Search

The deposited pesticide distribution in fruit tree canopies is crucial for evaluating the efficacy of air-assisted spraying in orchards. Most studies have determined the impact of pesticide application on pesticide deposition on canopies without a quantitative computational model. In this study, an air-assisted orchard sprayer with airflow control was used to perform spraying experiments on artificial and peach trees. In the spraying experiment on an artificial tree, a canopy with leaf areas ranging from 2.54~5.08 m2 was found to require an effective air speed of 18.12~37.05 m/s. The canopy leaf area, air speed at the sprayer fan outlet and spray distance were used as test factors in a three-factor five-level quadratic general rotational orthogonal test to develop a computational model for pesticide deposition at the inner, outer and middle regions of a fruit tree canopy with R2 values of 0.9042, 0.8575 and 0.8199, respectively. A significance analysis was used to rank the influencing factors for the deposited pesticide distribution in decreasing order of significance as follows: the spray distance, leaf area and air speed for the inner region of the canopy, followed by the spray distance, air speed and leaf area for the middle and outer regions of the canopy. The results of the verification test conducted in a peach orchard showed that the computational errors of the pesticide deposition model for the inner, middle and outer regions of the canopy were 32.62%, 22.38% and 23.26%, respectively. The results provide support for evaluating the efficacy of an air-assisted orchard sprayer and optimizing the sprayer parameters.

The hybrid LCC/VSC cascaded HVDC (HC-HVDC) scheme, in which the LCC is adopted as rectifier, and one LCC in series with multiple paralleled VSC based MMCs is adopted as inverter, combines the advantages of both LCC and VSC-MMC technologies. Presently, a HC-HVDC project is planned to deliver the power from Baihetan to Jiangsu in China. One of the major challenges is to suppress the DC overcurrent issue of VSC-MMCs especially under commutation failure condition of LCC at inverter side. Based on Baihetan-Jiangsu project under construction, the model of HC-HVDC system is developed in PSCAD/EMTDC. In the system, the rectifier employs a dual 12-pulse LCC, while the inverter adopts a hybrid scheme, i.e., one LCC converter in series with three paralleled VSC-MMC converters. This paper addresses the main cause and feature of the DC overcurrent, then proposes an overcurrent suppression control method for HC-HVDC system based on fuzzy clustering and identification approach. And the system transient performances with the proposed control approach are evaluated under AC fault conditions. The results show that the proposed approach can not only suppress the DC overcurrent following the commutation failure to a certain extent, but also improve the post-fault recovery performance of the overall system.

BackgroundBlisters are tense vesicles or bullae that arise on swollen skin and are found in a wide range of injuries. As a complication of fracture, fracture blisters are considered soft tissue injuries, which often lead to adverse effects such as prolonged preoperative waiting time and increased risk of surgical site infection. However, our previous study found that in patients with acute compartment syndrome, fracture blisters may be a form of compartment pressure release, but the specific mechanism has not been revealed. Here, we mapped out the proteomic landscape of fracture blister fluid for the first time and compared its expression profile to cupping and burn blisters.MethodsFirst, fluid samples were collected from 15 patients with fracture blisters, 7 patients with cupping blisters, and 9 patients with burn blisters. Then, the expression levels of 92 inflammatory proteins were measured using the Olink Target 96 Inflammation panel. Protein profiles were compared across the three groups using Differential Protein Expression Analysis and Principal Component Analysis (PCA).ResultsFracture blisters had significantly higher levels of 50 proteins in comparison to cupping and 26 proteins in comparison to burn blisters. Notably, PCA showed fracture blisters closely resembled the protein expression profile of burn blisters but were distinct from the protein expression profile of cupping blisters.ConclusionOur study provides the first characterization of fracture blister fluid using proteomics, which provides a valuable reference for further analysis of the difference between blisters caused by fractures and those caused by other pathogenic factors. This compendium of proteomic data provides valuable insights and a rich resource to better understand fracture blisters.

