Search

Recent studies have shown that higher accuracy on ImageNet usually leads to better robustness against different corruptions. Therefore, in this paper, instead of following the traditional research paradigm that investigates new out-of-distribution corruptions or perturbations deep models may encounter, we conduct model debugging in in-distribution data to explore which object attributes a model may be sensitive to. To achieve this goal, we create a toolkit for object editing with controls of backgrounds, sizes, positions, and directions, and create a rigorous benchmark named ImageNet-E(diting) for evaluating the image classifier robustness in terms of object attributes. With our ImageNet-E, we evaluate the performance of current deep learning models, including both convolutional neural networks and vision transformers. We find that most models are quite sensitive to attribute changes. A small change in the background can lead to an average of 9.23\% drop on top-1 accuracy. We also evaluate some robust models including both adversarially trained models and other robust trained models and find that some models show worse robustness against attribute changes than vanilla models. Based on these findings, we discover ways to enhance attribute robustness with preprocessing, architecture designs, and training strategies. We hope this work can provide some insights to the community and open up a new avenue for research in robust computer vision. The code and dataset are available at https://github.com/alibaba/easyrobust., Accepted by CVPR2023

It is necessary to assess the appropriate dietary protein level of the forest musk deer (FMD), as nutritional needs are unclear. The microbiome in gastrointestinal tracts plays an important role in regulating nutrient utilization, absorption and host growth or development. Thus, we aimed to evaluate growth performance, nutrient digestibility and fecal microbiome of growing FMD supplied with different protein levels of diets. Eighteen 6-month-old male FMD with an initial weight 5.0 ± 0.2 kg were used in a 62-day trial. The animals were randomly distributed to three groups, the dietary crude protein (CP) level was 11.51% (L), 13.37% (M), and 15.48% (H). The results showed that the CP digestibility decreased as dietary CP level increased (p p Ruminococcaceae_005, Ruminococcaceae_UCG-014 and uncultured_bacterium_f_Lachnospiraceae were significantly increased wtih rising CP, the proportions of Bacteroides and Rikenellaceae_RC9_gut_group were significantly decrease nevertheless at the genus level. The higher abundance of f_Prevotellaceae and g_Prevotellaceae_UCG_004 were found at M group by LEfSe analysis. The relative abundance of uncultured_bacterium_f_Ruminococcaceae was positively correlated with the average daily gain and feed conversion ratio (p Family_XIII_AD3011_group was negatively correlated with feed conversion ratio (p

With the recent worldwide COVID-19 pandemic, almost everyone wears a mask daily, leading to severe degradation in the accuracy of conventional face recognition systems. Several works improve the performance of masked faces by adopting synthetic masked face images for training. However, such methods often cause performance degradation on unmasked faces, raising the contradiction between the face recognition system's accuracy on unmasked and masked faces. In this paper, we propose a dual-proxy face recognition training method to improve masked faces' performance while maintaining unmasked faces' performance. Specifically, we design two fully-connected layers as the unmasked and masked feature space proxies to alleviate the significant difference between the two data distributions. The cross-space constraints are adopted to ensure the intra-class compactness and inter-class discrepancy. Extensive experiments on popular unmasked face benchmarks and masked face benchmarks, including real-world mask faces and the generated mask faces, demonstrate our method's superiority over the state-of-the-art methods on masked faces without incurring a notable accuracy degradation on unmasked faces.

This study extends our understanding of CEO inside debt compensation under an agency problem perspective by considering the impact of a behavioral trait, namely CEO overconfidence. Using a sample of US firms in Standard & Poor’s ExecuComp for the period 2006-2019, we find that overconfident CEOs exhibit greater inside debt incentives (i.e., incentives arising from definedbenefit pensions and deferred compensation). This relationship is more pronounced among firms with higher CEO overconfidence-induced agency costs of debt (e.g., financially unconstrained firms) managed by CEOs who are less able to align compensation with their own preferences (e.g., less powerful CEOs). The results are robust to endogeneity, self-selection concerns, and several alternative explanations. We contribute to the confluence of agency problem executive compensation under overconfident CEOs, and optimal contracting and managerial power theories.

