Search

Long non‐coding RNAs (lncRNAs) and microRNAs (miRNAs) are two typical types of non‐coding RNAs that interact and play important regulatory roles in many animal organisms. Exploring the unknown interactions between lncRNAs and miRNAs contributes to a better understanding of their functional involvement. Currently, studying the interactions between lncRNAs and miRNAs heavily relies on laborious biological experiments. Therefore, it is necessary to design a computational method for predicting lncRNA–miRNA interactions. In this work, we propose a method called MPGK‐LMI, which utilizes a graph attention network (GAT) to predict lncRNA–miRNA interactions in animals. First, we construct a meta‐path similarity matrix based on known lncRNA–miRNA interaction information. Then, we use GAT to aggregate the constructed meta‐path similarity matrix and the computed Gaussian kernel similarity matrix to update the feature matrix with neighbourhood information. Finally, a scoring module is used for prediction. By comparing with three state‐of‐the‐art algorithms, MPGK‐LMI achieves the best results in terms of performance, with AUC value of 0.9077, AUPR of 0.9327, ACC of 0.9080, F1‐score of 0.9143 and precision of 0.8739. These results validate the effectiveness and reliability of MPGK‐LMI. Additionally, we conduct detailed case studies to demonstrate the effectiveness and feasibility of our approach in practical applications. Through these empirical results, we gain deeper insights into the functional roles and mechanisms of lncRNA–miRNA interactions, providing significant breakthroughs and advancements in this field of research. In summary, our method not only outperforms others in terms of performance but also establishes its practicality and reliability in biological research through real‐case analysis, offering strong support and guidance for future studies and applications. [ABSTRACT FROM AUTHOR]

Background: Impaired cognitive ability is one of the most frequently reported neuropsychiatric symptoms in the post-COVID phase among patients. It is unclear whether this condition is related to structural or functional brain changes. Purpose: In this study, we present a multimodal magnetic resonance imaging study of 36 post-COVID patients and 36 individually matched controls who had a mild form of severe acute respiratory syndrome coronavirus 2 (SARS CoV-2) infection from March 2020 to February 2022. This study aimed to investigate structural and functional brain alterations and their correlation with post-COVID symptoms and neurocognitive functions. Materials and methods: The study protocol comprised an assessment of physical fatigue [Fatigue Severity Scale (FSS)], mental fatigue (Mental Fatigue Scale (MFS)], depression [Montgomery Asberg Depression Rating Scale (MADRS)], anxiety [Hospital Anxiety and Depression Scale (HAD)], post-COVID Symptoms Severity Score, and neurocognitive status [Repeatable Battery for the Assessment of Neuropsychological Status Update (RBANS)]. The magnetic resonance imaging protocol included morphological sequences, arterial spin labeling (ASL) and dynamic susceptibility contrast-enhanced (DSC) perfusion, diffusion tensor imaging (DTI), and resting-state functional magnetic resonance imaging (fMRI) sequences. Using these protocols, the assessments of macrostructural abnormalities, perfusion, gray matter density, white matter integrity, and brain connectivity were performed. Results: Post-COVID patients had higher levels of physical fatigue, mental fatigue, depression, and anxiety than controls and showed cognitive impairment in all the RBANS domains except in Visuospatial/Construction. The subjective mental fatigue correlated with objective impaired cognitive ability in the RBANS test, particularly in the Attention domain. There were no differences between patients and controls regardingmacrostructural abnormalities, regional volumes, regional perfusion metrics, gray matter density, or DTI parameters. We observed a significant positive correlation between RBANS Total Scale Index score and gray matter volume in the right superior/middle-temporal gyrus (p < 0.05) and a significant negative correlation between the white matter integrity and post-COVID symptoms (p < 0.05) in the same area. The connectivity differences were observed between patients and controls in a few regions, including the right middle frontal gyrus, an important area of convergence of the dorsal and ventral attention networks. We also noted a positive correlation between post-COVID symptoms and increased connectivity in the right temporoparietal junction, which is part of the ventral attention system. Conclusion: In non-hospitalized subjects with post-COVID, we did not find any structural brain changes or changes in perfusion, compared to controls. However, we noted differences in connectivity within an important area for attention processes, which may be associated with post-COVID brain fog. [ABSTRACT FROM AUTHOR]

Over the past decade, extensive research has been conducted on employing deep learning techniques to detect visual defects in structural facades during inspection. Although these models have shown accuracy in identifying defects from visual data, they encounter limitations in practical applications. This includes uncertainty with data that fall outside the trained distribution and their lack of explanation of detection results. In addition, owing to their extensive parameters, these models require substantial computational resources, which is impractical for visual inspection. These limitations impede immediate defect checking and misjudgment of the deep learning model. The study aims to address these challenges by optimizing a deep-learning-based defect recognition model using a selective layer attention network (SAN). This utilizes a selective feature extraction method to provide essential visual defect information through feature maps within a deep learning model. SAN can effectively represent defect information from building surface images across each layer using the gradient-weighted class activation-mapping visualization technique. These findings demonstrate that the SAN-based model offers clear visual information while significantly reducing the usage of computational resources by 90% compared with the original network, maintaining an equivalent performance level. [ABSTRACT FROM AUTHOR]

This paper presents an automatic speech recognition (ASR) system developed for the Indian language Odia. In recent years, deep learning models have been used widely to develop ASR systems in various languages and domains. These models demand huge training resources, primarily annotated continuous speech utterances collected from various speakers. However, sufficient speech corpus is not available in many Indian languages. This paper explores the effectiveness of data augmentation in overcoming data scarcity in the Odia ASR task. The baseline system is developed using BiLSTM and the Seq2Seq framework. Next, a portion of the training data is selected based on phonetic richness, and certain augmentation techniques like pitch alteration and time stretching are applied. The augmented data is used along with the actual training data, and a substantial performance improvement is observed. The effectiveness of the attention mechanism in Odia ASR is also explored. When the system is trained through an attention layer embedded with the baseline BiLSTM model, it outperforms the baseline model and existing Odia ASR systems in the literature. [ABSTRACT FROM AUTHOR]

