: Naive Bayes, Support Vector Machine,Gradient Boosting Decision Tree, k-Nearest Neighbors, Logistic regression, and DMAP (a Deep Neural Network model). Based on the 62,811 aftershocks of 171 distinct mainshocks in the past about 40 years in China, we conduct comprehensive experiments and analyses. We find that features play a more important role for this prediction task. Using the same feature type, different classifiers obtain quite similar performance. With different features, the same model performs differently. Taking the combined features as input, we achieve the state-of-the-art performance, with an AUC of 0.9530, about 4% higher than that of DeVries et al., showing the superiority of the combined features. Among all the features, adding the distance to the stress change sensors contributes the most to improve the prediction performance. In addition, it is found that the model prediction performance varies in terms of the time spans after mainshocks and the aftershock magnitudes. [ABSTRACT FROM AUTHOR]

Smartphones are changing humans’ lifestyles. Mobile applications (apps) on smartphones serve as entries for users to access a wide range of services in our daily lives. The apps installed on one’s smartphone convey lots of personal information, such as demographics, interests, and needs. This provides a new lens to understand smartphone users. However, it is difficult to compactly characterize a user with his/her installed app list. In this article, a user representation framework is proposed, where we model the underlying relations between apps and users with Boolean matrix factorization (BMF). It builds a compact user subspace by discovering basic components from installed app lists. Each basic component encapsulates a semantic interpretation of a series of special-purpose apps, which is a reflection of user needs and interests. Each user is represented by a linear combination of the semantic basic components. With this user representation framework, we use supervised and unsupervised learning methods to understand users, including mining user attributes, discovering user groups, and labeling semantic tags to users. Extensive experiments were conducted on three data subsets from a large-scale real-world dataset for evaluation, each consisting of installed app lists from over 10 000 users. The results demonstrated the effectiveness of our user representation framework. [ABSTRACT FROM AUTHOR]

Solving the optimization problem to approach a Nash Equilibrium point plays an important role in imperfect information games, e.g., StarCraft and poker. Neural Fictitious Self-Play (NFSP) is an effective algorithm that learns approximate Nash Equilibrium of imperfect-information games from purely self-play without prior domain knowledge. However, it needs to train a neural network in an off-policy manner to approximate the action values. For games with large search spaces, the training may suffer from unnecessary exploration and sometimes fails to converge. In this paper, we propose a new Neural Fictitious Self-Play algorithm that combines Monte Carlo tree search with NFSP, called MC-NFSP, to improve the performance in real-time zero-sum imperfect-information games. With experiments and empirical analysis, we demonstrate that the proposed MC-NFSP algorithm can approximate Nash Equilibrium in games with large-scale search depth while the NFSP can not. Furthermore, we develop an Asynchronous Neural Fictitious Self-Play framework (ANFSP). It uses asynchronous and parallel architecture to collect game experience and improve both the training efficiency and policy quality. The experiments with th e games with hidden state information (Texas Hold’em), and the FPS (firstperson shooter) games demonstrate effectiveness of our algorithms. [ABSTRACT FROM AUTHOR]