Supplemental material, sj-pdf-1-imr-10.1177_03000605231174313 for Capillary leak syndrome presenting as refractory ascites in a patient with lymphoma: A rare case report by Xu Han, Huanjun Huang, Mingyu Zhang, Yue Wen, Shuhui Wang, Zheng Huang, Dean Tian and Yan Yu in Journal of International Medical Research

Research Data, sj-pdf-3-imr-10.1177_03000605231174313 for Capillary leak syndrome presenting as refractory ascites in a patient with lymphoma: A rare case report by Xu Han, Huanjun Huang, Mingyu Zhang, Yue Wen, Shuhui Wang, Zheng Huang, Dean Tian and Yan Yu in Journal of International Medical Research

Few-shot image generation aims to generate data of an unseen category based on only a few samples. Apart from basic content generation, a bunch of downstream applications hopefully benefit from this task, such as low-data detection and few-shot classification. To achieve this goal, the generated images should guarantee category retention for classification beyond the visual quality and diversity. In our preliminary work, we present an ``editing-based'' framework Attribute Group Editing (AGE) for reliable few-shot image generation, which largely improves the generation performance. Nevertheless, AGE's performance on downstream classification is not as satisfactory as expected. This paper investigates the class inconsistency problem and proposes Stable Attribute Group Editing (SAGE) for more stable class-relevant image generation. SAGE takes use of all given few-shot images and estimates a class center embedding based on the category-relevant attribute dictionary. Meanwhile, according to the projection weights on the category-relevant attribute dictionary, we can select category-irrelevant attributes from the similar seen categories. Consequently, SAGE injects the whole distribution of the novel class into StyleGAN's latent space, thus largely remains the category retention and stability of the generated images. Going one step further, we find that class inconsistency is a common problem in GAN-generated images for downstream classification. Even though the generated images look photo-realistic and requires no category-relevant editing, they are usually of limited help for downstream classification. We systematically discuss this issue from both the generative model and classification model perspectives, and propose to boost the downstream classification performance of SAGE by enhancing the pixel and frequency components.

Supplemental material, sj-pdf-2-imr-10.1177_03000605231174313 for Capillary leak syndrome presenting as refractory ascites in a patient with lymphoma: A rare case report by Xu Han, Huanjun Huang, Mingyu Zhang, Yue Wen, Shuhui Wang, Zheng Huang, Dean Tian and Yan Yu in Journal of International Medical Research

Supplemental material, sj-pdf-1-imr-10.1177_03000605231174313 for Capillary leak syndrome presenting as refractory ascites in a patient with lymphoma: A rare case report by Xu Han, Huanjun Huang, Mingyu Zhang, Yue Wen, Shuhui Wang, Zheng Huang, Dean Tian and Yan Yu in Journal of International Medical Research

Weakly supervised video anomaly detection aims to identify abnormal events in videos using only video-level labels. Recently, two-stage self-training methods have achieved significant improvements by self-generating pseudo labels and self-refining anomaly scores with these labels. As the pseudo labels play a crucial role, we propose an enhancement framework by exploiting completeness and uncertainty properties for effective self-training. Specifically, we first design a multi-head classification module (each head serves as a classifier) with a diversity loss to maximize the distribution differences of predicted pseudo labels across heads. This encourages the generated pseudo labels to cover as many abnormal events as possible. We then devise an iterative uncertainty pseudo label refinement strategy, which improves not only the initial pseudo labels but also the updated ones obtained by the desired classifier in the second stage. Extensive experimental results demonstrate the proposed method performs favorably against state-of-the-art approaches on the UCF-Crime, TAD, and XD-Violence benchmark datasets.

Few-shot image generation is a challenging task since it aims to generate diverse new images for an unseen category with only a few images. Existing methods suffer from the trade-off between the quality and diversity of generated images. To tackle this problem, we propose Hyperbolic Attribute Editing (HAE), a simple yet effective method. Unlike other methods that work in Euclidean space, HAE captures the hierarchy among images using data from seen categories in hyperbolic space. Given a well-trained HAE, images of unseen categories can be generated by moving the latent code of a given image toward any meaningful directions in the Poincar\'e disk with a fixing radius. Most importantly, the hyperbolic space allows us to control the semantic diversity of the generated images by setting different radii in the disk. Extensive experiments and visualizations demonstrate that HAE is capable of not only generating images with promising quality and diversity using limited data but achieving a highly controllable and interpretable editing process.

