Search

Radar jamming recognition is a key step in electronic countermeasures, and accurate and sufficient labeled samples are essential for supervised learning-based recognition methods. However, in real practice, collected radar jamming samples often have weak labels (i.e., noisy-labeled or unlabeled ones), which degrade recognition performance. Additionally, recognition performance is hindered by limitations in capturing the global features of radar jamming. The Transformer (TR) has advantages in modeling long-range relationships. Therefore, a weakly supervised Transformer is proposed to address the issues of performance degradation under weak supervision. Specifically, complementary label (CL) TR, called RadarCL-TR, is proposed to improve radar jamming recognition accuracy with noisy samples. CL learning and a cleansing module are successively utilized to detect and remove potentially noisy samples. Thus, the adverse influence of noisy samples is mitigated. Additionally, semi-supervised learning (SSL) TR, called RadarSSL-PL-TR, is proposed to boost recognition performance under unlabeled samples via pseudo labels (PLs). Network generalization is improved by training with pseudo-labeling unlabeled samples. Moreover, the RadarSSL-PL-S-TR is proposed to further promote recognition performance, where a selection module identifies reliable pseudo-labeling samples. The experimental results show that the proposed RadarCL-TR and RadarSSL-PL-S-TR outperform comparison methods in recognition accuracy by at least 7.07 % and 6.17 % with noisy and unlabeled samples, respectively. [ABSTRACT FROM AUTHOR]

Recently, transformer has gradually attracted interest for its excellence in modeling the long-range dependencies of spatial-spectral features in HSI. However, transformer has the problem of the quadratic computational complexity due to the self-attention mechanism, which is heavier than other models and thus has limited adoption in HSI processing. Fortunately, the recently emerging state space model-based Mamba shows great computational efficiency while achieving the modeling power of transformers. Therefore, in this paper, we first proposed spectral-spatial Mamba (SS-Mamba) for HSI classification. Specifically, SS-Mamba mainly includes a spectral-spatial token generation module and several stacked spectral-spatial Mamba blocks. Firstly, the token generation module converts any given HSI cube to spatial and spectral tokens as sequences. And then these tokens are sent to stacked spectral-spatial mamba blocks (SS-MB). Each SS-MB includes two basic mamba blocks and a spectral-spatial feature enhancement module. The spatial and spectral tokens are processed separately by the two basic mamba blocks, correspondingly. Moreover, the feature enhancement module modulates spatial and spectral tokens using HSI sample's center region information. Therefore, the spectral and spatial tokens cooperate with each other and achieve information fusion within each block. The experimental results conducted on widely used HSI datasets reveal that the proposed SS-Mamba requires less processing time compared with transformer. The Mamba-based method thus opens a new window for HSI classification. [ABSTRACT FROM AUTHOR]

A radiation-hardened delay-locked loop (DLL) applied for multiple frequency ranges is proposed in this paper. Novel and effective radiation-hardened techniques are applied in charge pump (CP) and voltage-controlled delay line (VCDL) to avoid inverted lock error and missing pulses error caused by single-event transient (SET). In order to ensure a wide operation frequency range, a frequency detection module is designed to detect the frequency of input signals and adjust the current source array of VCDL so as to change its delay time. Besides, an accelerate module is designed to reduce the lock time of DLL by providing an extra discharge current. According to the simulation results, the proposed DLL shows good performance when ion strikes occur and can work properly in a wide operation frequency range from 3.7 GHz to 7.4 GHz. Also, the proposed accelerate module helps to reduce lock time by 22.2%. [ABSTRACT FROM AUTHOR]

Background: Current artificial intelligence (AI) in histopathology typically specializes on a single task, resulting in a heavy workload of collecting and labeling a sufficient number of images for each type of cancer. Heterogeneous transfer learning (HTL) is expected to alleviate the data bottlenecks and establish models with performance comparable to supervised learning (SL). Methods: An accurate source domain model was trained using 28,634 colorectal patches. Additionally, 1000 sentinel lymph node patches and 1008 breast patches were used to train two target domain models. The feature distribution difference between sentinel lymph node metastasis or breast cancer and CRC was reduced by heterogeneous domain adaptation, and the maximum mean difference between subdomains was used for knowledge transfer to achieve accurate classification across multiple cancers. Result: HTL on 1000 sentinel lymph node patches (L-HTL-1000) outperforms SL on 1000 sentinel lymph node patches (L-SL-1-1000) (average area under the curve (AUC) and standard deviation of L-HTL-1000 vs. L-SL-1-1000: 0.949 ± 0.004 vs. 0.931 ± 0.008, p value = 0.008). There is no significant difference between L-HTL-1000 and SL on 7104 patches (L-SL-2-7104) (0.949 ± 0.004 vs. 0.948 ± 0.008, p value = 0.742). Similar results are observed for breast cancer. B-HTL-1008 vs. B-SL-1-1008: 0.962 ± 0.017 vs. 0.943 ± 0.018, p value = 0.008; B-HTL-1008 vs. B-SL-2-5232: 0.962 ± 0.017 vs. 0.951 ± 0.023, p value = 0.148. Conclusions: HTL is capable of building accurate AI models for similar cancers using a small amount of data based on a large dataset for a certain type of cancer. HTL holds great promise for accelerating the development of AI in histopathology. [ABSTRACT FROM AUTHOR]

This paper presents a delay-locked loop (DLL) with multiple radiation-hardened techniques. A radiation-hardened charge pump (RH-CP) and a radiation-hardened voltage-controlled delay line (RH-VCDL) are proposed to mitigate effects on DLLs caused by single-event transient (SET). A lock detection module (LDM) is designed in the RH-CP to separate CP and VCDL when the DLL is locked. Thanks to LDM, the voltage transients caused by ion strikes on CP can be prevented from effecting the control voltage of VCDL and causing inverted lock error. A SET correction module (SCM) combined with sliced delay cells is used in the RH-VCDL to select the uncorrupted output and avoid missing pulses errors occurring at the output of DLL. The proposed DLL is designed in 22-nm CMOS process with an operation frequency of 5 GHz. Simulations at linear energy transfer (LET) between 40 and 100 MeV-cm2/mg show that RH-CP eliminates inverted lock error and RH-VCDL mitigates missing pulses generated by the DLL after ion strikes effectively. [ABSTRACT FROM AUTHOR]

In this paper, a programmable low-power analog front-end (AFE) integrated circuit (IC) for biomedical signal monitoring systems is presented. The whole system includes four parts, instrumentation amplifiers (IA), programmable gain amplifiers (PGA), programmable bandwidth filters (PBF) and successive approximation register analog-to-digital convertors (SAR ADC). The proposed IA employs a pseudo-differential structure to enhance the input impedance of the system. A novel hybrid bandwidth extension technology is introduced in the proposed PGA to correct the low-frequency distortion and improve the low-frequency bandwidth. In the design of PBF, the current steering technology is applied to achieve an area-efficient and energy-efficient architecture. Furthermore, a novel efficient switching scheme and a self-calibration dynamic element matching method are presented in the design of SAR ADC to reduce the power consumption and eliminate the effect of capacitor mismatch. In 0.18 μm CMOS technology, this AFE IC consumes 46.8 μW and occupies 0.36 mm2. It achieves a variable gain from 40.4 to 55.1 dB under different modes designed for different biomedical signals. The spurious-free dynamic range and signal-to-noise plus distortion ratio of the proposed SAR ADC are 71.3 and 53.3 dB, respectively, along with a figure of merit of 0.18 pJ/conv. [ABSTRACT FROM AUTHOR]