APS (antiphospholipid syndrome) is a systematic autoimmune disease presenting with the high levels of aPLs (antiphospholipid antibodies). These autoantibodies are involved in various clinical manifestations, mainly including arterial or venous thrombosis formation, proinflammatory response, and recurrent pregnant loss. Pyroptosis is a form of lytic programmed cell death, and it aggravates autoimmune diseases progression via activating NOD-like receptors, especially the NLRP3 inflammasome and its downstream inflammatory factors IL (interleukin)-1β and IL-18. However, the underlying mechanisms of pyroptosis-induced APS progression remain to be elucidated. ECs (endothelial cells), monocytes, platelets, trophoblasts, and neutrophils are prominent participants in APS development. Of significance, pyroptosis of APS-related cells leads to the excessive release of proinflammatory and prothrombotic factors, which are the primary contributors to APOs (adverse pregnancy outcomes), thrombosis formation, and autoimmune dysfunction in APS. Furthermore, pyroptosis-associated medicines have made encouraging advancements in attenuating inflammation and thrombosis. Given the potential of pyroptosis in regulating APS development, this review would systematically expound the molecular mechanisms of pyroptosis, and elaborate the role of pyroptosis-mediated cellular effects in APS progression. Lastly, the prospective therapeutic approaches for APS would be proposed based on the regulation of pyroptosis.

Electroencephalography (EEG) inverse problem is a typical inverse problem, in which the electrical activity within the brain is reconstructed based on EEG data collected from the scalp electrodes. In this paper, the four-layer concentric head model is used for simulation firstly, four deep neural network models including a multilayer perceptron (MLP) model and three convolutional neural networks (CNNs) are adopted to solve EEG inverse problem based on equal current dipole (ECD) model. In the simulations, 100,000 samples are generated randomly, of which 60% are used for network training and 20% are used for cross-validation. Eventually, the generalization performance of the model using the optimal function is measured by the errors in the rest 20% testing set. The experimental results show that the absolute error, relative error, mean positioning error and standard deviation of the four models are extremely low. The CNN with 6 convolutional layers and 3 pooling layers (CNN-3) is the best model. Its absolute error is about 0.015, its relative error is about 0.005, and its dipole position error is 0.040±0.029 cm. Furthermore, we use CNN-3 for source localization of the real EEG data in Working Memory. The results are in accord with physiological experience. The deep neural network method in our study needs fewer calculation parameters, takes less time, and has better positioning results.

Imitation learning has been widely applied to various autonomous systems thanks to recent development in interactive algorithms that address covariate shift and compounding errors induced by traditional approaches like behavior cloning. However, existing interactive imitation learning methods assume access to one perfect expert. Whereas in reality, it is more likely to have multiple imperfect experts instead. In this paper, we propose MEGA-DAgger, a new DAgger variant that is suitable for interactive learning with multiple imperfect experts. First, unsafe demonstrations are filtered while aggregating the training data, so the imperfect demonstrations have little influence when training the novice policy. Next, experts are evaluated and compared on scenarios-specific metrics to resolve the conflicted labels among experts. Through experiments in autonomous racing scenarios, we demonstrate that policy learned using MEGA-DAgger can outperform both experts and policies learned using the state-of-the-art interactive imitation learning algorithm. The supplementary video can be found at https://youtu.be/pYQiPSHk6dU.

We present SLoMo: a first-of-its-kind framework for transferring skilled motions from casually captured "in the wild" video footage of humans and animals to legged robots. SLoMo works in three stages: 1) synthesize a physically plausible reconstructed key-point trajectory from monocular videos; 2) optimize a dynamically feasible reference trajectory for the robot offline that includes body and foot motion, as well as contact sequences that closely tracks the key points; 3) track the reference trajectory online using a general-purpose model-predictive controller on robot hardware. Traditional motion imitation for legged motor skills often requires expert animators, collaborative demonstrations, and/or expensive motion capture equipment, all of which limits scalability. Instead, SLoMo only relies on easy-to-obtain monocular video footage, readily available in online repositories such as YouTube. It converts videos into motion primitives that can be executed reliably by real-world robots. We demonstrate our approach by transferring the motions of cats, dogs, and humans to example robots including a quadruped (on hardware) and a humanoid (in simulation). To the best knowledge of the authors, this is the first attempt at a general-purpose motion transfer framework that imitates animal and human motions on legged robots directly from casual videos without artificial markers or labels.