Background: In laparoscopic liver surgery, accurately predicting the displacement of key intrahepatic anatomical structures is crucial for informing the doctor's intraoperative decision‐making. However, due to the constrained surgical perspective, only a partial surface of the liver is typically visible. Consequently, the utilization of non‐rigid volume to surface registration methods becomes essential. But traditional registration methods lack the necessary accuracy and cannot meet real‐time requirements. Purpose: To achieve high‐precision liver registration with only partial surface information and estimate the displacement of internal liver tissues in real‐time. Methods: We propose a novel neural network architecture tailored for real‐time non‐rigid liver volume to surface registration. The network utilizes a voxel‐based method, integrating sparse convolution with the newly proposed points of interest (POI) linear attention module. POI linear attention module specifically calculates attention on the previously extracted POI. Additionally, we identified the most suitable normalization method RMSINorm. Results: We evaluated our proposed network and other networks on a dataset generated from real liver models and two real datasets. Our method achieves an average error of 4.23 mm and a mean frame rate of 65.4 fps in the generation dataset. It also achieves an average error of 8.29 mm in the human breathing motion dataset. Conclusions: Our network outperforms CNN‐based networks and other attention networks in terms of accuracy and inference speed. [ABSTRACT FROM AUTHOR]

Image restoration is an inevitable pre-processing step in most satellite imaging applications. The satellite imaging modality such as Synthetic Aperture Radar (SAR) is prone to speckle distortions due to constructive and destructive interference of the probing signals. Speckles being data correlated and multiplicative, their reduction is not so trivial. Since speckles are not purely noise interventions, a blind reduction process leads to spurious analysis at the later stages. Moreover, the image details are liable to get compromised during such a noise reduction process. An attention-based deep image prior (DIP) model with U-Net architecture has been proposed in this work to carefully address these setbacks. The attention block is used to scale the features extracted from the encoder, and they are concatenated with the features from the decoder to obtain both low- and high-level features. The attention module incorporated in the model helps to extract significant complex structures in SAR images. Further, the DIP model duly respects the noise distribution of speckles while performing the despeckling task. Various synthetic, natural, aerial, and satellite images are subjected to the testing and verification process, and the results obtained are in favor of the proposed model. The quantitative analysis carried out using various statistical metrics in this study also reveals the restoration ability of the proposed method in terms of both despeckling and structure preservation. [ABSTRACT FROM AUTHOR]

Detecting wafer map anomalies is crucial for preventing yield loss in semiconductor fabrication, although intricate patterns and resource-intensive labeled data prerequisites hinder precise deep-learning segmentation. This paper presents an innovative, unsupervised method for detecting pixel-level anomalies in wafer maps. It utilizes an efficient dual attention module with a knowledge distillation network to learn defect distributions without anomalies. Knowledge transfer is achieved by distilling information from a pre-trained teacher into a student network with similar architecture, except an efficient dual attention module is incorporated atop the teacher network’s feature pyramid hierarchies, which enhances feature representation and segmentation across pyramid hierarchies that selectively emphasize relevant and discard irrelevant features by capturing contextual associations in positional and channel dimensions. Furthermore, it enables student networks to acquire an improved knowledge of hierarchical features to identify anomalies across different scales accurately. The dissimilarity in feature pyramids acts as a discriminatory function, predicting the likelihood of an abnormality, resulting in highly accurate pixel-level anomaly detection. Consequently, our proposed method excelled on the WM-811K and MixedWM38 datasets, achieving AUROC, AUPR, AUPRO, and F1-Scores of (99.65%, 99.35%), (97.31%, 92.13%), (90.76%, 84.66%) respectively, alongside an inference speed of 3.204 FPS, showcasing its high precision and efficiency. [ABSTRACT FROM AUTHOR]

Copyright of Journal of Mine Automation is the property of Industry & Mine Automation Editorial Department 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.)

This paper proposes a novel approach that leverages a hybrid deep learning framework called the Squirrel Search-optimized Attention-Deep Recurrent Neural Network (SS-ADRNN) to optimize the management of charging stations, ensuring efficient resource allocation while safeguarding user data and minimizing operational costs. The SS-ADRNN model incorporates squirrel search optimization, which is inspired by the foraging behaviour of squirrels, to dynamically adjust charging station operations based on environmental conditions and demand patterns. Additionally, attention mechanisms are employed to prioritize relevant input features, enabling the model to focus on critical information during decision-making processes. Deep recurrent neural networks (RNNs) are utilized to capture temporal dependencies in charging station data, allowing for more accurate predictions and adaptive control strategies. Experimental evaluations demonstrate the effectiveness and feasibility of the proposed SS-ADRNN-based approach in real-world scenarios. The results showcase significant improvements in the detection of malicious traffic and cost minimization compared to traditional charging station management methods. Overall, this research contributes to advancing the field of intelligent charging station optimization, offering a robust and adaptable solution for EV and UAV charging infrastructures that prioritize both security and operational efficiency.