Convolutional Neural Network (CNN) has been proved excellent for hyperspectral image(HSI) classification on the basis of large training set. However, the lack of available samples degrades the performance of CNNs in practical HSI classification. Although transfer learning is a typical strategy to address the issue of limited training samples, such methods tend to reload source model's feature extractors directly ignoring their generalization. Furthermore, most existing adaptive transfer learning algorithms pay attention to spectral features of HSI while the obvious difference between RGB image and HSI's spatial features is overlooked. In this paper, the proposed method can solve the mentioned problems to some extent. Firstly, the proposed adaptive heterogeneous transfer (AHT) can measure their transferability and find appropriate solution to transfer knowledge. It computes feature extractors' transferability based on feature maps and influences the learning process of target model through backpropagation, therefore, target model is prone to receive knowledge from feature extractors with high generalization. Secondly, the proposed method lays emphasis on HSI's spatial features. It adopts the methodology of spatial attention and calculates the regularization loss based on spatial features. Experiments were conducted on two widely used HSI datasets. The results demonstrate the method's effectiveness when few samples are available. [ABSTRACT FROM AUTHOR]

Recently, deep learning methods, especially convolutional neural networks (CNNs), have achieved good performance for hyperspectral image (HSI) classification. However, due to limited training samples of HSIs and the high volume of trainable parameters in deep models, training deep CNN-based models is still a challenge. To address this issue, this study investigates contrastive learning (CL) as a pre-training strategy for HSI classification. Specifically, a supervised contrastive learning (SCL) framework, which pre-trains a feature encoder using an arbitrary number of positive and negative samples in a pair-wise optimization perspective, is proposed. Additionally, three techniques for better generalization in the case of limited training samples are explored in the proposed SCL framework. First, a spatial–spectral HSI data augmentation method, which is composed of multiscale and 3D random occlusion, is designed to generate diverse views for each HSI sample. Second, the features of the augmented views are stored in a queue during training, which enriches the positives and negatives in a mini-batch and thus leads to better convergence. Third, a multi-level similarity regularization method (MSR) combined with SCL (SCL–MSR) is proposed to regularize the similarities of the data pairs. After pre-training, a fully connected layer is combined with the pre-trained encoder to form a new network, which is then fine-tuned for final classification. The proposed methods (SCL and SCL–MSR) are evaluated on four widely used hyperspectral datasets: Indian Pines, Pavia University, Houston, and Chikusei. The experiment results show that the proposed SCL-based methods provide competitive classification accuracy compared to the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Accurate classification of remote sensing (RS) images is a perennial topic of interest in the RS community. Recently, transfer learning, especially for fine-tuning pretrained convolutional neural networks (CNNs), has been proposed as a feasible strategy for RS scene classification. However, because the target domain (i.e., the RS images) and the source domain (e.g., ImageNet) are quite different, simply using the model pretrained on an ImageNet dataset presents some difficulties. The RS images and the pretrained models need to be properly adjusted to build a better classification system. In this study, an adaptive learning strategy for transferring a CNN-based model is proposed. First, an adaptive transform is used to adjust the original size of the RS image to a certain size, which is tailored to the input of the subsequent pretrained model. Then, an adaptive transferring model is proposed to automatically learn what knowledge from the pretrained model should be transferred to the RS scene classification model. Finally, in combination with a label smoothing approach, an adaptive label is presented to generate soft labels based on the statistics of the classification model predictions for each category, which is beneficial for learning the relationships between the target and nontarget categories of scenes. In general, the proposed methods adaptively manage the input, model, and label simultaneously, which leads to better classification performance for RS scene classification. The proposed methods are tested on three widely used datasets, and the obtained results show that the proposed methods provide competitive classification accuracy compared to the state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Recently, deep convolutional neural networks (CNNs) have achieved a high classification accuracy of hyperspectral images (HSIs). However, high accuracy is not the only goal of a good HSI classifier. In real-world applications, it is necessary to tell whether the classifier is certain about its classification result, which is critical for safe usage. Unfortunately, most of the existing models do not consider the issue. In this study, uncertainty is estimated and reduced to build a trustworthy HSI classifier. First, since the output probabilities of the softmax layer cannot represent the confidence scores, the distance measurement scheme is used to measure the confidence scores. Then, a trustworthy HSI classifier, which reduces the predictive uncertainty in CNN (i.e., PU-CNN), is obtained by minimizing the distance to the correct centroid. Second, the fact that a training sample of HSI usually contains many pixel vectors that belong to different classes brings label uncertainty. Then, label uncertainty CNN (i.e., LU-CNN), which uses a classifier-consistent estimator to recover the multiple classes in each HSI sample, is proposed. LU-CNN computes loss over candidate label sets to find the optimal classes, which leads to a trustworthy HSI classifier. Finally, the combination of PU-CNN and LU-CNN [i.e., predictive uncertainty and label uncertainty (PL-CNN)] is proposed to address predictive uncertainty and label uncertainty at the same time. Experimental results on the three popular hyperspectral datasets show that the proposed methods yield improvements in both accuracy and confidence. The proposed trustworthy classifier opens a new window for the safe usage of HSI. [ABSTRACT FROM AUTHOR]

This paper presents a low-power asynchronous 8-bit 500MS/s 2b/cycle successive-approximation-register (SAR) analog-to-digital converter (ADC) in 40 nm CMOS process. The proposed ADC shows high energy efficiency and good performance against process–voltage–temperature (PVT) variations by a background offset mismatch calibration technique, a supply voltage calibration technique and an efficient 2b/cycle switching scheme. The background offset mismatch calibration eliminates the effect of offset mismatches of the comparator array without any extra phase. The supply voltage calibration is presented to monitor the conversion speed of the ADC and adjust the speed of comparators to ensure the ADC can work properly at different process corners. The improved switching scheme helps cut down the switch times and simplifies the switch logic, which saves 87.8% switching energy compared with conventional VCM-based scheme. The proposed SAR ADC simulated a signal-to-noise plus distortion ratio (SNDR) of 45.8 dB and a spurious-free dynamic range (SFDR) of 55.5 dB for a near-Nyquist input while consuming a total power of 2.59 mW, culminating in a Walden figure of merit (FoM) of 32 fJ/conversion-step. [ABSTRACT FROM AUTHOR]

Summary: Motivation: Most low‐ and middle‐income countries face an urgent need to scale up sustainable finance for low‐carbon and climate‐resilient infrastructure investment, yet underdeveloped capital markets tend to inhibit domestic resource mobilization for infrastructure investment. At the same time, domestic savers face a scarcity of "safe" local‐currency assets, resulting in the export of capital. Purpose: This article explores options for mobilizing domestic savings through financial technology (fintech) solutions to scale up sustainable investment. Methods and approach: The article discusses how fintech can help to complement conventional capital markets and mobilize financial resources for sustainable infrastructure investments. Findings: The article puts forward a proposal for blockchain‐based project bonds to raise finance through a digital crowdfunding platform which is able to record transparently and certify the use of proceeds, sustainability impact, and revenue streams of projects by combining timestamp, public and private key mechanisms, and smart contract technologies. Policy implications: The proposed approach would not only provide investors of different sizes with the opportunity to purchase local‐currency assets and issuers such as municipalities to raise funds for sustainable infrastructure investment, it would also facilitate project management once the project is operational, for example through metering and billing, and create full transparency over the life cycle of the investment, reducing problems concerning the misuse of funds. [ABSTRACT FROM AUTHOR]

Necroptosis is a form of regulated cell death that depends on the receptor-interacting serine-threonine kinase 3 (RIPK3) and mixed lineage kinase domain-like (MLKL). The molecular mechanisms underlying distinct instances of necroptosis have only recently begun to emerge. In the present study, we characterized RABGEF1 as a positive regulator of RIPK1/RIPK3 activation in vitro. Based on the overexpression and knockdown experiments, we determined that RABGEF1 accelerated the phosphorylation of RIPK1 and promoted necrosome formation in L929 cells. The pro-necrotic effect of RABGEF1 is associated with its E3 ubiquitin ligase activity and guanine nucleotide exchange factor (GEF) activity. We further confirmed that RABGEF1 interacts with cIAP1 protein by inhibiting its function and plays a regulatory role in necroptosis, which can be abolished by treatment with the antagonist Smac mimetic (SM)-164. In conclusion, our study highlights a potential and novel role of RABGEF1 in promoting TNF-induced cell necrosis. • RABGEF1 promotes the phosphorylation of RIPK1 and formation of complex Ⅱb in L929 cells. • RABGEF1 does not affect TNF-induced activation of NF-kB and MAPK signaling pathways. • Knockdown of RABGEF1 levels inhibits TNF-induced necroptosis and necrosome formation. • The pro-necrotic effect of RABGEF1 is associated with the A20-type zinc finger and Vps9 domains. • RABGEF1 interacts with cIAP1 but does not affect cIAP1 protein levels. [ABSTRACT FROM AUTHOR]