In neuroimaging analysis, fMRI can well assess the function changes for brain diseases with no obvious structural lesions. To date, most deep-learning-based fMRI studies have employed functional connectivity (FC) as the basic feature for disease classification. However, FC is calculated on time series of predefined regions of interest and neglects detailed information contained in each voxel. Another drawback of using FC is the limited sample size for the training of deep models. The low representation ability of FC leads to poor performance in clinical practice, especially when dealing with multimodal medical data involving multiple types of visual signals and textual records for brain diseases. To overcome this bottleneck problem in the fMRI feature modality, we propose BrainFormer, an end-to-end functional interaction learning method for brain disease classification with single fMRI volume. Unlike traditional deep learning methods that construct convolution and transformers on FC, BrainFormer learns the functional interaction from fMRI signals, by modeling the local cues within each voxel with 3D convolutions and capturing the global correlations among distant regions with specially designed global attention mechanisms from shallow layers to deep layers. Meanwhile, BrainFormer can deal with multimodal medical data including fMRI volume, structural MRI, FC features and phenotypic data to achieve more comprehensive brain disease diagnosis. We evaluate BrainFormer on five independent multi-site datasets on autism, Alzheimer's disease, depression, attention deficit hyperactivity disorder and headache disorders. The results demonstrate its effectiveness and generalizability for multiple brain diseases diagnosis with multimodal features. BrainFormer may promote precision of neuroimaging-based diagnosis in clinical practice and motivate future studies on fMRI analysis.

Few-shot image generation is a challenging task even using the state-of-the-art Generative Adversarial Networks (GANs). Due to the unstable GAN training process and the limited training data, the generated images are often of low quality and low diversity. In this work, we propose a new editing-based method, i.e., Attribute Group Editing (AGE), for few-shot image generation. The basic assumption is that any image is a collection of attributes and the editing direction for a specific attribute is shared across all categories. AGE examines the internal representation learned in GANs and identifies semantically meaningful directions. Specifically, the class embedding, i.e., the mean vector of the latent codes from a specific category, is used to represent the category-relevant attributes, and the category-irrelevant attributes are learned globally by Sparse Dictionary Learning on the difference between the sample embedding and the class embedding. Given a GAN well trained on seen categories, diverse images of unseen categories can be synthesized through editing category-irrelevant attributes while keeping category-relevant attributes unchanged. Without re-training the GAN, AGE is capable of not only producing more realistic and diverse images for downstream visual applications with limited data but achieving controllable image editing with interpretable category-irrelevant directions., CVPR2022

Additional file 1. Additional figures and tables.

Referring video object segmentation aims to segment the object referred by a given language expression. Existing works typically require compressed video bitstream to be decoded to RGB frames before being segmented, which increases computation and storage requirements and ultimately slows the inference down. This may hamper its application in real-world computing resource limited scenarios, such as autonomous cars and drones. To alleviate this problem, in this paper, we explore the referring object segmentation task on compressed videos, namely on the original video data flow. Besides the inherent difficulty of the video referring object segmentation task itself, obtaining discriminative representation from compressed video is also rather challenging. To address this problem, we propose a multi-attention network which consists of dual-path dual-attention module and a query-based cross-modal Transformer module. Specifically, the dual-path dual-attention module is designed to extract effective representation from compressed data in three modalities, i.e., I-frame, Motion Vector and Residual. The query-based cross-modal Transformer firstly models the correlation between linguistic and visual modalities, and then the fused multi-modality features are used to guide object queries to generate a content-aware dynamic kernel and to predict final segmentation masks. Different from previous works, we propose to learn just one kernel, which thus removes the complicated post mask-matching procedure of existing methods. Extensive promising experimental results on three challenging datasets show the effectiveness of our method compared against several state-of-the-art methods which are proposed for processing RGB data. Source code is available at: https://github.com/DexiangHong/MANet., Comment: Accepted by ACM MM 2022

Additional file 3: Table S1. Association of CARM1 level with clinicopathological parameters of patients with gastric cancer.

Additional file 5: Table S3. Antibodies and regents used in experiments.

Additional file 4: Table S2. Sequences of primers and targets for siRNA.

Dense video captioning (DVC) aims to generate multi-sentence descriptions to elucidate the multiple events in the video, which is challenging and demands visual consistency, discoursal coherence, and linguistic diversity. Existing methods mainly generate captions from individual video segments, lacking adaptation to the global visual context and progressive alignment between the fast-evolved visual content and textual descriptions, which results in redundant and spliced descriptions. In this paper, we introduce the concept of information flow to model the progressive information changing across video sequence and captions. By designing a Cross-modal Information Flow Alignment mechanism, the visual and textual information flows are captured and aligned, which endows the captioning process with richer context and dynamics on event/topic evolution. Based on the Cross-modal Information Flow Alignment module, we further put forward DVCFlow framework, which consists of a Global-local Visual Encoder to capture both global features and local features for each video segment, and a pre-trained Caption Generator to produce captions. Extensive experiments on the popular ActivityNet Captions and YouCookII datasets demonstrate that our method significantly outperforms competitive baselines, and generates more human-like text according to subject and objective tests.