The existing coal and gangue segmentation and recognition technology has a large number of parameters, slow classification speed, and low recognition accuracy. The YOLOv5-seg model is prone to losing texture details and grayscale feature information on the image surface during up and down sampling operations, which reduces the efficiency of coal and gangue recognition. The YOLOv5-seg model overly focuses on global features during training, while neglecting the locally significant regions and features that are crucial for coal and gangue recognition. In order to solve the above problems, a coal and gangue segmentation and recognition method based on YOLOv5-SEDC model is proposed. Firstly, the method receives an image containing the shape information of coal and gangue, and uses the backbone network for feature extraction to generate a feature map. The method integrates the SENet module into the YOLOv5-seg model to preserve the texture details and grayscale features of coal and gangue surfaces, avoiding information loss caused by down sampling. The method adopts a dilated convolution strategy with different dilation rates instead of traditional convolution kernels. It not only expands the receptive field of the model, but also effectively reduces the number of model parameters. Finally, the segmentation detection head finely processes the fused features to achieve precise segmentation and recognition of coal and gangue. A coal and gangue image acquisition experimental platform is established at the actual coal and gangue sorting site of Daliuta Coal Mine. The ablation experiment results show that the accuracy of coal and gangue recognition of YOLOv5-SEDC model is improved by an average of 1.3% compared to YOLOv5-seg model. The parameter quantity is reduced by 0.7×106, and the detection speed is increased by 1.4 frames/s. The comparative experimental results show the following points. ① The accuracy of the YOLOv5-SEDC model is improved by 10.7%, 2.7%, 1.9% compared to the YOLOv3-tiny, YOLOv5-seg, and Mask-RCNN models, respectively, reaching 95.8%. ② The recall rate of the YOLOv5-SEDC model has increased by 3.0%, 2.1%, and 0.9% compared to the YOLOv3-tiny, YOLOv5-seg, and Mask-RCNN models, respectively, reaching 89.1%. ③ The mAP of the YOLOv5-SEDC model has increased by 6.4%, 6.3%, and 1.8% compared to the YOLOv3-tiny, YOLOv5-seg, and Mask-RCNN models, respectively, reaching 95.5%. ④ The F1 value of the YOLOv5-SEDC model has increased by 5.2%, 4.2%, 2.1% compared to the YOLOv3-tiny, YOLOv5-seg, and Mask-RCNN models, respectively, reaching 92.2%. ⑤ The detection speed of the YOLOv5-SEDC model is reduced by 1.9, 1.4, and 2.7 frames/s compared to the YOLOv3-tiny, YOLOv5-seg, and Mask-RCNN models, respectively. The visualization results show that the YOLOv5-SEDC model has higher detection accuracy for coal and gangue than the YOLOv5-seg and Mask-RCNN models. It indicates that the YOLOv5-SEDC model has good performance in coal gangue segmentation and recognition.

Recommender system has attracted increasing attentions of many service providers, as it plays an important role in helping user filter irrelevant information. As an important application in daily life, point-of-interest (POI) recommendation system has become a powerful tool for assisting users make travel decisions, by modeling the impact of external factors on user behavior, such as time, geographical location, to predict future check-ins. However, the influence of intention, an important internal factor, on user check-in behavior has not been well explored. Existing research lacks methods for intention representing learning in POI recommendation, and has not explore the relationship between intention prediction and check-in behavior prediction. Motivated by this, this paper develops a novel sequential-hierarchical attention neural network based recommendation method (SH-Rec), which learns the hierarchy association of intention and sequential dependency of behavior and its interactions to improve user representation in POI recommendation. The main idea of the proposed SH-Rec is to describe user intentions from both hierarchical and sequential aspects using historical check-in sequence and side information, such as POI category attributes. Specifically, we design a novel sequential-hierarchical attention network to model the interaction of hierarchical intention features and sequential behavior features, by stacking several LSTM and self-attention layers. Besides, we model user’s behavior patterns by extracting sequential preference features using memory network. To utilize the contribution of intention learning in recommendation, we propose a weighted optimization function by employing multi-task learning strategy, to migrate knowledge from intention prediction to check-in prediction. Extensive experiments over three real-world datasets evaluate the better performance of the proposed model than the state-of-the-art methods in terms of various evaluation metrics. A series of ablation experiments and parameter experiments verify the better robustness and stability of the proposed model. [ABSTRACT FROM AUTHOR]

Abstract Defect detection in complex background is a critical issue. To address this issue, this paper proposes the mixture attention mechanism cascade network, in which the new channel attention network is linked with the spatial attention network to create an effective mixed attention network that takes advantage of their respective advantages, adaptively suppresses background noise, and highlights defect features. To ensure the efficiency and effectiveness of effective mixed attention network, the new channel attention network splices the output features of the global average pooling layer and the global maximum pooling layer and then sends the spliced features into a shared network, which is a one‐dimensional convolutional network, and uses cross‐channel interaction for fusion. Furthermore, in order to provide more discriminative feature representation, the authors extract the intermediate features of the region proposal network and input them into effective mixed attention network. Finally, the cascade head is used to refine the predicted bounding box to achieve high‐quality defect location. To demonstrate the superiority and usefulness of this method, it is compared to the latest method using widely used PCB and NEU data sets. A large number of trials demonstrate that this strategy outperforms other methods for detecting defects in complicated backgrounds.