Background: Obesity leads to a decline in the exercise capacity of skeletal muscle, thereby reducing mobility and promoting obesity-associated health risks. Dietary intervention has been shown to be an important measure to regulate skeletal muscle function, and previous studies have demonstrated the beneficial effects of docosahexaenoic acid (DHA; 22:6 ω-3) on skeletal muscle function. At the molecular level, DHA and its metabolites were shown to be extensively involved in regulating epigenetic modifications, including DNA methylation, histone modifications, and small non-coding microRNAs. However, whether and how epigenetic modification of mRNA such as N6-methyladenosine (m6A) mediates DHA regulation of skeletal muscle function remains unknown. Here, we analyze the regulatory effect of DHA on skeletal muscle function and explore the involvement of m6A mRNA modifications in mediating such regulation. Results: DHA supplement prevented HFD-induced decline in exercise capacity and conversion of muscle fiber types from slow to fast in mice. DHA-treated myoblasts display increased mitochondrial biogenesis, while slow muscle fiber formation was promoted through DHA-induced expression of PGC1α. Further analysis of the associated molecular mechanism revealed that DHA enhanced expression of the fat mass and obesity-associated gene (FTO), leading to reduced m6A levels of DNA damage-induced transcript 4 (Ddit4). Ddit4 mRNA with lower m6A marks could not be recognized and bound by the cytoplasmic m6A reader YTH domain family 2 (YTHDF2), thereby blocking the decay of Ddit4 mRNA. Accumulated Ddit4 mRNA levels accelerated its protein translation, and the consequential increased DDIT4 protein abundance promoted the expression of PGC1α, which finally elevated mitochondria biogenesis and slow muscle fiber formation. Conclusions: DHA promotes mitochondrial biogenesis and skeletal muscle fiber remodeling via FTO/m6A/DDIT4/PGC1α signaling, protecting against obesity-induced decline in skeletal muscle function. [ABSTRACT FROM AUTHOR]

Capacitor mismatch plays an important part in the spurious-free dynamic range (SFDR) performance of high-resolution successive approximation register (SAR) analog-to-digital converters (ADCs). This paper presents a symmetric group control dynamic element matching (SGCDEM) method aiming at enhancing SFDR of SAR ADCs applied in biomedical applications. The proposed scheme symmetrically divides the capacitor array into four sub-DACs, using quantization results corresponding to each sub-DAC as rotary control codes to symmetrically control the rotation of the capacitors. In order to improve the energy efficiency and save chip area, the proposed method replaces the pseudo-random number generator (PRNG) with quantization results of ADC as the rotary control code. Moreover, split rotator technology is applied in the proposed scheme to reduce the number of multiplexers (MUX) used in the rotator, which introduces a low logic depth. The proposed architecture has been implemented on a 12-bit SAR ADC model in MATLAB in order to verify its performances. After taking a standard deviation of 1% for the unit capacitance into consideration, this method achieves above 77.3 dB SFDR, enabling to correct the SFDR by more than 23.6 dB compared with the structure without DEM. [ABSTRACT FROM AUTHOR]

Juvenile brook trout were fed on six isonitrogenous and isolipidic diets, containing graded levels of gelatinized corn starch (50, 100, 150, 200, 250, 300 g/kg diet) for 81 days. Cellulose was used to compensate carbohydrate loss. The weight gain, specific growth rate and protein efficiency ratio increased as the dietary carbohydrate level increased from 50 to 150 g/kg, but decreased thereafter (p <.05). Quadratic regression analysis revealed that the optimum dietary carbohydrate level was in the range of 187.1–194.1 g/kg. In addition, serum alanine transaminase and aspartate transaminase activities, as well as glucose content, were highest in the group fed on 300 g/kg carbohydrate (p <.05). The hepatic malondialdehyde level increased with dietary carbohydrate levels (p <.05). The hepatic lysozyme activity increased as dietary carbohydrate level increased from 50 to 150 g/kg and decreased thereafter (p <.05). The overall survival rate after hypoxia challenge (45 min; dissolved oxygen content: 2.2 mg L−1) decreased with dietary carbohydrate levels (p <.05). Taken together, these results suggested that optimal level of dietary carbohydrate could improve growth performance. However, excessive dietary carbohydrate intake (> 250 g/kg) may decrease innate immunity status, increase oxidative stress and reduce resistance to hypoxia stress in brook trout. [ABSTRACT FROM AUTHOR]

Tumor cells, especially drug-resistant cells, predominately support growth by glycolysis even under the condition of adequate oxygen, which is known as the Warburg effect. Glucose metabolism reprogramming is one of the main factors causing tumor resistance. Previous studies on Shenmai injection (SMI), a Chinese herbal medicine, have shown enhanced efficacy in the treatment of tumors in combination with chemotherapy drugs, but the mechanism is not clear. In this study, we investigated the effect of SMI combined with cisplatin on cisplatin-resistant lung adenocarcinoma A549/DDP cells. Our results showed that cisplatin-resistant A549/DDP cells exhibited increased glucose consumption, lactate production, and expression levels of key glycolytic enzymes, including hexokinase 2 (HK2), pyruvate kinase M1/2 (PKM1/2), pyruvate kinase M2 (PKM2), glucose transporter 1 (GLUT1), and lactate dehydrogenase A (LDHA), compared with cisplatin-sensitive A549 cells. SMI combined with cisplatin in A549/DDP cells, led to significantly lower expression levels of key glycolytic enzymes, such as HK2, PKM1/2, GLUT1, and pyruvate dehydrogenase (PDH). In addition, we found that the combination of SMI and cisplatin could inhibit cell proliferation and promote apoptosis by reducing the expression levels of p-Akt, p-mTOR, and c-Myc, and then, it reduced the glycolysis level. These results suggest that SMI enhances the antitumor effect of cisplatin via glucose metabolism reprogramming. Therefore, the combination of SMI and cisplatin may be a potential therapeutic strategy to treat cisplatin-resistant nonsmall cell lung cancer. [ABSTRACT FROM AUTHOR]

An ultra-low power consumption high-linearity switching scheme for successive approximation register (SAR) analog-to-digital converter (ADC) is presented with a mixed switching method. Based on the combination of C-2C dummy capacitors, the charge sharing technique and monotonic switching method, the proposed switching method achieves high-energy saving and high linearity. Compared with the conventional SAR ADC, the proposed method consumes no reset energy and achieves 98.9% less switching energy and 87.2% reduction in capacitor area. Moreover, the proposed scheme obtains good performance in linearity. Furthermore, the common-mode voltage variation of the proposed scheme is smaller than other published schemes, which is important for decreasing input-dependent offset of the comparator. [ABSTRACT FROM AUTHOR]

Deep convolutional neural networks (CNNs) have shown their outstanding performance in the hyperspectral image (HSI) classification. The success of CNN-based HSI classification relies on the availability sufficient training samples. However, the collection of training samples is expensive and time consuming. Besides, there are many pretrained models on large-scale data sets, which extract the general and discriminative features. The proper reusage of low-level and midlevel representations will significantly improve the HSI classification accuracy. The large-scale ImageNet data set has three channels, but HSI contains hundreds of channels. Therefore, there are several difficulties to simply adapt the pretrained models for the classification of HSIs. In this article, heterogeneous transfer learning for HSI classification is proposed. First, a mapping layer is used to handle the issue of having different numbers of channels. Then, the model architectures and weights of the CNN trained on the ImageNet data sets are used to initialize the model and weights of the HSI classification network. Finally, a well-designed neural network is used to perform the HSI classification task. Furthermore, attention mechanism is used to adjust the feature maps due to the difference between the heterogeneous data sets. Moreover, controlled random sampling is used as another training sample selection method to test the effectiveness of the proposed methods. Experimental results on four popular hyperspectral data sets with two training sample selection strategies show that the transferred CNN obtains better classification accuracy than that of state-of-the-art methods. In addition, the idea of heterogeneous transfer learning may open a new window for further research. [ABSTRACT FROM AUTHOR]