Clinical treatment of triple negative breast cancer (TNBC) is very poor for lack of effective treatment combination selection. Protein C receptor (PROCR) is a novel cancer stem marker in TNBC patients tumor tissues. Developed based on peptide BP10 with affinity to PROCR as a targeting element, constructing a peptide drug conjugate of BP10 covalently coupling doxorubicin with disulfide bonds. This study demonstrated that the constructed BP10-DOX can selectively target Triplenegative breast cancer cells expressing PROCR and controlled release of DOX in response to the GSH environment. Moreover, BP10-DOX improves the therapeutic efficiency on MDA-MB-231 cells in vitro. Further evidence obtained from in vivo xenograft experiments revealed that administration of BP10-DOX enhanced the antitumor efficacy. This study developed a promising chemotherapy strategy for TNBC.

Detecting faults in axial piston pumps is of significance to enhance the reliability and security of hydraulic systems. However, it is difficult to detect multiple faults in the hydraulic electromechanical coupling systems because the fault mechanism of some faults is unclear. In this paper, a method using deep belief networks (DBNs) is proposed to detect multiple faults in axial piston pumps. Firstly, for each individual fault, all the data indicators extracted from the raw signals in time domain, frequency domain and time-frequency domain are calculated to construct training and testing samples. Then, the constructed samples are fed into DBNs to classify the multiple faults in axial piston pumps. With restricted Boltzmann machine (RBM) stacked layer by layer, DBNs can automatically learn fault features. Numerical simulations using the benchmark data of five faults in rolling bearings are classified by the present method to select the relative optimal combination of indicators. The classification results are also compared with those commonly used support vector machine (SVM) and artificial neural network (ANN) to manifest the classification accuracy of the present method. Experimental investigations are performed to classify four faults in an axial piston pump. The classification accuracy ratio is 97.40%, which confirms the feasibility and effectiveness of multiple faults detection in axial piston pumps using DBNs.

Gouty arthritis (GA) is an inflammatory disease owing to the accumulation of monosodium urate (MSU) in joints, leading to redness and burning pain. In this study, the effect of Zisheng Shenqi Decoction (ZSD) on a rat model of MSU-induced GA was investigated. ZSD obviously diminished the right paw thickness, the degree of the swelling of the paw, and the infiltration of the inflammatory cell, as well as cartilage erosion, and widened the joint space in MSU-treated rats. Besides, MSU remarkably elevated the release of tumor necrosis factor-α (TNF-α), interleukin-1β (IL-1β), IL-6, and IL-18; however, ZSD treatment dose dependently lowered these levels and resulted in a significant decrease in articular elastase activity. Also, ZSD administration increased the activities of superoxide dismutase (SOD), glutathione peroxidase (GSH-Px), and catalase (CAT) but declined malondialdehyde (MDA) and nitrogen monoxide (NO) contents. Importantly, western blotting analysis revealed that NOD-like receptor protein 3 (NLRP3), cleaved caspase-1, IL-1β, nuclear factor-E2-related factor 2 (Nrf2) in the cytoplasm, phosphorylated mammalian target of rapamyclin (p-mTOR), and p62 expressions were downregulated, whereas the levels of nuclear Nrf2, phosphorylated AMP-activated protein kinase (p-AMPK), Beclin-1, and LC3II/I were upregulated by ZSD. Immunofluorescence assay indicated that ZSD evidently promoted nuclear translocation of LC3. Taken together, ZSD inhibited inflammation and oxidative stress and facilitated autophagy through the activation of the AMPK pathway and suppression of the mTOR signaling pathway, demonstrating its potential for preventing and curing GA.

Additional file 1: Figure S1. Screening of reporter genes. Figure S2. Characterization of different green fluorescent proteins expressed under the promoter PUAS1B4-LEUm. Figure S3. Schematic of theTEF core promoter. Figure S4. Time course of fluorescence by Po1g PUAS1B4-LEUm + hrGFPO. Table S1. Primers used in PCR. Table S2. Plasmids used in this study. Table S3. Strains used in this study.