As one of the most fundamental techniques in multimodal learning, cross-modal matching aims to project various sensory modalities into a shared feature space. To achieve this, massive and correctly aligned data pairs are required for model training. However, unlike unimodal datasets, multimodal datasets are extremely harder to collect and annotate precisely. As an alternative, the co-occurred data pairs (e.g., image-text pairs) collected from the Internet have been widely exploited in the area. Unfortunately, the cheaply collected dataset unavoidably contains many mismatched data pairs, which have been proven to be harmful to the model's performance. To address this, we propose a general framework called BiCro (Bidirectional Cross-modal similarity consistency), which can be easily integrated into existing cross-modal matching models and improve their robustness against noisy data. Specifically, BiCro aims to estimate soft labels for noisy data pairs to reflect their true correspondence degree. The basic idea of BiCro is motivated by that -- taking image-text matching as an example -- similar images should have similar textual descriptions and vice versa. Then the consistency of these two similarities can be recast as the estimated soft labels to train the matching model. The experiments on three popular cross-modal matching datasets demonstrate that our method significantly improves the noise-robustness of various matching models, and surpass the state-of-the-art by a clear margin., Comment: CVPR 2023

There are various research strategies used for non-Hermitian systems, which typically involve introducing non-Hermitian terms to pre-existing Hermitian Hamiltonians. It can be challenging to directly design non-Hermitian many-body models that exhibit unique features not found in Hermitian systems. In this Letter, we propose a new method to construct non-Hermitian many-body systems by generalizing the parent Hamiltonian method into non-Hermitian regimes. This allows us to build a local Hamiltonian using given matrix product states as its left and right ground states. We demonstrate this method by constructing a non-Hermitian spin-$1$ model from the asymmetric Affleck-Kennedy-Lieb-Tasaki (AKLT) state, which preserves both chiral order and symmetry-protected topological order. Our approach opens up a new paradigm for systematically constructing and studying non-Hermitian many-body systems, providing guiding principles to explore new properties and phenomena in non-Hermitian physics., Comment: 14 pages, 10 figures, to appear in Physical Review Letters

Supplementary Material 9

We consider perception-based control using state estimates that are obtained from high-dimensional sensor measurements via learning-enabled perception maps. However, these perception maps are not perfect and result in state estimation errors that can lead to unsafe system behavior. Stochastic sensor noise can make matters worse and result in estimation errors that follow unknown distributions. We propose a perception-based control framework that i) quantifies estimation uncertainty of perception maps, and ii) integrates these uncertainty representations into the control design. To do so, we use conformal prediction to compute valid state estimation regions, which are sets that contain the unknown state with high probability. We then devise a sampled-data controller for continuous-time systems based on the notion of measurement robust control barrier functions. Our controller uses idea from self-triggered control and enables us to avoid using stochastic calculus. Our framework is agnostic to the choice of the perception map, independent of the noise distribution, and to the best of our knowledge the first to provide probabilistic safety guarantees in such a setting. We demonstrate the effectiveness of our proposed perception-based controller for a LiDAR-enabled F1/10th car.