An ultra-low power consumption two-stage mixed switching scheme for successive approximation register (SAR) analog-to-digital converter (ADC) is presented. Using two simple switches, the novel switching scheme divides the capacitor arrays into two sub-arrays: stage-one and stage-two arrays. The two sub-arrays convert high and low bits cycles, respectively. Once the high bits conversion cycles are completed, the corresponding sub-arrays are split off from the capacitor arrays. This decreases the number of capacitors used in the rest conversion procedure. Thus, the proposed method improves the energy efficiency of SAR ADC. Thanks to C–2C dummy capacitors and two-stage capacitor arrays, the novel architecture achieves 86% reduction in capacitor area than conventional SAR ADC. Furthermore, based on the charge sharing technique and monotonic switching method, the proposed switching scheme does not consume reset energy and achieves 99.8% less switching energy than the conventional switching method. In addition, the proposed scheme is less sensitive to capacitor mismatch because of its great performance in linearity. [ABSTRACT FROM AUTHOR]

Multistage interfacial desalination is a promising strategy to address global freshwater scarcity due to its ultra-high energy conversion efficiency. While numerous researches focus on materials improvement and system design, the underlying mechanisms of energy efficiency improvement are less explored. In this work, we establish a theoretical framework based on energy conservation and heat and mass transfer to comprehensively analyze the energy efficiency of multistage interfacial desalination. We demonstrate that there is an efficiency upper limit of 894 % under ideal conditions, which is mainly limited by the unavoidable sensible heat loss from the fresh water. Furthermore, we show that a typical ten-stage desalination system with a lab scale (10 cm × 10 cm) can reach a calculated efficiency of 232 %. An energy loss analysis allows us to identify another important heat transfer term, the wick where the evaporator takes seawater, which also causes significant heat loss. Experiments are conducted to show that better thermal insulation on wick can enhance energy efficiency. Moreover, we found that the system efficiency can be improved by enhancing the internal convective heat and mass transfer. This work provides insights into physical processes involved in multistage interfacial desalination and offers important guidance for further performance improvement towards its application. • Comprehensive analysis of the energy efficiency of multistage interfacial desalination was conducted. • Reveal that there is a solar-to-vapor efficiency upper limit of 894 % under ideal conditions • Quantitatively prove that better thermal insulation on wick can enhance energy efficiency under practical conditions • Suggest the system efficiency can be improved by enhancing the internal convective flow [ABSTRACT FROM AUTHOR]

Hyperspectral image (HSI) classification is a core task in the remote sensing community, and recently, deep learning-based methods have shown their capability of accurate classification of HSIs. Among the deep learning-based methods, deep convolutional neural networks (CNNs) have been widely used for the HSI classification. In order to obtain a good classification performance, substantial efforts are required to design a proper deep learning architecture. Furthermore, the manually designed architecture may not fit a specific data set very well. In this paper, the idea of automatic CNN for the HSI classification is proposed for the first time. First, a number of operations, including convolution, pooling, identity, and batch normalization, are selected. Then, a gradient descent-based search algorithm is used to effectively find the optimal deep architecture that is evaluated on the validation data set. After that, the best CNN architecture is selected as the model for the HSI classification. Specifically, the automatic 1-D Auto-CNN and 3-D Auto-CNN are used as spectral and spectral–spatial HSI classifiers, respectively. Furthermore, the cutout is introduced as a regularization technique for the HSI spectral–spatial classification to further improve the classification accuracy. The experiments on four widely used hyperspectral data sets (i.e., Salinas, Pavia University, Kennedy Space Center, and Indiana Pines) show that the automatically designed data-dependent CNNs obtain competitive classification accuracy compared with the state-of-the-art methods. In addition, the automatic design of the deep learning architecture opens a new window for future research, showing the huge potential of using neural architectures’ optimization capabilities for the accurate HSI classification. [ABSTRACT FROM AUTHOR]

Background: Previous experimental studies have suggested that the consumption of soy isoflavones may have a potential impact on lowering blood pressure. Nevertheless, epidemiological studies have presented conflicting outcomes concerning the correlation between soy isoflavone consumption and blood pressure levels. Consequently, a comprehensive meta-analysis of all eligible randomized controlled trials (RCTs) was conducted to explore the influence of soy isoflavones on systolic blood pressure (SBP) and diastolic blood pressure (DBP) in adults. Methods: A thorough search of PubMed, Embase, and the Cochrane Library for relevant literature up to April 30, 2023 was conducted. RCTs involving adults that compared soy isoflavone supplementation with a placebo (the same matrix devoid of soy isoflavone) were included. The combined effect size was presented as the weighted mean difference (WMD) along with 95% confidence interval (CI), employing a fixed-effects model. Results: Our meta-analysis included a total of 24 studies involving 1945 participants. The results revealed a significant reduction in both SBP and DBP with soy isoflavone supplementation. Subgroup analyses suggested more pronounced reductions in SBP and DBP for interventions lasting ≥6 months, in individuals receiving mixed-type soy isoflavone, and among patients with metabolic syndrome or prehypertension. However, we did not detect significant nonlinear associations between supplementation dosage and intervention duration concerning both SBP and DBP. The overall quality of evidence was deemed moderate. Conclusions: The current meta-analysis revealed that supplementation with soy isoflavones alone effectively reduces blood pressure. Additional high-quality studies are required to investigate the efficacy of blood pressure reduction through supplementation with an optimal quantity and proportion of soy isoflavone. [ABSTRACT FROM AUTHOR]

A generative adversarial network (GAN) usually contains a generative network and a discriminative network in competition with each other. The GAN has shown its capability in a variety of applications. In this paper, the usefulness and effectiveness of GAN for classification of hyperspectral images (HSIs) are explored for the first time. In the proposed GAN, a convolutional neural network (CNN) is designed to discriminate the inputs and another CNN is used to generate so-called fake inputs. The aforementioned CNNs are trained together: the generative CNN tries to generate fake inputs that are as real as possible, and the discriminative CNN tries to classify the real and fake inputs. This kind of adversarial training improves the generalization capability of the discriminative CNN, which is really important when the training samples are limited. Specifically, we propose two schemes: 1) a well-designed 1D-GAN as a spectral classifier and 2) a robust 3D-GAN as a spectral–spatial classifier. Furthermore, the generated adversarial samples are used with real training samples to fine-tune the discriminative CNN, which improves the final classification performance. The proposed classifiers are carried out on three widely used hyperspectral data sets: Salinas, Indiana Pines, and Kennedy Space Center. The obtained results reveal that the proposed models provide competitive results compared to the state-of-the-art methods. In addition, the proposed GANs open new opportunities in the remote sensing community for the challenging task of HSI classification and also reveal the huge potential of GAN-based methods for the analysis of such complex and inherently nonlinear data. [ABSTRACT FROM AUTHOR]

Triplicate groups of juvenile yellow catfish Pelteobagrus fulvidraco were exposed to three levels of α-ethinylestradiol (EE2) (0, 0.1 and 1 ng L −1 ) for 56 days. Fish survival rate (>93.33%) was not different among experimental groups. Weight gain and specific growth rate of fish exposed to EE2 were higher than those of control fish. Hepatosomatic index of fish exposed to 1 ng L −1 EE2 was the highest. Serum total protein, albumin, globulin, alanine aminotransferase, aspartate transaminase, cholesterol and triglyceride increased with increasing EE2 exposure levels. Liver total anti-oxidative capacity, malondialdehyde content and lysozyme activity of fish exposed to EE2 were higher than those of control fish. Phagocytic indices of fish exposed to 1 ng L −1 EE2 was lower than that of control fish. This study indicates that although EE2 exposure can promote the growth of yellow catfish in short-term, EE2 exerts its toxic effects by inducing reactive oxygen species generation and malondialdehyde accumulation, leading to blood deterioration and interfering with immune response. [ABSTRACT FROM AUTHOR]