Due to the domain discrepancy in visual domain adaptation, the performance of source model degrades when bumping into the high data density near decision boundary in target domain. A common solution is to minimize the Shannon Entropy to push the decision boundary away from the high density area. However, entropy minimization also leads to severe reduction of prediction diversity, and unfortunately brings harm to the domain adaptation. In this paper, we investigate the prediction discriminability and diversity by studying the structure of the classification output matrix of a randomly selected data batch. We find by theoretical analysis that the prediction discriminability and diversity could be separately measured by the Frobenius-norm and rank of the batch output matrix. The nuclear-norm is an upperbound of the former, and a convex approximation of the latter. Accordingly, we propose Batch Nuclear-norm Maximization and Minimization, which performs nuclear-norm maximization on the target output matrix to enhance the target prediction ability, and nuclear-norm minimization on the source batch output matrix to increase applicability of the source domain knowledge. We further approximate the nuclear-norm by L_{1,2}-norm, and design multi-batch optimization for stable solution on large number of categories. The fast approximation method achieves O(n^2) computational complexity and better convergence property. Experiments show that our method could boost the adaptation accuracy and robustness under three typical domain adaptation scenarios. The code is available at https://github.com/cuishuhao/BNM., Comment: TPAMI under revivew. arXiv admin note: text overlap with arXiv:2003.12237

Additional file 2: Supplementary Figure S2. Principal component analysis (PCA) of the Chinese indigenous rabbit breeds based on autosome (PC2=4.10%, PC3=3.41%).

Additional file 1: Supplementary Figure S1. Principal component analysis (PCA) of the Chinese indigenous rabbit breeds based on autosome (PC1=5.59%, PC3=3.41%).

In visual domain adaptation (DA), separating the domain-specific characteristics from the domain-invariant representations is an ill-posed problem. Existing methods apply different kinds of priors or directly minimize the domain discrepancy to address this problem, which lack flexibility in handling real-world situations. Another research pipeline expresses the domain-specific information as a gradual transferring process, which tends to be suboptimal in accurately removing the domain-specific properties. In this paper, we address the modeling of domain-invariant and domain-specific information from the heuristic search perspective. We identify the characteristics in the existing representations that lead to larger domain discrepancy as the heuristic representations. With the guidance of heuristic representations, we formulate a principled framework of Heuristic Domain Adaptation (HDA) with well-founded theoretical guarantees. To perform HDA, the cosine similarity scores and independence measurements between domain-invariant and domain-specific representations are cast into the constraints at the initial and final states during the learning procedure. Similar to the final condition of heuristic search, we further derive a constraint enforcing the final range of heuristic network output to be small. Accordingly, we propose Heuristic Domain Adaptation Network (HDAN), which explicitly learns the domain-invariant and domain-specific representations with the above mentioned constraints. Extensive experiments show that HDAN has exceeded state-of-the-art on unsupervised DA, multi-source DA and semi-supervised DA. The code is available at https://github.com/cuishuhao/HDA.

Semantic editing on segmentation map has been proposed as an intermediate interface for image generation, because it provides flexible and strong assistance in various image generation tasks. This paper aims to improve quality of edited segmentation map conditioned on semantic inputs. Even though recent studies apply global and local adversarial losses extensively to generate images for higher image quality, we find that they suffer from the misalignment of the boundary area in the mask area. To address this, we propose MExGAN for semantic editing on segmentation map, which uses a novel Multi-Expansion (MEx) loss implemented by adversarial losses on MEx areas. Each MEx area has the mask area of the generation as the majority and the boundary of original context as the minority. To boost convenience and stability of MEx loss, we further propose an Approximated MEx (A-MEx) loss. Besides, in contrast to previous model that builds training data for semantic editing on segmentation map with part of the whole image, which leads to model performance degradation, MExGAN applies the whole image to build the training data. Extensive experiments on semantic editing on segmentation map and natural image inpainting show competitive results on four datasets.

Sleep disorder significantly affects the life quality of a large number of people but is still an underrecognized disease. Dietary nutrition is believed to play a significant impact on sleeping wellness. Many nutritional supplements have been used trying to benefit sleep wellness. However, the relationship between nutritional components and sleep is complicated. Nutritional factors vary dramatically with different diet patterns and depend significantly on the digestive and metabiotic functions of each individual. Moreover, nutrition can profoundly affect the hormones and inflammation status which directly or indirectly contribute to insomnia. In this review, we summarized the role of major nutritional factors, carbohydrates, lipids, amino acids, and vitamins on sleep and sleep disorders and discussed the potential mechanisms.