Bilayer plates are slender structures made of two thin layers of different materials. They react to environmental stimuli and undergo large bending deformations with relatively small actuation. The reduced model is a constrained minimization problem for the second fundamental form, with a given spontaneous curvature that encodes material properties, subject to an isometry constraint. We design a local discontinuous Galerkin (LDG) method which imposes a linearized discrete isometry constraint and controls deformation gradients at barycenters of elements. We prove $Γ$-convergence of LDG, design a fully practical gradient flow, which gives rise to a linear scheme at every step, and show energy stability and control of the isometry defect. We extend the $Γ$-convergence analysis to piecewise quadratic creases. We also illustrate the performance of the LDG method with several insightful simulations of large deformations, one including a curved crease., 41 pages

This thesis utilizes a combined numerical and machine learning approach to explorethe performance of an alumina ceramic tile undergoing high-velocity impact. The fi-nite element model is established by incorporating a user-defined Johnson-Holmquist-Beissel (JHB) material model within the framework of smoothed particle hydrody-namics (SPH) in LS-DYNA finite element software. The computational frameworkis validated across a range of conditions by matching the simulation results fromboth plate impact experiments and ballistic testing from the literature. Once vali-dated, the model is used to generate training data sets for an artificial neural network(ANN) to predict the residual velocity and projectile erosion of an alumina ceramictile undergoing high-velocity impact in the SPH framework. The ANN is then used toperform a sensitivity analysis involving exploring the effect of mechanical properties(e.g., strength and shear modulus) and impact simulation geometries (e.g., the thick-ness of ceramic tile) on material performance (i.e., residual projectile velocity anderosion). Overall, this study shows the capability of the hybrid FEM-ANN approachin studying the high-velocity impact on ceramic tiles and is applicable to guide thestructural-scale design of ceramic-based protection systems.

Supplementary Material 10

Network Intrusion Detection (NID) works as a kernel technology for the security network environment, obtaining extensive research and application. Despite enormous efforts by researchers, NID still faces challenges in deploying on resource-constrained devices. To improve detection accuracy while reducing computational costs and model storage simultaneously, we propose a lightweight intrusion detection approach based on self-knowledge distillation, namely LNet-SKD, which achieves the trade-off between accuracy and efficiency. Specifically, we carefully design the DeepMax block to extract compact representation efficiently and construct the LNet by stacking DeepMax blocks. Furthermore, considering compensating for performance degradation caused by the lightweight network, we adopt batch-wise self-knowledge distillation to provide the regularization of training consistency. Experiments on benchmark datasets demonstrate the effectiveness of our proposed LNet-SKD, which outperforms existing state-of-the-art techniques with fewer parameters and lower computation loads., Comment: Accepted to IEEE ICC 2023

Recently, the birth of non-fungible tokens (NFTs) has attracted great attention. NFTs are capable of representing users' ownership on the blockchain and have experienced tremendous market sales due to their popularity. Unfortunately, the high value of NFTs also makes them a target for attackers. The defects in NFT smart contracts could be exploited by attackers to harm the security and reliability of the NFT ecosystem. Despite the significance of this issue, there is a lack of systematic work that focuses on analyzing NFT smart contracts, which may raise worries about the security of users' NFTs. To address this gap, in this paper, we introduce 5 defects in NFT smart contracts. Each defect is defined and illustrated with a code example highlighting its features and consequences, paired with possible solutions to fix it. Furthermore, we propose a tool named NFTGuard to detect our defined defects based on a symbolic execution framework. Specifically, NFTGuard extracts the information of the state variables from the contract abstract syntax tree (AST), which is critical for identifying variable-loading and storing operations during symbolic execution. Furthermore, NFTGuard recovers source-code-level features from the bytecode to effectively locate defects and report them based on predefined detection patterns. We run NFTGuard on 16,527 real-world smart contracts and perform an evaluation based on the manually labeled results. We find that 1,331 contracts contain at least one of the 5 defects, and the overall precision achieved by our tool is 92.6%., Comment: Accepted by ISSTA 2023