It is useful and challenging to analyze and select object features of very high resolution (VHR) remote sensing imagery. The overwhelming majority of existing feature selection methods always concatenate all of the features into a long feature vector and then select features from the vector, ignoring the homogeneity and heterogeneity of underlying feature subspaces. In this paper, we propose a supervised multiview feature selection (SMFS) method. Unlike the existing multiview methods, SMFS requires no prior knowledge of the number of views, and is independent of a prefixed classifier. By utilizing homogeneity and heterogeneity of the data, SMFS employs affinity propagation to automatically decompose features into multiple disjoint and meaningful feature groups or views without any prior knowledge. A group or view consists of homogeneous features, describing a unique data characteristic. Different views represent heterogeneous data characteristics. Then, features are evaluated and selected based on joint \ell 1,2 -norm minimization of a loss function and a regularization term. Different from the popular \ell 2,1 -norm, joint \ell 1,2 -norm enforces the intraview sparsity, instead of interview sparsity. Consequently, a view can be represented by a few representative features in each view, and the information of heterogeneous views can be well kept by the remaining representative features. The experimental results on four VHR satellite images attest to the effectiveness and practicability of SMFS in comparison with single-view algorithms. Furthermore, some discussions are conducted to give insights into homogeneity and heterogeneity of features. [ABSTRACT FROM PUBLISHER]

In this letter, we present a new hyperspectral data-clustering method, named density analysis ensemble, from a different perspective. Instead of distance-based metrics in traditional clustering methods, we use density analysis for hyperspectral data clustering. Moreover, in order to improve the performance, we use the random subspace ensemble method to formulate a set of clustering systems. The final results are retrieved through majority voting. Compared to thek-means method, the overall accuracies have been improved by 7.05% and 6.93% for the Salinas and Pavia University data sets, respectively. [ABSTRACT FROM AUTHOR]

Due to the advantages of deep learning, in this paper, a regularized deep feature extraction (FE) method is presented for hyperspectral image (HSI) classification using a convolutional neural network (CNN). The proposed approach employs several convolutional and pooling layers to extract deep features from HSIs, which are nonlinear, discriminant, and invariant. These features are useful for image classification and target detection. Furthermore, in order to address the common issue of imbalance between high dimensionality and limited availability of training samples for the classification of HSI, a few strategies such as L2 regularization and dropout are investigated to avoid overfitting in class data modeling. More importantly, we propose a 3-D CNN-based FE model with combined regularization to extract effective spectral–spatial features of hyperspectral imagery. Finally, in order to further improve the performance, a virtual sample enhanced method is proposed. The proposed approaches are carried out on three widely used hyperspectral data sets: Indian Pines, University of Pavia, and Kennedy Space Center. The obtained results reveal that the proposed models with sparse constraints provide competitive results to state-of-the-art methods. In addition, the proposed deep FE opens a new window for further research. [ABSTRACT FROM AUTHOR]

The effects of electromagnetic stirring (EMS) including stirring time, applied voltage and rotational frequency on the microstructure of semisolid slurries of the Mg–Nd–Zn–(Zr) alloys were investigated. The results indicate that all of the parameters have large effects on the microstructure of the slurries. After being treated by EMS, the morphology of primary α-Mg phases in the Mg–3Nd–0.2Zn (NZ30) alloy slurries evolves from dendrite to three types: rosette, dendrite, and spheroid, while a semisolid microstructure with small and spheroidal particles is obtained in the Mg–3Nd–0.2Zn–0.4Zr (NZ30K) alloy slurries. The data for solid fraction with stirring time can be fitted to linear equations. The increase of applied voltage and rotational frequency makes the primary α-Mg phases initially refined and then coarsened. The optimal processing parameters are stirring time 120–180 s, applied voltage 300–350 V, rotational frequency ~ 20 Hz for NZ30 alloy, and stirring time 30–180 s, applied voltage ~ 350 V, rotational frequency ~ 20 Hz for NZ30K alloy. The average particle size of NZ30 alloy can be refined from ~ 879 to ~ 457 μm. Furthermore, the effects of applied voltage and rotational frequency on undercooling and melt temperature were also discussed on the basis of thermodynamic theory. [ABSTRACT FROM AUTHOR]

Object detection in remote-sensing images (RSIs) is always a vibrant research topic in the remote-sensing community. Recently, deep-convolutional-neural-network (CNN)-based methods, including region-CNN-based and You-Only-Look-Once-based methods, have become the de-facto standard for RSI object detection. CNNs are good at local feature extraction but they have limitations in capturing global features. However, the attention-based transformer can obtain the relationships of RSI at a long distance. Therefore, the Transformer for Remote-Sensing Object detection (TRD) is investigated in this study. Specifically, the proposed TRD is a combination of a CNN and a multiple-layer Transformer with encoders and decoders. To detect objects from RSIs, a modified Transformer is designed to aggregate features of global spatial positions on multiple scales and model the interactions between pairwise instances. Then, due to the fact that the source data set (e.g., ImageNet) and the target data set (i.e., RSI data set) are quite different, to reduce the difference between the data sets, the TRD with the transferring CNN (T-TRD) based on the attention mechanism is proposed to adjust the pre-trained model for better RSI object detection. Because the training of the Transformer always needs abundant, well-annotated training samples, and the number of training samples for RSI object detection is usually limited, in order to avoid overfitting, data augmentation is combined with a Transformer to improve the detection performance of RSI. The proposed T-TRD with data augmentation (T-TRD-DA) is tested on the two widely-used data sets (i.e., NWPU VHR-10 and DIOR) and the experimental results reveal that the proposed models provide competitive results (i.e., centuple mean average precision of 87.9 and 66.8 with at most 5.9 and 2.4 higher than the comparison methods on the NWPU VHR-10 and the DIOR data sets, respectively) compared to the competitive benchmark methods, which shows that the Transformer-based method opens a new window for RSI object detection. [ABSTRACT FROM AUTHOR]

In recent years, supervised learning-based methods have achieved excellent performance for hyperspectral image (HSI) classification. However, the collection of training samples with labels is not only costly but also time-consuming. This fact usually causes the existence of weak supervision, including incorrect supervision where mislabeled samples exist and incomplete supervision where unlabeled samples exist. Focusing on the inaccurate supervision and incomplete supervision, the weakly supervised classification of HSI is investigated in this paper. For inaccurate supervision, complementary learning (CL) is firstly introduced for HSI classification. Then, a new method, which is based on selective CL and convolutional neural network (SeCL-CNN), is proposed for classification with noisy labels. For incomplete supervision, a data augmentation-based method, which combines mixup and Pseudo-Label (Mix-PL) is proposed. And then, a classification method, which combines Mix-PL and CL (Mix-PL-CL), is designed aiming at better semi-supervised classification capacity of HSI. The proposed weakly supervised methods are evaluated on three widely-used hyperspectral datasets (i.e., Indian Pines, Houston, and Salinas datasets). The obtained results reveal that the proposed methods provide competitive results compared to the state-of-the-art methods. For inaccurate supervision, the proposed SeCL-CNN has outperformed the state-of-the-art method (i.e., SSDP-CNN) by 0.92%, 1.84%, and 1.75% in terms of OA on the three datasets, when the noise ratio is 30%. And for incomplete supervision, the proposed Mix-PL-CL has outperformed the state-of-the-art method (i.e., AROC-DP) by 1.03%, 0.70%, and 0.82% in terms of OA on the three datasets, with 25 training samples per class. [ABSTRACT FROM AUTHOR]