The pervasive nature of short-form video platforms has seamlessly integrated into daily routines, yet it is important to recognize their potential adverse effects on both physical and mental health. Prior research has identified a detrimental impact of excessive short-form video consumption on attentional behavior, but the underlying neural mechanisms remain unexplored. In the current study, we aimed to investigate the effect of short-form video use on attentional functions, measured through the attention network test (ANT). A total of 48 participants, consisting of 35 females and 13 males, with a mean age of 21.8 years, were recruited. The mobile phone short video addiction tendency questionnaire (MPSVATQ) and self-control scale (SCS) were conducted to assess the short video usage behavior and self-control ability. Electroencephalogram (EEG) data were recorded during the completion of the ANT task. The correlation analysis showed a significant negative relationship between MPSVATQ and theta power index reflecting the executive control in the prefrontal region (r = -0.395, p = 0.007), this result was not observed by using theta power index of the resting-state EEG data. Furthermore, a significant negative correlation was identified between MPSVATQ and SCS outcomes (r = -0.320, p = 0.026). These results suggest that an increased tendency toward mobile phone short video addiction could negatively impact self-control and diminish executive control within the realm of attentional functions. This study sheds light on the adverse consequences stemming from short video consumption and underscores the importance of developing interventions to mitigate short video addiction. [ABSTRACT FROM AUTHOR]

While significant strides have been made in predicting neoepitopes that trigger autologous CD4+ T cell responses, accurately identifying the antigen presentation by human leukocyte antigen (HLA) class II molecules remains a challenge. This identification is critical for developing vaccines and cancer immunotherapies. Current prediction methods are limited, primarily due to a lack of high-quality training epitope datasets and algorithmic constraints. To predict the exogenous HLA class II-restricted peptides across most of the human population, we utilized the mass spectrometry data to profile >223 000 eluted ligands over HLA-DR, -DQ, and -DP alleles. Here, by integrating these data with peptide processing and gene expression, we introduce HLAIImaster, an attention-based deep learning framework with adaptive domain knowledge for predicting neoepitope immunogenicity. Leveraging diverse biological characteristics and our enhanced deep learning framework, HLAIImaster is significantly improved against existing tools in terms of positive predictive value across various neoantigen studies. Robust domain knowledge learning accurately identifies neoepitope immunogenicity, bridging the gap between neoantigen biology and the clinical setting and paving the way for future neoantigen-based therapies to provide greater clinical benefit. In summary, we present a comprehensive exploitation of the immunogenic neoepitope repertoire of cancers, facilitating the effective development of "just-in-time" personalized vaccines. [ABSTRACT FROM AUTHOR]

Defect detection in complex background is a critical issue. To address this issue, this paper proposes the mixture attention mechanism cascade network, in which the new channel attention network is linked with the spatial attention network to create an effective mixed attention network that takes advantage of their respective advantages, adaptively suppresses background noise, and highlights defect features. To ensure the efficiency and effectiveness of effective mixed attention network, the new channel attention network splices the output features of the global average pooling layer and the global maximum pooling layer and then sends the spliced features into a shared network, which is a one‐dimensional convolutional network, and uses cross‐channel interaction for fusion. Furthermore, in order to provide more discriminative feature representation, the authors extract the intermediate features of the region proposal network and input them into effective mixed attention network. Finally, the cascade head is used to refine the predicted bounding box to achieve high‐quality defect location. To demonstrate the superiority and usefulness of this method, it is compared to the latest method using widely used PCB and NEU data sets. A large number of trials demonstrate that this strategy outperforms other methods for detecting defects in complicated backgrounds. [ABSTRACT FROM AUTHOR]

Given an object of interest, visual navigation aims to reach the object's location based on a sequence of partial observations. To this end, an agent needs to (1) acquire specific knowledge about the relations of object categories in the world during training and (2) locate the target object based on the pre-learned object category relations and its trajectory in the current unseen environment. In this article, we propose a Category Relation Graph (CRG) to learn the knowledge of object category layout relations and a Temporal-Spatial-Region attention (TSR) architecture to perceive the long-term spatial-temporal dependencies of objects, aiding navigation. We establish CRG to learn prior knowledge of object layout and deduce the positions of specific objects. Subsequently, we propose the TSR architecture to capture relationships among objects in temporal, spatial, and regions within observation trajectories. Specifically, we implement a Temporal attention module (T) to model the temporal structure of the observation sequence, implicitly encoding historical moving or trajectory information. Then, a Spatial attention module (S) uncovers the spatial context of the current observation objects based on CRG and past observations. Last, a Region attention module (R) shifts the attention to the target-relevant region. Leveraging the visual representation extracted by our method, the agent accurately perceives the environment and easily learns a superior navigation policy. Experiments on AI2-THOR demonstrate that our CRG-TSR method significantly outperforms existing methods in both effectiveness and efficiency. The supplementary material includes the code and will be publicly available. [ABSTRACT FROM AUTHOR]

Several studies on attention-deficit hyperactivity disorder (ADHD) have suggested a developmental sequence of brain changes: subcortico-subcortical connectivity in children, evolving to subcortico-cortical in adolescence, and culminating in cortico-cortical connectivity in young adulthood. This study hypothesized that children with ADHD would exhibit decreased functional connectivity (FC) between the cortex and striatum compared to adults with ADHD, who may show increased FC in these regions. Seventy-six patients with ADHD (26 children, 26 adolescents, and 24 adults) and 74 healthy controls (25 children, 24 adolescents, and 25 adults) participated in the study. Resting state magnetic resonance images were acquired using a 3.0 T Philips Achieva scanner. The results indicated a gradual decrease in the number of subcategories representing intelligence quotient deficits in the ADHD group with age. In adulthood, the ADHD group exhibited lower working memory compared to the healthy control group. The number of regions showing decreased FC from the cortex to striatum between the ADHD and control groups reduced with age, while regions with increased FC from the default mode network and attention network in the ADHD group increased with age. In adolescents and adults, working memory was positively associated with brain activity in the postcentral gyrus and negatively correlated with ADHD clinical symptoms. In conclusion, the findings suggest that intelligence deficits in certain IQ subcategories may diminish as individuals with ADHD age. Additionally, the study indicates an increasing anticorrelation between cortical and subcortical regions with age in individuals with ADHD. [ABSTRACT FROM AUTHOR]