The brain is the most important and complex organ in most living creatures which serves as the center of the nervous system. The function of human brain includes controlling of the motion of the body and different organs and maintaining basic homeostasis. The disorders of the brain caused by a variety of reasons often severely impact the patients’ normal life or lead to death in extreme cases. Monocyte is an important immune cell which is often recruited to the brain in a number of brain disorders. However, the role of monocytes may not be simply described as beneficial or detrimental. It significantly depends on the disease models and the stages of disease progression. In this review, we summarized the current knowledge about the role of monocytes and monocyte-derived macrophages during several common brain disorders. Major focuses include ischemic stroke, Alzheimer’s disease, multiple sclerosis, intracerebral hemorrhage, and insomnia. The recruitment, differentiation, and function of monocyte in these diseases are reviewed.

Most of existing salient object detection models have achieved great progress by aggregating multi-level features extracted from convolutional neural networks. However, because of the different receptive fields of different convolutional layers, there exists big differences between features generated by these layers. Common feature fusion strategies (addition or concatenation) ignore these differences and may cause suboptimal solutions. In this paper, we propose the F3Net to solve above problem, which mainly consists of cross feature module (CFM) and cascaded feedback decoder (CFD) trained by minimizing a new pixel position aware loss (PPA). Specifically, CFM aims to selectively aggregate multi-level features. Different from addition and concatenation, CFM adaptively selects complementary components from input features before fusion, which can effectively avoid introducing too much redundant information that may destroy the original features. Besides, CFD adopts a multi-stage feedback mechanism, where features closed to supervision will be introduced to the output of previous layers to supplement them and eliminate the differences between features. These refined features will go through multiple similar iterations before generating the final saliency maps. Furthermore, different from binary cross entropy, the proposed PPA loss doesn't treat pixels equally, which can synthesize the local structure information of a pixel to guide the network to focus more on local details. Hard pixels from boundaries or error-prone parts will be given more attention to emphasize their importance. F3Net is able to segment salient object regions accurately and provide clear local details. Comprehensive experiments on five benchmark datasets demonstrate that F3Net outperforms state-of-the-art approaches on six evaluation metrics., Accepted by AAAI2020, https://github.com/weijun88/F3Net

Vibration signals of faulty rolling element bearings usually exhibit non-linear and non-stationary characteristics caused by the complex working environment. It is difficult to develop a robust method to detect faults in bearings based on signal processing techniques. In this paper, convolutional neural network -based hidden Markov models (CNN HMMs) are presented to classify multi-faults in mechanical systems. In CNN HMMs, a CNN model is first employed to learn data features automatically from raw vibration signals. By utilizing the t-distributed stochastic neighbor embedding (t-SNE) technique, feature visualization is constructed to manifest the powerful learning ability of CNN. Then, HMMs are employed as a strong stability tool to classify faults. Both the benchmark data and experimental data are applied to the CNN HMMs. Classification results confirm the superior performance of the present combination model by comparing with CNN model alone, support vector machine (SVM) and back propagation (BP) neural network. It is shown that the average classification accuracy ratios are 98.125% and 98% for two data series with agreeable error rate reductions.

The fault diagnosis of axial piston pumps is of significance for enhancing the reliability and security of hydraulic systems. Most of the faults occurring in the mechanical components of piston pumps are exhibited as fault-excited impulses. However, the strong impact-induced natural periodic impulses under the common working conditions (i.e. reciprocating motion of pistons) inevitably cause interference that considerably affects the fault detection performance. In this study, a simulation-determined band pass filter is employed to improve the performance of minimum entropy deconvolution (MED) for the fault diagnosis of axial piston pump bearings. First, a finite element method (FEM) simulation is performed to determine the possible carrier frequency. Second, the carrier frequency is used as the center frequency in association with a fixed bandwidth to determine the band pass filter parameters. Finally, the MED technique is applied to enhance weak fault-excited impulses by means of kurtosis maximization. Thereafter, envelope spectrum analysis is applied to the enhanced signals to obtain faulty feature frequencies. Two case studies are conducted, using bearings with faults in the outer and inner races of an axial piston pumps under common working conditions. The case studies confirm the necessity and effectiveness of the proposed method for detecting bearings faults in axial piston pumps.