Recently, many convolutional neural network (CNN)-based methods have been proposed to tackle the classification task of hyperspectral images (HSI). In fact, CNN has become the de-facto standard for HSI classification. It seems that the traditional neural networks such as multi-layer perceptron (MLP) are not competitive for HSI classification. However, in this study, we try to prove that the MLP can achieve good classification performance of HSI if it is properly designed and improved. The proposed Modified-MLP for HSI classification contains two special parts: spectral–spatial feature mapping and spectral–spatial information mixing. Specifically, for spectral–spatial feature mapping, each input sample of HSI is divided into a sequence of 3D patches with fixed length and then a linear layer is used to map the 3D patches to spectral–spatial features. For spectral–spatial information mixing, all the spectral–spatial features within a single sample are feed into the solely MLP architecture to model the spectral–spatial information across patches for following HSI classification. Furthermore, to obtain the abundant spectral–spatial information with different scales, Multiscale-MLP is proposed to aggregate neighboring patches with multiscale shapes for acquiring abundant spectral–spatial information. In addition, the Soft-MLP is proposed to further enhance the classification performance by applying soft split operation, which flexibly capture the global relations of patches at different positions in the input HSI sample. Finally, label smoothing is introduced to mitigate the overfitting problem in the Soft-MLP (Soft-MLP-L), which greatly improves the classification performance of MLP-based method. The proposed Modified-MLP, Multiscale-MLP, Soft-MLP, and Soft-MLP-L are tested on the three widely used hyperspectral datasets. The proposed Soft-MLP-L leads to the highest OA, which outperforms CNN by 5.76%, 2.55%, and 2.5% on the Salinas, Pavia, and Indian Pines datasets, respectively. The obtained results reveal that the proposed models provide competitive results compared to the state-of-the-art methods, which shows that the MLP-based methods are still competitive for HSI classification. [ABSTRACT FROM AUTHOR]

In recent years, many convolutional neural network (CNN)-based methods have been proposed to address the scene classification tasks of remote sensing images. Since the number of training samples in RS datasets is generally small, data augmentation is often used to expand the training set. It is, however, not appropriate when original data augmentation methods keep the label and change the content of the image at the same time. In this study, label augmentation (LA) is presented to fully utilize the training set by assigning a joint label to each generated image, which considers the label and data augmentation at the same time. Moreover, the output of images obtained by different data augmentation is aggregated in the test process. However, the augmented samples increase the intra-class diversity of the training set, which is a challenge to complete the following classification process. To address the above issue and further improve classification accuracy, Kullback–Leibler divergence (KL) is used to constrain the output distribution of two training samples with the same scene category to generate a consistent output distribution. Extensive experiments were conducted on widely-used UCM, AID and NWPU datasets. The proposed method can surpass the other state-of-the-art methods in terms of classification accuracy. For example, on the challenging NWPU dataset, competitive overall accuracy (i.e., 91.05%) is obtained with a 10% training ratio. [ABSTRACT FROM AUTHOR]

Obesity has become a health threat and hard enough to deal with. Evidences show that metformin could inhibit adipogenesis and combat obesity, while its mechanisms remain to be elucidated more comprehensively. In this study, we found that administration of metformin could combat obesity of mice induced by high-fat diet (HFD), indicated by strikingly decreased body weight and weight of inguinal white adipose tissue (iWAT) and epidydimal white adipose tissue (eWAT) compared with the control group. Mechanically, we revealed that metformin could inhibit protein expression of FTO, leading to increased m6A methylation levels of cyclin D1 (Ccnd1) and cyclin dependent kinase 2 (Cdk2), two crucial regulators in cell cycle. Ccnd1 and Cdk2 with increased m6A levels were recognised by YTH m6A RNA binding protein 2 (YTHDF2), causing an YTHDF2-dependent decay and decreased protein expressions. In consequence, mitotic clonal expansion (MCE) process was blocked and adipogenesis was inhibited. [ABSTRACT FROM AUTHOR]

• Breakthrough nanoparticle to distinguish hemorrhagic plaque from other vulnerable plaques via ultrasound, photoacoustic, and magnetic resonance multimodality imaging. • Effective strategy to fill the gap of treatment about intraplaque hemorrhage. • Novel strategy for early diagnosis and visualized treatment of intraplaque hemorrhage. Intraplaque hemorrhage (IPH) is a critical contributor to atherosclerotic plaque progression and destabilization; however, its timely detection and treatment of IPH are limited. In this study, novel macrophage-targeted nanoparticles (NPs) are developed that encapsulate perfluoropentane (PFP) in the core, load MnFe 2 O 4 and ferrostatin-1 (Fer-1) in the polylactic acid-glycolic acid (PLGA) shell, and conjugate CTHRSSVVC peptide on the surface. These NPs exhibit highly sensitive and specific magnetic resonance/photoacoustic/ultrasound multimodal imaging to distinguish the hemorrhagic plaque from other vulnerable plaques accurately. Further, the distribution could be observed in real-time, conducive to individual visualized therapy. The NPs could actively target the M(Hb) cell, which plays a critical role in clearing erythrocyte lysed products within IPH area. Additionally, Fer-1 loaded in the NPs exhibits a controlled release with non-invasive low-intensity focused ultrasound and further rescues the M(Hb) ferroptosis due to overloaded hemoglobin by scavenging lipid peroxides and chelating iron directly, subsequently stabilizing the hemorrhagic plaque without any toxic biological effects. Overall, these intricately designed NPs introduce a safe and effective theranostic strategy for IPH in susceptible populations, which is promising for further clinical applications. [ABSTRACT FROM AUTHOR]

The long-term impact of COVID-19 among those with mild infections is not well characterized. Among 81 adults who completed online assessments at 3- and 12-months following infection, quality of life scores did not significantly improve over time. Among 62 subjects who also completed telephone interviews, respiratory symptoms or exercise limitation were reported by 42% at a median follow-up of 387 days (IQR 251–402 days). Those with persistent respiratory symptoms scored lower on the EQ-5D visual analog score compared to those without. Persistent respiratory symptoms were associated with a lower likelihood of full-time employment at 1 year (aOR 0.09, 95%CI 0.01–0.91; P = 0.041). In an adjusted linear regression, persistent respiratory symptoms (P = 0.037) and female sex (P = 0.016) were both independent risks for increased visits to a primary care provider. This cohort study demonstrates that respiratory symptoms are frequent at 1 year following COVID-19 and more importantly, are associated with negative impacts on employment, quality of life, and health care utilization. Further research is needed to determine the pathophysiology and risk factors for persistent symptoms as well as optimal management strategies to improve the level of functioning and quality of life. [ABSTRACT FROM AUTHOR]

Recently, a great many deep convolutional neural network (CNN)-based methods have been proposed for hyperspectral image (HSI) classification. Although the proposed CNN-based methods have the advantages of spatial feature extraction, they are difficult to handle the sequential data with and CNNs are not good at modeling the long-range dependencies. However, the spectra of HSI are a kind of sequential data, and HSI usually contains hundreds of bands. Therefore, it is difficult for CNNs to handle HSI processing well. On the other hand, the Transformer model, which is based on an attention mechanism, has proved its advantages in processing sequential data. To address the issue of capturing relationships of sequential spectra in HSI in a long distance, in this study, Transformer is investigated for HSI classification. Specifically, in this study, a new classification framework titled spatial-spectral Transformer (SST) is proposed for HSI classification. In the proposed SST, a well-designed CNN is used to extract the spatial features, and a modified Transformer (a Transformer with dense connection, i.e., DenseTransformer) is proposed to capture sequential spectra relationships, and multilayer perceptron is used to finish the final classification task. Furthermore, dynamic feature augmentation, which aims to alleviate the overfitting problem and therefore generalize the model well, is proposed and added to the SST (SST-FA). In addition, to address the issue of limited training samples in HSI classification, transfer learning is combined with SST, and another classification framework titled transferring-SST (T-SST) is proposed. At last, to mitigate the overfitting problem and improve the classification accuracy, label smoothing is introduced for the T-SST-based classification framework (T-SST-L). The proposed SST, SST-FA, T-SST, and T-SST-L are tested on three widely used hyperspectral datasets. The obtained results reveal that the proposed models provide competitive results compared to the state-of-the-art methods, which shows that the concept of Transformer opens a new window for HSI classification. [ABSTRACT FROM AUTHOR]