In the real world, data tends to follow long-tailed distributions w.r.t. class or attribution, motivating the challenging Long-Tailed Recognition (LTR) problem. In this paper, we revisit recent LTR methods with promising Vision Transformers (ViT). We figure out that 1) ViT is hard to train with long-tailed data. 2) ViT learns generalized features in an unsupervised manner, like mask generative training, either on long-tailed or balanced datasets. Hence, we propose to adopt unsupervised learning to utilize long-tailed data. Furthermore, we propose the Predictive Distribution Calibration (PDC) as a novel metric for LTR, where the model tends to simply classify inputs into common classes. Our PDC can measure the model calibration of predictive preferences quantitatively. On this basis, we find many LTR approaches alleviate it slightly, despite the accuracy improvement. Extensive experiments on benchmark datasets validate that PDC reflects the model's predictive preference precisely, which is consistent with the visualization., Comment: Accepted by ICASSP 2023

In this work, we study the generalization of decohered average symmetry-protected topological (ASPT) phases to open quantum systems with a combination of subsystem symmetries and global symmetries. In particular, we provide examples of two types of intrinsic average higher-order topological phases with average subsystem symmetries. A classification scheme for these phases based on generalized anomaly cancellation criteria of average symmetry is also discussed., Comment: 11+6 pages, 4+1 figures

Non-Hermitian systems have attracted considerable interest in recent years owing to their unique topological properties that are absent in Hermitian systems. While such properties have been thoroughly characterized in free fermion models, they remain an open question for interacting bosonic systems. In this Letter, we present a precise definition of quantum phases for non-Hermitian systems and propose a new family of phases referred to as composite quantum phases. We demonstrate the existence of these phases in a one-dimensional spin-$1$ system and show their robustness against perturbations through numerical simulations. Furthermore, we investigate the phase diagram of our model, indicating the extensive presence of these new phases in non-Hermitian systems. Our work establishes a new framework for studying and constructing quantum phases in non-Hermitian interacting systems, revealing exciting possibilities beyond the single-particle picture., Comment: 9 pages, 5 figures

Background: Completing the transition from a sitting position to a standing position is a basic skill in people’s daily lives and is crucial for independent living. Lower limb dysfunction will bring many inconveniences into a person’s life and greatly affect their quality of life. Patients with lower limb dysfunction are a specialized group, and nursing problems for this group are becoming increasingly serious. Helping patients with lower limb dysfunction restore their lower limb mobility or assisting them to walk is a social problem necessary to be solved. Objective: : To review the recent sit-to-stand assistive devices based on hip support, classify them systematically and introduce their characteristics, including the mechanisms and the types of patients for which such mechanisms are applicable; to help patients with lower limb dysfunction or doctors (therapists) understand and choose a reasonable sit-to-stand assist device based on hip support. Methods: This paper summarizes literatures and patents about sit-to-stand assistive devices. From the aspects of structural characteristics, drive type and support modes based on the hip and applications situation, the advantages and disadvantages of the typical sit-to-stand assist devices are represented. Results: Current and future development trends on the structural characteristics, drive type and support modes based on the hip and applications situation of sit-to-stand assist devices are discussed to improve the humanization, modularization and applicability of sit-to-stand assist devices. Conclusion: Sit-to-stand assistive devices based on hip support can help patients improve the quality of their life, assist patients carrying out rehabilitation training, and delay the decline of lower limb function. However, the existing sit-to-stand assistive devices based on hip support need further improvement in the aspects of motion mechanism, new technology application and ergonomics design.

This study combines applied mathematics, visual analysis technology, information science with an approach of Scientometrics to systematically analyze the development status, research distribution and future trend of intelligent vehicles research. A total number of 3933 published paper index by SCIE and SSCI from 2000 to 2019 are researched based on Mapping Knowledge Domain (MKD) and Scientometrics approaches. Firstly, this paper analyzes the literature content in the field of intelligent vehicles by including the literature number, literature productive countries, research organization, co-authorship of main research groups and the journals from which the articles are mainly sourced. Then, co-citation analysis is used to obtain five major research directions in the field of intelligent vehicles, which include “system framework”, “internet of vehicles”, “intersection control algorithms”, “influence on traffic flow”, and “policies and barriers”, respectively. The keyword co-occurrence analysis is applied to identify four dominant clusters: “planning and control system”, “autonomous vehicle questionnaire”, “sensor and vision”, and “connected vehicles”. Finally, we divide burst keywords into three phases according to the publication date to show more clearly the change of research focus and direction over time.