MicroRNAs (miRNAs) are small endogenous conserved RNAs regulating genes expression through base pairing with the 3'-untranslated region (3'-UTR) of target messenger RNAs. MiR-214-5p is a newly identified miRNA with its biological role largely unknown. In this study, we explored miR-214-5p expression status in 78 paired tumor and nontumor tissues obtained from patients with hepatocellular carcinoma (HCC) by RT-qPCR. The effects of miR-214-5p expression on HCC cell proliferation, cell cycle progression, and cell migration were measured by CCK-8 assay, flow cytometry, and wound-healing assay. A dual-luciferase activity assay was performed to identify whether KLF5 was a target of miR-214-5p. Kaplan-Meier curve and log-rank test were used to investigate the effects of miR-214-5p and KLF5 on overall survival and disease-free survival of patients with HCC. We found miR-214-5p expression was sharply reduced in HCC tissues and cell lines compared with the normal tissues and cell lines. Functional assay revealed that miR-214-5p overexpression could downregulate cell proliferation, cell migration, and arrested cell cycle at G0/G1 phase. Further, we validated Kruppel-like factor 5 (KLF5) as a direct target of miR-214-5p, and was upregulated in HCC and inversely correlated with the expression of miR-214-5p. Moreover, we found the low expression of miR-214-5p and high expression of KLF5 were correlated with tumor size, tumor stage, and poorer 5-year overall survival and disease-free survival of patients with HCC. In conclusion, our results suggested miR-214-5p functions as a tumor suppressor through targeting KLF5 in HCC. Also, miR-214-5p and KLF5 were identified as potential prognostic markers and might be therapeutic targets in HCC.

In video captioning task, the best practice has been achieved by attention-based models which associate salient visual components with sentences in the video. However, existing study follows a common procedure which includes a frame-level appearance modeling and motion modeling on equal interval frame sampling, which may bring about redundant visual information, sensitivity to content noise and unnecessary computation cost. We propose a plug-and-play PickNet to perform informative frame picking in video captioning. Based on a standard Encoder-Decoder framework, we develop a reinforcement-learning-based procedure to train the network sequentially, where the reward of each frame picking action is designed by maximizing visual diversity and minimizing textual discrepancy. If the candidate is rewarded, it will be selected and the corresponding latent representation of Encoder-Decoder will be updated for future trials. This procedure goes on until the end of the video sequence. Consequently, a compact frame subset can be selected to represent the visual information and perform video captioning without performance degradation. Experiment results shows that our model can use 6-8 frames to achieve competitive performance across popular benchmarks., Comment: 10 pages, 7 figures, 2 tables

HO_2 and OH, also known as HO_x, play an important role in controlling middle atmospheric O_3. Due to their photochemical production and short chemical lifetimes, HO_x are expected to respond rapidly to solar irradiance changes, resulting in O_3 variability. While OH solar cycle signals have been investigated, HO_2 studies have been limited by the lack of reliable observations. Here we present the first evidence of HO_2 variability during solar 27 day cycles by investigating the recently developed HO_2 data from the Aura Microwave Limb Sounder (MLS). We focus on 2012–2015, when solar variability is strong near the peak of Solar Cycle 24. The features of HO_2 variability, with the strongest signals at 0.01–0.068 hPa, correlate well with those of solar Lyman α. When continuous MLS OH observations are not available, the new HO_2 data could be a promising alternative for investigating HO_x variability and the corresponding impacts on O_3 and the climate.

Observations have shown that mesospheric hydroxyl (OH) is affected by energetic electron precipitation (EEP) at magnetic latitudes connected to the outer radiation belt. It is not clear, however, if the current satellite-based electron flux observations can be used to accurately describe EEP in atmospheric models. We use the Sodankylä Ion and Neutral Chemistry (SIC) model to reproduce the changes in OH and ozone observed by the Microwave Limb Sounder (MLS/Aura) during four strong EEP events. The daily mean electron energy-flux spectrum, needed for ionization rate calculations, is determined by combining the Medium Energy Proton and Electron Detector fluxes and spectral form from the instrument for the detection of particles high-energy electron detector on board the DEMETER satellite. We show that in general SIC is able to reproduce the observed day-to-day variability of OH and ozone. In the lower mesosphere, the model tends to underestimate the OH concentration, possibly because of uncertainties in the electron spectra for energies >300 keV. The model predicts OH increases at 60–80 km, reaching several hundred percent at 70–80 km during peak EEP forcing. Increases in OH are followed by ozone depletion, up to several tens of percent. The magnitude of modeled changes is similar to those observed by MLS and comparable to effects of individual solar proton events. Our results suggest that the combined satellite observations of electrons can be used to model the EEP effects above 70 km during geomagnetic storms, without a need for significant adjustments. However, for EEP energies >300 keV impacting altitudes

Since the 1970s, international trade and investment patterns have occurred great changes, also the production organization of international enterprises had undergone profound changes. Intra-industry trade gradually replace the inter-industry trade, has become the main form of international trade. After 1980s, the multinational enterprises got unprecedented development. All the data are from the U.S. Bureau of Economic Analysis and the World Bank.