• A three-bed six-step temperature swing adsorption cycle is used to save energy. • Two cases of the three-bed cycle are numerically studied by a mathematic model. • Around 40% reduction in heat consumption is achieved by the three-bed cycle. To avoid the waste heat exhaust of a gas dryer based on the conventional two-bed five-step temperature swing adsorption (TSA) cycle, we designed and developed a three-bed six-step TSA cycle with heat carrier gas (HCG) recycling. To clarify the impact of the retrofit from the two-bed system to the three-bed system on the gas dryer performance, numerical studies on the three- and two-bed systems were carried out in this work. Firstly, a nonisothermal, nonadiabatic, nonequilibrium mathematical model of the TSA process was established. Secondly, an industrial-scale dryer with electrical HCG heaters was provided for case study, and two concrete cases were studied to examine the effect of regeneration pressure on the energy-saving potential of the three-bed system. Thirdly, the bed profiles and historical curves of the online adsorption and offline regeneration of the two systems were obtained by simulations, thereby the two systems were comparatively analyzed in terms of energy efficiency, drying performance and head loss. Moreover, the economic analysis on the retrofit from the two-bed system to the three-bed system were carried out. In conclusion, the case studies show that, with a comparable drying performance, the three-bed system saved around 40% energy compared with the conventional two-bed system. The economic analysis shows that the additional investment for the retrofit from the two-bed system to the three-bed system is convenient since the payback periods are around one year. [ABSTRACT FROM AUTHOR]

Hyperspectral remote sensing obtains abundant spectral and spatial information of the observed object simultaneously. It is an opportunity to classify hyperspectral imagery (HSI) with a fine-grained manner. In this study, the fine-grained classification of HSI, which contains a large number of classes, is investigated. On one hand, traditional classification methods cannot handle fine-grained classification of HSI well; on the other hand, deep learning methods have shown their powerfulness in fine-grained classification. So, in this paper, deep learning is explored for HSI supervised and semi-supervised fine-grained classification. For supervised HSI fine-grained classification, densely connected convolutional neural network (DenseNet) is explored for accurate classification. Moreover, DenseNet is combined with pre-processing technique (i.e., principal component analysis or auto-encoder) or post-processing technique (i.e., conditional random field) to further improve classification performance. For semi-supervised HSI fine-grained classification, a generative adversarial network (GAN), which includes a discriminative CNN and a generative CNN, is carefully designed. The GAN fully uses the labeled and unlabeled samples to improve classification accuracy. The proposed methods were tested on the Indian Pines data set, which contains 33,3951 samples with 52 classes. The experimental results show that the deep learning-based methods provide great improvements compared with other traditional methods, which demonstrate that deep models have huge potential for HSI fine-grained classification. [ABSTRACT FROM AUTHOR]

Capsule networks can be considered to be the next era of deep learning and have recently shown their advantages in supervised classification. Instead of using scalar values to represent features, the capsule networks use vectors to represent features, which enriches the feature presentation capability. This paper introduces a deep capsule network for hyperspectral image (HSI) classification to improve the performance of the conventional convolutional neural networks (CNNs). Furthermore, a modification of the capsule network named Conv-Capsule is proposed. Instead of using full connections, local connections and shared transform matrices, which are the core ideas of CNNs, are used in the Conv-Capsule network architecture. In Conv-Capsule, the number of trainable parameters is reduced compared to the original capsule, which potentially mitigates the overfitting issue when the number of available training samples is limited. Specifically, we propose two schemes: (1) A 1D deep capsule network is designed for spectral classification, as a combination of principal component analysis, CNN, and the Conv-Capsule network, and (2) a 3D deep capsule network is designed for spectral-spatial classification, as a combination of extended multi-attribute profiles, CNN, and the Conv-Capsule network. The proposed classifiers are tested on three widely-used hyperspectral data sets. The obtained results reveal that the proposed models provide competitive results compared to the state-of-the-art methods, including kernel support vector machines, CNNs, and recurrent neural network. [ABSTRACT FROM AUTHOR]

The study evaluated the effects of different types of dietary carbohydrates on growth, immunity, antioxidant capacity and glucose metabolism enzyme activities in the juvenile brook trout. Fish (13.65 ± 0.06 g) were fed five isonitrogenous and isolipidic diets that contained 150g kg‐1 glucose, sucrose, dextrin, pregelatinized corn starch (PS) or raw corn starch (RS) respectively. Weight gain, specific growth rate and protein efficiency ratio were significantly higher in fish fed dextrin than the other treatments. The serum alanine aminotransferase (ALT), aspartate transaminase (AST), alkaline phosphatase (AKP) and glucose were higher in fish fed glucose than those fed dextrin, PS or RS. Hepatic superoxide dismutase (SOD), catalase (CAT) and lysozyme activity, as well as total antioxidant capacity, were lower in fish fed glucose or sucrose than the other treatments. Moreover, the highest malondialdehyde (MDA) level was reported in the glucose group. Furthermore, fish fed glucose had significantly lower hepatic glucokinase (GK), phosphofructokinase (PFK) and phosphoenolpyruvate carboxykinase (PEPCK) activities than those fed dextrin, PS or RS. The results suggest that dextrin is the optimal carbohydrate source for brook trout, while incorporating glucose into the diet suppresses innate immunity, increases oxidative stress and interferes with glucose metabolism, resulting into a reduction in growth. [ABSTRACT FROM AUTHOR]

Deep learning-based methods, especially deep convolutional neural network (CNN), have proven their powerfulness in hyperspectral image (HSI) classification. On the other hand, ensemble learning is a useful method for classification task. In this letter, in order to further improve the classification accuracy, the combination of CNN and random forest (RF) is proposed for HSI classification. The well-designed CNN is used as individual classifier to extract the discriminant features of HSI and RF randomly selects the extracted features and training samples to formulate a multiple classifier system. Furthermore, the learned weights of CNN are adopted to initialize other individual CNN. Experimental results with two hyperspectral data sets indicate that the proposed method provides competitive classification results compared with state-of-the-art methods. [ABSTRACT FROM AUTHOR]

Necroptosis is a form of regulated cell death that has been implicated in multiple diseases. TNF-induced necroptosis is regulated by necrosomes, complexes consisting of RIPK1, RIPK3 and MLKL. In this study, by screening of a small-compound library, we identified dozens of compounds that inhibited TNF-induced necroptosis. According to the mechanisms by which they inhibited necroptosis, these compounds were classified into different groups. We then identified Ibrutinib as an inhibitor of RIPK3 and found that Quizartinib protected against the TNF-induced systemic inflammatory response syndrome in mice by inhibiting the activation of RIPK1. Altogether, our work revealed dozens of necroptosis inhibitors, suggesting new potential approaches for treating necroptosis-related diseases. Screening a library of compounds identifies several inhibitors of TNF-induced necroptosis, including Ibrutinib as an inhibitor of RIPK3 and Quizartinib as an indirect RIPK1 kinase activity inhibitor. [ABSTRACT FROM AUTHOR]

The fundamental behaviour of stainless steel tubular hollow section members under axial cyclic loading is presented in this paper. An experimental investigation on eight test specimens with varying global and local slenderness is conducted first, where the strength, fracture life, ductility, and energy dissipation capability are discussed in detail. A numerical study is then carried out to further elaborate the deformation/stress pattern, global and local buckling behaviour, and imperfection sensitivity of the members. The subsequent parametric study considers a broadened parameter matrix and thus enables a more comprehensive understanding of the global and local buckling behaviour of stainless steel tubular hollow section members. It is generally observed that the major stainless steel design codes provide conservative predictions for the compressive resistance of the members under cyclic loading, and a large dispersion of the FE-to-predicted ratio exists. A Continuous Strength Method (CSM) is shown to provide more consistent predictions. Due to the diverse failure modes observed in the parametric study models, large inconsistency also exists in the prediction of the post-buckling resistance of the FE models. A ductility-oriented design approach is proposed to enable a quick yet reliable prediction of the available compressive ductility of cyclically loaded stainless steel members. [ABSTRACT FROM AUTHOR]