BackgroundImpaired cognitive ability is one of the most frequently reported neuropsychiatric symptoms in the post-COVID phase among patients. It is unclear whether this condition is related to structural or functional brain changes.PurposeIn this study, we present a multimodal magnetic resonance imaging study of 36 post-COVID patients and 36 individually matched controls who had a mild form of severe acute respiratory syndrome coronavirus 2 (SARS CoV-2) infection from March 2020 to February 2022. This study aimed to investigate structural and functional brain alterations and their correlation with post-COVID symptoms and neurocognitive functions.Materials and methodsThe study protocol comprised an assessment of physical fatigue [Fatigue Severity Scale (FSS)], mental fatigue (Mental Fatigue Scale (MFS)], depression [Montgomery Asberg Depression Rating Scale (MADRS)], anxiety [Hospital Anxiety and Depression Scale (HAD)], post-COVID Symptoms Severity Score, and neurocognitive status [Repeatable Battery for the Assessment of Neuropsychological Status Update (RBANS)]. The magnetic resonance imaging protocol included morphological sequences, arterial spin labeling (ASL) and dynamic susceptibility contrast-enhanced (DSC) perfusion, diffusion tensor imaging (DTI), and resting-state functional magnetic resonance imaging (fMRI) sequences. Using these protocols, the assessments of macrostructural abnormalities, perfusion, gray matter density, white matter integrity, and brain connectivity were performed.ResultsPost-COVID patients had higher levels of physical fatigue, mental fatigue, depression, and anxiety than controls and showed cognitive impairment in all the RBANS domains except in Visuospatial/Construction. The subjective mental fatigue correlated with objective impaired cognitive ability in the RBANS test, particularly in the Attention domain. There were no differences between patients and controls regarding macrostructural abnormalities, regional volumes, regional perfusion metrics, gray matter density, or DTI parameters. We observed a significant positive correlation between RBANS Total Scale Index score and gray matter volume in the right superior/middle-temporal gyrus (p < 0.05) and a significant negative correlation between the white matter integrity and post-COVID symptoms (p < 0.05) in the same area. The connectivity differences were observed between patients and controls in a few regions, including the right middle frontal gyrus, an important area of convergence of the dorsal and ventral attention networks. We also noted a positive correlation between post-COVID symptoms and increased connectivity in the right temporoparietal junction, which is part of the ventral attention system.ConclusionIn non-hospitalized subjects with post-COVID, we did not find any structural brain changes or changes in perfusion, compared to controls. However, we noted differences in connectivity within an important area for attention processes, which may be associated with post-COVID brain fog.

Convolutional neural networks （CNNs） struggle to efficiently capture contextual information of power equipment such as arresters and GIS inlet casings due to their limited receptive fields， thereby affecting detection performance. To address this issue， the paper introduces a Transformer-based voxel-graph attention network. Local attention and dilated attention mechanisms are proposed to respectively capture short-range and long-range feature correlations within image volume pixels， effectively expanding the attention scope while keeping computational costs unchanged. Additionally， submanifold voxel modules and sparse voxel modules are designed to extract feature information from non-empty voxel positions and empty voxel positions， respectively. Finally， through comparative analysis with mainstream models on the general datasets Waymo and KITTI， as well as on an image dataset from a transmission and transformation area in Yunnan Province， the superior performance of the proposed model in detecting power equipment is demonstrated.

The accurate segmentation of the tissue layers of Optical Coherence Tomography (OCT) esophageal images has vital guiding significance for esophageal diseases research and computer-aided diagnosis. Existing automatic segmentation algorithms based on fully convolutional networks due to the lack of context information resulting in inaccurate segmentation performance. In this paper, we propose a hybrid attention network (HAN) to address this problem. The proposed framework includes an encoder attention module, a channel attention module and a tailored pyramid pooling module (TPPM). The encoder attention module can make up for the loss of image details due to down-sampling operation. The channel attention module is designed to selectively emphasize interdependent channel by integrating associated features among all channel maps. The tailored pyramid pooling module is designed to enhance the spatial contextual information. The combination of channel and spatial contextual information helps to boost feature discriminability, the framework has high segmentation accuracy. Through experiments, it can be verified that the network outperforms several advanced deep learning frameworks in segmentation. The clinical diagnostic potential of HAN for eosinophilic esophagitis (EOE), an esophageal disease, is also demonstrated in the experiment. [ABSTRACT FROM AUTHOR]

Copyright of Zhejiang Electric Power is the property of Zhejiang Electric Power Editorial Office 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.)

Knowledge Graphs can not contain all the knowledge during the construction process, so needs to be completed to enhance its integrity. In real knowledge graphs, different relationships often show apparent long-tail distributions, i.e., many relationships have only a small number of entity pairs. Therefore, it is an urgent need to study few-shot knowledge graph completion. Existing methods generally complete the knowledge graph by learning representations of entities and relationships, but ignore the impact of the similarity of neighbor relations between triple entity pairs on completion. In this paper, we propose an implicit relational attention network to address this limitation. First, we propose a heterogeneous entity and relational encoder to mine one-hop neighbor information and enhance entity and relational representations through attention mechanism and convolution. Next, we propose an implicit relationship aware encoder to mine the neighbor relationship similarity information of triple entity pairs and obtain the triple dynamic relationship representation. Then we propose an adaptive relationship fusion network, which fuses the triple dynamic relationship representation and the original information of the neighbor relationship similarity of entity pairs, enhances the relationship representation of the query set to the reference set, so as to improve the accuracy of the few-shot knowledge graph completion. On two benchmark datasets, by comparing with well-known completion methods, the experimental results show that the proposed method achieves very competitive performance. [ABSTRACT FROM AUTHOR]