Energetic particle precipitation leads to enhancement of odd hydrogen (HOx) below 80 km altitude through water cluster ion chemistry. Using measurements from the Microwave Limb Sounder (MLS/Aura) and Medium Energy Proton and Electron Detector (MEPED/POES) between 2004–2009, we study variations of nighttime OH caused by radiation belt electrons at geomagnetic latitudes 55–65°. For those months with daily mean 100–300 keV electron count rate exceeding 150 counts/s in the outer radiation belt, we find a strong correlation (r ≥ 0.6) between OH mixing ratios at 70–78 km (0.046–0.015 hPa) and precipitating electrons. Correlations r ≥ 0.35, corresponding to random chance probability p ≤ 5%, are observed down to52 km (0.681 hPa), while no clear correlation is observed at altitudes below. This suggests that the fluxes of ≥3 MeV electrons were not high enough to cause observable changes in OH mixing ratios. At 75 km, in about 34% of the 65 months analyzed we find a correlation r ≥ 0.35. Although similar results are obtained for both hemispheres in general, in some cases the differences in atmospheric conditions make the OH response more difficult to detect in the South. Considering the latitude extent of electron forcing, we find clear effects on OH at magnetic latitudes 55–72°, while the lower latitudes are influenced much less. Because the time period 2004–2009 analyzed here coincided with an extended solar minimum, and the year 2009 was anomalously quiet, it is reasonable to assume that our results provide a lower-limit estimation of the importance of energetic electron precipitation at the latitudes considered.

Since Sokolov[1] proposed the version of the Ultrasonic Microscope with an ultrasonic converter tube 40 years ago, many studies have been done. Smyth[2] produced the camera with the low-voltage electronic beam scanning. Jacobs[3] improved the effective aperture and increased the field of view of the tube using the discrete matching elements. Addison[4] invented the metal fiber faceplate which acts as a part of envelope and makes the construction of the tube’s inner target become outer target so that the chosing of the piezoelectric plate is easier. Nigam[5] adopted the foil-electret to make the condenser-microphone array. His tube has a large aperture up to 60. Brown[6] got the largest area of view of the tube up to 210 mm × 150 mm. Wardley[7] obtained the mechanically strong, long life Sokolov tube. The operating time can be up to 5000 hours. There are still many problems, such as resolution, sensitivity, effective aperture, large field of view, service life and so on, to be improved.

In order to develop a new vaccine candidate for equine infectious anemia virus (EIAV), gag gene of Chinese donkey leukocyte attenuated strain (EIAV DLV) and its parental virulent strain (EIAV LN) were inserted respectively into the TK region of the Tiantan strain (VV) of vaccinia virus by homologous recombination and the positive clone was confirmed by blue plaque assay. Protein expression was examined by Western blot. Prime and prime-boost procedures were used to immunize mice with two DNA vectors and two recombinant vaccinia viruses expressing EIAV Gag proteins. The results showed that the specific lysis of CTL responses in the DNA+rVV groups was stronger than those in the DNA groups, amounting to 31%. Although the levels of specific antibodies were not significantly different, we could conclude that the recombinant vaccinia virus could boost the cellular responses following DNA vector priming. There was no detectable difference between the immune responses induced by DLV and LN Gag proteins. This data demonstrates that the combined immunity of DNA vector and recombinant vaccinia virus expressing EIAV gag proteins, utilizing the prime-boost procedure, can drive immunized mice to produce powerful cellular responses. These results lay an important foundation for the development of a new EIAV genetic engineering vaccine.

A survey was conducted by the Sichuan Provincial Women's Federation from the winter of last year to the spring of this year. It was designed to probe into rural women's life in the province and covered ten townships in six counties with different levels of economic growth. Findings of the investigation indicate a general trend of decline in regard to the educational structure of female teenagers and young adults.