First-line treatment for osteosarcoma includes chemotherapy and surgery. However, the five-year survival rate of refractory osteosarcoma remains unsatisfactory. Osteosarcoma cancer stem cells, possessing stemness and chemoresistance, are one of the critical causes of poor response to chemotherapy. Elucidating regulatory signaling pathways of osteosarcoma cancer stem cells may provide a rationale for improving regimens against chemoresistant osteosarcoma. Methotrexate (MTX)-resistant osteosarcoma cells were established. microRNA expression profiles were used for detecting differentially expressed microRNA in resistant clones and the parental cells. microRNA target databases were employed to predict potential microRNA and mRNA interactions. Flow cytometry was performed to measure stem cell marker Prominin-1 (CD133)-positive cells. Immunofluorescence staining was applied to detect CD133 expression. miR-197-3p mimic or anti-miR-197-3p stably transfected cells were used to generate xenograft models. In the study, we found that miR-197-3p was increased in MTX-resistant cell lines. Overexpression of miR-197-3p enhanced the expression of cancer stem cell markers CD133, Octamer-binding protein 4 (OCT4), Transcription factor SOX-2 (SOX2), and Homeobox protein NANOG (NANOG), as well as chemoresistance-associated genes ATP-dependent translocase ABCB1 (ABCB1) and Broad substrate specificity ATP-binding cassette transporter ABCG2 (ABCG2), whereas miR-197-3p knockdown inhibited stemness and recovered sensitivity to MTX. We also classified the tumor suppressor Speckle-type POZ protein-like (SPOPL) as a target of miR-197-3p. The miR-197-3p mutation that could not combine SPOPL promoter regions was unable to sustain stemness or chemoresistance. Collectively, we discovered miR-197-3p conferred osteosarcoma stemness and chemotherapy resistance by targeting SPOPL, prompting promising therapeutic candidates for refractory osteosarcoma treatment. [ABSTRACT FROM AUTHOR]

Although 5-methylcytosine (m5C) has been identified as a novel and abundant mRNA modification and associated with energy metabolism, its regulation function in adipose tissue and skeletal muscle is still limited. This study aimed at investigating the effect of mRNA m5C on adipogenesis and myogenesis using Jinhua pigs (J), Yorkshire pigs (Y) and their hybrids Yorkshire–Jinhua pigs (YJ). We found that Y grow faster than J and YJ, while fatness-related characteristics observed in Y were lower than those of J and YJ. Besides, total mRNA m5C levels and expression rates of NSUN2 were higher both in backfat layer (BL) and longissimus dorsi muscle (LDM) of Y compared to J and YJ, suggesting that higher mRNA m5C levels positively correlate with lower fat and higher muscle mass. RNA bisulfite sequencing profiling of m5C revealed tissue-specific and dynamic features in pigs. Functionally, hyper-methylated m5C-containing genes were enriched in pathways linked to impaired adipogenesis and enhanced myogenesis. In in vitro, m5C inhibited lipid accumulation and promoted myogenic differentiation. Furthermore, YBX2 and SMO were identified as m5C targets. Mechanistically, YBX2 and SMO mRNAs with m5C modification were recognized and exported into the cytoplasm from the nucleus by ALYREF, thus leading to increased YBX2 and SMO protein expression and thereby inhibiting adipogenesis and promoting myogenesis, respectively. Our work uncovered the critical role of mRNA m5C in regulating adipogenesis and myogenesis via ALYREF-m5C-YBX2 and ALYREF-m5C-SMO manners, providing a potential therapeutic target in the prevention and treatment of obesity, skeletal muscle dysfunction and metabolic disorder diseases. [ABSTRACT FROM AUTHOR]

• FTO knockout increased the secretion of IL-6 in adipocytes. • Adipocyte-derived IL-6 promoted lipolysis genes expressions in the liver. • FTOAKO alleviated HFD-induced hepatic steatosis in mice. Adipose dysfunction affects the secretion of adipokines and mediates the hepatic physiological changes. Fat mass and obesity associated protein (FTO) plays a crucial part in fat deposition but the crosstalk between FTO-mediated secretion of adipokines and hepatic steatosis is not clear. Firstly, adipose-selective FTO knockout (FTOAKO) and control (FTOflox/flox) mice were induced by high fat diet (HFD). Then qRT-PCR assay was performed to analyze the expressions of hepatic lipid metabolism genes and adipocytokines gene of inguinal white adipose tissue (iWAT) and epididymal white adipose tissue (eWAT). Afterwards, 3T3-L1 cells were knocked out IL-6 and co-cultured with AML12 cells (3T3-L1 siIL-6/AML12) and the expressions of hepatic lipid lipolysis genes were measured. Finally, we detected the hepatic lipid metabolism genes expressions in AML12 cells with the medium from 3T3-L1 cells or IL-6 treatment. FTOAKO effectively alleviated HFD-induced hepatic steatosis in mice and improved the transcription level of genes involved in hepatic lipolysis. Further investigation demonstrated that FTO knockout increased level of IL-6 in adipose tissues and 3T3-L1 cells. Compared to 3T3-L1/AML12, our results showed lipolysis-related genes expressions were dramatically inhibited in 3T3-L1 siIL-6/AML12. Finally, both depletion of FTO in adipocytes and IL-6 supplement led to increased lipolysis genes expressions in AML12 cells. FTO knockout in adipose tissue alleviated hepatic steatosis via targeting adipocyte-derived IL-6. [ABSTRACT FROM AUTHOR]

Jamming is a big threat to the survival of a radar system. Therefore, the recognition of radar jamming signal type is a part of radar countermeasure. Recently, convolutional neural networks (CNNs) have shown their effectiveness in radar signal processing, including jamming signal recognition. However, most of existing CNN methods do not regard radar jamming as a complex value signal. In this study, a complex-valued CNN (CV-CNN) is investigated to fully explore the inherent characteristics of a radar jamming signal, and we find that we can obtain better recognition accuracy using this method compared with a real-valued CNN (RV-CNN). CV-CNNs contain more parameters, which need more inference time. To reduce the parameter redundancy and speed up the recognition time, a fast CV-CNN (F-CV-CNN), which is based on pruning, is proposed for radar jamming signal fast recognition. The experimental results show that the CV-CNN and F-CV-CNN methods obtain good recognition performance in terms of accuracy and speed. The proposed methods open a new window for future research, which shows a huge potential of CV-CNN-based methods for radar signal processing. [ABSTRACT FROM AUTHOR]

• MILD combustion technology applies to the heat flue of a coke oven. • A novel type of diffusion heat flue structure and combustion method is proposed. • Effects of gas composition on the thermal process of the heat flue were studied. • Low NOx combustion performance with different ratios of mixed gas was compared. • The influence of gas composition on NOx emission is also studied. Based on high-temperature and low-oxygen diffusion clean combustion technology, a novel-type of heat flue in a coke oven was designed in this paper. By changing the internal structure of the flue and the supply mode of the combustion-supporting gas, fuel zoning and diffusion combustion in the heat flue are achieved. The combustion temperature in the flue and the formation of thermal NOx are greatly reduced. The flow, temperature and NOx concentration fields inside the flue are studied though a three-dimensional finite element numerical model. The results show that when coke oven gas is used as the fuel, the NOx concentration at the outlet of the diffused heat flue can be reduced to 312.89 mg/m3, which reduces the emission of nitrogen oxides by nearly 35%, compared with the result using graded heat flue. In addition, the influence of gas composition on NOx emission is also studied, and the mathematical relationship involving the gas composition (α), the maximum temperature (T max), and the outlet NOx concentration (C NOx) of the flue is obtained. When the preheating temperature reaches 550 K, T ave reaches 1626.76 K, and the outlet NOx concentration reaches 144.07 mg/m3, which can meet the standards for emissions and the heating requirements. [ABSTRACT FROM AUTHOR]