Two-dimensional phase unwrapping (2-D PU) is vital for reconstructing Earth's surface topography and displacement from interferometric synthetic aperture radar (InSAR) data. Conventional algorithms rely on the postulate, but this assumption is often insufficient due to abrupt topographic changes and severe noise. To address this challenge, our research proposes a novel approach utilizing deep convolutional neural networks inspired by the U-Net architecture to estimate phase gradient information. Our approach involves downsampling the input data to extract crucial features, followed by upsampling to restore spatial resolution. We incorporate two attention mechanisms—feature pyramid attention (FPA) and global attention upsample (GAU)—and a residual structure in the network's structure. Thus, we construct ResDANet (residual and dual attention net). We rigorously train ResDANet utilizing simulated datasets and employ an L1-norm objective function to minimize the disparity between unwrapped phase gradients and those calculated by ResDANet, yielding the final 2-D PU results. The network is rigorously trained using two distinct training strategies and encompassing three types of simulated datasets. ResDANet exhibits excellent robust performance and efficiency on simulated data and real data, such as China's Three Gorges and an Italian volcano. [ABSTRACT FROM AUTHOR]

Debugging is an intellectually rich and a challenging process when learning a programming language. This process is important for increasing the quality of the program and making it functional. Debugging, by its nature, is thought to be a practice with a state of focus and concentration. This study explored whether the debugging performance could be predicted by attention networks of students studying at the IT department of vocational high schools. 108 vocational high school IT department students participated in the research. First, students' attention levels were determined by utilizing the Attention Network Test. Second, a Debugging Performance Test was administered to determine their debugging performance. The analysis result indicated that the general attention network statistically predicted the debugging performance but the effect size was found to be small. Furthermore, the singular and dual interactions of the alerting attention network, the orienting attention network, and the executive attention network did not predict debugging performances. Yet, the interactions of the three attention networks on predicting the debugging performances were found to be statistically significant. These findings were discussed together with the existing studies in the literature and suggestions to obtain more in-depth information regarding these results were provoked. [ABSTRACT FROM AUTHOR]

In this work, DDTraMIS architecture based on vision has been designed for medical image segmentation for different medical imaging such as MRI and CT-scan. This methodology contributes novel hybrid features extracted with a bi-directional attention-based transformer encoder-decoder network along with all stage features fused with an approximation fusing algorithm. The novelty is to combine Convolution neural networks (CNNs) and shift-invariant methods to develop hybrid features. For the verification of this novel network experiment, three different datasets, such as the ACDC, LiTS, and BraTS datasets have been used. Performance analysis has been conducted using the dice similarity coefficient (DSC) and the Hausdorff distance (HD) metrics. The proposed architecture accomplished DSC values of 92.80, 96.45, 72.80, and 73.12 for ACDC, LiTS Liver, LiTS tumor, and BraTS tumor segmentation analysis, respectively. Similarly, HD metric numeric values are 8.53, 12.84, 9.69, and 10.38 for ACDC, LiTS Liver, LiTS tumor, and BraTS tumor segmentation analysis, respectively. With advantageous performance for highly features incorporated, this method learned from tiny to extensive scaling information. Comparative and quantitative analysis has proven its superior performance and effective segmenting architecture. [ABSTRACT FROM AUTHOR]

The next Point-of-Interest (POI) recommendation, in recent years, has attracted an extensive amount of attention from the academic community. RNN-based methods cannot establish effective long-term dependencies among the input sequences when capturing the user's motion patterns, resulting in inadequate exploitation of user preferences. Besides, the majority of prior studies often neglect high-order neighborhood information in users' check-in trajectory and their social relationships, yielding suboptimal recommendation efficacy. To address these issues, this paper proposes a novel Double-Layer Attention Network model, named DLAN. Firstly, DLAN incorporates a multi-head attention module that can combine first-order and high-order neighborhood information in user check-in trajectories, thereby effectively and parallelly capturing both long- and short-term preferences of users and overcoming the problem that RNN-based methods cannot establish long-term dependencies between sequences. Secondly, this paper designs a user similarity weighting layer to measure the influence of other users on the target users leverage the social relationships among them. Finally, comprehensive experiments are conducted on user check-in data from two cities, New York (NYC) and Tokyo (TKY), and the results demonstrate that DLAN achieves a performance in Accuracy and Mean Reverse Rank enhancement by 8.07% -36.67% compared to the state-of-the-art method. Moreover, to investigate the effect of dimensionality and the number of heads of the multi-head attention mechanism on the performance of the DLAN model, we have done sufficient sensitivity experiments. [ABSTRACT FROM AUTHOR]

This study examined the impact of brief mindfulness meditation (BMM) training on attention function and dispositional mindfulness in young males. 126 male participants aged 18–26 from the security industry were recruited, with 66 participants (M = 22.84, SD = 2.41) undergoing 4-week mindfulness meditation training and 60 participants (M = 23.07, SD = 2.29) in the active control group. The intervention was integrated into the participants' schedules. Measures included Five Facets Mindfulness Questionnaires (FFMQ), concentration and assignment attention tasks, Attention Network Test (ANT), and saliva cortisol concentration. Findings indicate that brief mindfulness meditation training led to significant improvements in participants' FFMQ scores), with marginally significant enhancements in the executive control network. However, it had no discernible effect on alertness and orientation networks. Additionally, brief mindfulness meditation training enhanced attention allocation to light stimulation and prolonged individual attention. Surprisingly, there was no observed decrease in saliva cortisol concentration among meditation training participants. However, this study did not find a decrease in saliva cortisol concentration in the brief mindfulness meditation group. In conclusion, this study highlights the potential of a 4-week brief mindfulness meditation training program to enhance dispositional mindfulness and specific aspects of attention function in young men, offering practical insights into the benefits of mindfulness meditation practices for this demographic.

In recent years, deep learning algorithms, particularly convolutional neural networks, have significantly improved the performance of the hyperspectral image (HSI) classification. However, due to the high dimensionality of HSI and limited training samples, the deep neural network causes model overfitting. In addition, considering all the bands of HSI datasets equally for feature learning and being unable to distinguish between the edge and the center pixels of a neighborhood reduces classification accuracy. Thus, in this article, we propose an end-to-end deep spectral–spatial residual attention network (DSSpRAN) motivated by the attention mechanism of the human visual system for HSI classification. The DSSpRAN considers input HSI data as a 3-D cube instead of using dimensionality reduction methods. The proposed model simultaneously incorporates spectral and spatial features by considering a spectral residual attention network (SRAN) and a spatial residual attention network (SpRAN). In SRAN, the weights are assigned and learned adaptively to select essential features from each band. The SpRAN enhances the importance of classifying each nearby pixel to the center pixel. It assigns the same label as that of the center pixel to the surrounding pixels, thus limiting pixels with different labels. The proposed method has been evaluated on five different datasets to prove the state of the art for various land use land cover scenarios. A comprehensive qualitative and quantitative analysis of the results shows that the proposed method significantly outperforms other state-of-the-art methods.

In recent years, convolutional neural network (CNN) based methods have been widely used in remote sensing image compression tasks. However, CNN is commonly used to extract local information and does not fully utilize global contextual information. The transformer model can effectively extract the latent contextual information in remote sensing images due to its multihead self-attention mechanisms. Due to the multiscale local features and global low-frequency information of remote sensing images, existing deep-learning-based compression methods have not effectively combined CNN and transformer. In order to overcome the limitations of the above methods, a multihead global attention and spatial spectral information fusion network (MGSSNet) is proposed for remote sensing image compression. First, a spatial spectral information fusion attention module (SSIF-AM) is constructed to obtain multiscale local information. Second, a multihead global attention module (MHG-AM) is proposed to capture rich global context information. Third, a local global collaboration module is developed to explore the correlation between the multiscale local features obtained by SSIF-AM and the global visual features obtained by MHG-AM, and to efficiently model the intrinsic relationships between them to achieve effective feature fusion. Experimental results show that compared with advanced compression models, the proposed MGSSNet method achieves better compression performance. In addition, using reconstructed images obtained by different compression methods for classification tasks has proven that the proposed method can help achieve better classification performance, indicating that the proposed compression method can more fully preserve important information in the image.

There has been a surge in research focused on the analysis of vital signs using remote photoplethysmography (rPPG) sensors, as opposed to traditional photoplethysmography (PPG) methods. Unlike PPG, rPPG imposes no spatial constraints and employs a straightforward measurement technique, making it increasingly prevalent. Its integration into image processing, harnesses the remarkable advances in artificial-intelligence technology, achieving accuracy that is comparable to that of traditional PPG sensors. In prior studies, obtaining vital signs often necessitated an unnecessary and procedural fixation of facial positions within frames to enhance predictive accuracy. Despite such fixation, achieving notably high accuracy remained elusive. Here, we introduce a simple yet robust approach utilizing videos captured by an rPPG sensor, ensuring both high accuracy and resilience to noise. We propose a convolutional neural network model meticulously designed to resist interference from noise data that may arise in the initial stages, coupled with effective preprocessing techniques to attain superior predictive accuracy. Data extracted by a facial extractor undergoes preprocessing via normalization. Leveraging the Temporal Shift Module (TSM), this normalization efficiently captures temporal relationships without incurring additional computational overhead. Mitigating noise signal interference from non-facial data through the use of multiple attention masks and augmenting prediction accuracy via skip connections. Moreover, we compile a specially tailored dataset for pulse rate and breath rate data, catering specifically to the East Asian population. The proposed process demonstrates outstanding performance in predicting both pulse rate and breath rate.

Bone age assessment plays a vital role in monitoring the growth and development of adolescents. However, it is still challenging to obtain precise bone age from hand radiography due to these problems: 1) Hand bone varies greatly and is always masked by the background; 2) the hand bone radiographs with successive ages offer high similarity. To solve such issues, a region fine-grained attention network (RFGA-Net) was proposed for bone age assessment, where the region aware attention (RAA) module was developed to distinguish the skeletal regions from the background by modeling global spatial dependency; then the fine-grained feature attention (FFA) module was devised to identify similar bone radiographs by recognizing critical fine-grained feature regions. The experimental results demonstrate that the proposed RFGA-Net shows the best performance on the Radiological Society of North America (RSNA) pediatric bone dataset, achieving the mean absolute error (MAE) of 3.34 and the root mean square error (RMSE) of 4.02, respectively.

Given the recent increase in the prevalence of online food ordering apps, food delivery has become an emerging service. Intelligent dispatch systems have been widely deployed by many on-demand logistics companies to maximize delivery efficiency. Predicting the food delivery time is a key module that provides critical information at the decision-making stage of order dispatch to ensure the punctual delivery service for each customer. In this study, we propose a multitask attention network for food delivery time prediction, mimicking the driver's decision-making process during delivery. First, an attention mechanism is employed to capture mutual influences among orders and evaluate the importance of each order. Then, a multitask learning method is used to simultaneously train delivery time prediction and delivery priority prediction. Finally, a specific loss function is designed to further improve the accuracy of prediction. Extensive modeling demonstrates that our model greatly outperforms other state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Detecting violent behavior in videos to ensure public safety and security poses a significant challenge. Precisely identifying and categorizing instances of violence in real-life closed-circuit television, which vary across specifications and locations, requires comprehensive understanding and processing of the sequential information embedded in these videos. This study aims to introduce a model that adeptly grasps the spatiotemporal context of videos within diverse settings and specifications of violent scenarios. We propose a method to accurately capture spatiotemporal features linked to violent behaviors using optical flow and RGB data. The approach leverages a Conv3D-based ResNet-3D model as the foundational network, capable of handling high-dimensional video data. The efficiency and accuracy of violence detection are enhanced by integrating an attention mechanism, which assigns greater weight to the most crucial frames within the RGB and optical-flow sequences during instances of violence. Our model was evaluated on the UBI-Fight, Hockey, Crowd, and Movie-Fights datasets; the proposed method outperformed existing state-of-the-art techniques, achieving area under the curve scores of 95.4, 98.1, 94.5, and 100.0 on the respective datasets. Moreover, this research not only has the potential to be applied in real-time surveillance systems but also promises to contribute to a broader spectrum of research in video analysis and understanding. [ABSTRACT FROM AUTHOR]

Aim: Alzheimer's disease is a neurodegenerative disease that causes 60-70% of all cases of dementia. This study is to provide a novel method that can identify AD more accurately. Methods: We first propose a VGG-inspired network (VIN) as the backbone network and investigate the use of attention mechanisms. We proposed an Alzheimer's Disease VGG-Inspired Attention Network (ADVIAN), where we integrate convolutional block attention modules on a VIN backbone. Also, 18-way data augmentation is proposed to avoid overfitting. Ten runs of 10-fold cross-validation are carried out to report the unbiased performance. Results: The sensitivity and specificity reach 97.65 ± 1.36 and 97.86 ± 1.55, respectively. Its precision and accuracy are 97.87 ± 1.53 and 97.76 ± 1.13, respectively. The F1 score, MCC, and FMI are obtained as 97.75 ± 1.13, 95.53 ± 2.27, and 97.76 ± 1.13, respectively. The AUC is 0.9852. Conclusion: The proposed ADVIAN gives better results than 11 state-of-the-art methods. Besides, experimental results demonstrate the effectiveness of 18-way data augmentation. [ABSTRACT FROM AUTHOR]

Hyperspectral images (HSIs) are widely used to identify and characterize objects in scenes of interest, but they are associated with high acquisition costs and low spatial resolutions. With the development of deep learning, HSI reconstruction from low-cost and high-spatial-resolution RGB images has attracted widespread attention. It is an inexpensive way to obtain HSIs via the spectral reconstruction (SR) of RGB data. However, due to a lack of consideration of outdoor solar illumination variation in existing reconstruction methods, the accuracy of outdoor SR remains limited. In this paper, we present an attention neural network based on an adaptive weighted attention network (AWAN), which considers outdoor solar illumination variation by prior illumination information being introduced into the network through a basic 2D block. To verify our network, we conduct experiments on our Variational Illumination Hyperspectral (VIHS) dataset, which is composed of natural HSIs and corresponding RGB and illumination data. The raw HSIs are taken on a portable HS camera, and RGB images are resampled directly from the corresponding HSIs, which are not affected by illumination under CIE-1964 Standard Illuminant. Illumination data are acquired with an outdoor illumination measuring device (IMD). Compared to other methods and the reconstructed results not considering solar illumination variation, our reconstruction results have higher accuracy and perform well in similarity evaluations and classifications using supervised and unsupervised methods. [ABSTRACT FROM AUTHOR]

Polygraphs are used in criminal interrogations to detect deception. However, polygraphs can be difficult to administer under circumstances that prevent the use of biosensors. To address the shortcomings of the biosensors, deception-detection technology without biosensors is needed. We propose a deception-detection method, FacialCueNet, which is a multi-modal network that utilizes both facial images and facial cues based on deep-learning technology. FacialCueNet incorporates facial cues that indicate deception, such as action-unit frequency, symmetry, gaze pattern, and micro-expressions extracted from videos. Additionally, the spatial-temporal attention module, based on convolutional neural network and convolutional long short-term memory, is applied to FacialCueNet to provide interpretable information from interrogations. Because our goal was developing an algorithm applicable to criminal interrogations, we trained and evaluated FacialCueNet using the DDCIT dataset, which was collected using a data acquisition protocol similar to those used in actual investigations. To compare deception-detection performance with state-of-the-art works, a public dataset was also used. As a result, the mean deception-detection F1 score using the DDCIT dataset was 81.22%, with an accuracy of 70.79%, recall of 0.9476, and precision of 0.7107. When evaluating against the public database, our method demonstrated an evaluation accuracy of 88.45% and achieved an AUC of 0.9541, indicating a improvement of 1.25% compared to the previous results. We also present interpretive results of deception detection by analyzing the influence of spatial and temporal factors. These results show that FacialCueNet has the potential to detect deception using only facial videos. By providing interpretation of predictions, our system could be useful tool for criminal interrogation. [ABSTRACT FROM AUTHOR]

Accurate and rapid censuses can provide detailed basic information for a country, which is useful for resource allocation, disease control, disaster prevention, urban planning, and business management. However, traditional censuses often take up much time, manpower, and financial resources. Population maps are created by national statistical institutes at statistical units. Remote sensing imagery combined with end-to-end deep learning models makes it possible to estimate a wide range of populations at a low cost. This study demonstrates the effectiveness of a local–global dual attention network (LGANet) for population estimation using remote sensing images. The LGANet contains a local attention embranchment and a global attention embranchment on the top of the backbone to adaptively learn and integrate two discriminative features simultaneously. To enhance the precision of population estimation, the outputs from the two attention modules are combined. This method utilizes daytime remote sensing images as input, complemented by nighttime light data, to estimate the population on 1 km grids. Our method exhibits superior accuracy compared to other deep learning methods, as evidenced by an experimental comparison between the estimated population and the ground-truth population in 1 km grids.