myotube guidance in Drosophila

To quantitatively investigate the system stability margin and provide systematic design guidance for the hybrid dual-infeed high-voltage direct current (HVdc) system, this article develops a dynamic multi-input–multi-output (MIMO) transfer function model for the hybrid dual-infeed HVdc system, consisting of a line-commutated converter-based HVdc (LCC-HVdc) link and a modular multilevel converter-based HVdc (MMC-HVdc) link located in close mutual proximity at the receiving side. Then, based on the individual channel analysis and design (ICAD) theory, the equivalent single-input–single-output (SISO) feedback control model is developed without structural information loss, which has the advantage of quantitative evaluation for the system stability margin. Then, the impacts of the ac system strength and control system on the system stability margin are quantitatively evaluated by gain margin (GM) and phase margin (PM) indexes based on the derived SISO model. Moreover, the coupling effect between the ac system strength and the control parameters is also studied. The obtained results from the theoretical analysis and the detailed electromagnetic transient (EMT) simulation in PSCAD/EMTDC validate the accuracy of the derived SISO feedback control model and the correctness of the theoretical findings.

The voltage source converter (VSC) system has stability issue under weak AC grid conditions. In this paper, the multi-input multi-output (MIMO) transfer function model of VSC system in frequency-domain is derived. Then, the individual channel analysis and design (ICAD) approach is employed to develop the equivalent single-input single-output (SISO) feedback control model for stability analysis. On that basis, the stability of VSC system and the coupling interaction intensity among the multiple control loops, including the active power control (APC), reactive power control (RPC) and phase-locked-loop (PLL) are investigated. The results show that larger PLL bandwidth increases the coupling degree of active power control (APC) and reactive power control (RPC) closed loops, resulting in the instability of the system under weak AC grid conditions. Finally, a derivative feedback compensation (DFC) approach is proposed to enhance the stability margin for VSC system connected to weak AC grid. The results based on MATLAB-based theoretical analysis and electromagnetic transient (EMT) simulations show that the proposed DFC approach can effectively suppress the instability of VSC system due to the close coupling of APC and RPC closed loops caused by the large PLL bandwidth, thus effectively improve the stability margin of the system.

Drill string vibration during the drilling must be measured in real-time as it will cause damage to the construction. This paper proposed a self-powered downhole vibration sensor based on the triboelectric nanogenerator. The downhole vibration sensor relies on the drill string vibration to induce triboelectric charge and electrostatic induction of nanomaterials, thereby realizing self-powered vibration measurement. Sensing performance test results show that the measurement range is between 0 to 5 Hz, the measurement error does not exceed 3.5%, and the output voltage amplitude with a range of 2 V to 5.5 V decreases with the increases of vibration frequency. Self-powered performance test results show that the output current can reach a maximum value of about 35 × E−8 A when a 50 ohm resistance is connected in series, the output power can reach a maximum value of about 924.5 × E−12 W when a 20k ohm resistance is connected in series at a vibration frequency of 0.8 Hz, and the output current and power all decrease with the increase of the vibration frequency.

AimThe association between gamma-glutamyl transferase (GGT) and type 2 diabetes mellitus (T2DM) is controversial. In this study, we investigated the association between GGT and the risk of T2DM using real-world data, Mendelian randomization (MR) analysis, and literature mining.MethodsA cross-sectional study enrolled 3,048 participants (>40 years) from a community in Northeastern China was conducted. A generalized additive model was used to examine the relation between GGT and T2DM. A two-sample MR was performed to investigate the causal effect of GGT (61,089 individuals, mostly of European ancestry) on T2DM (29,193 cases and 182,573 controls of European ancestry).ResultsGGT was related to glucose metabolism indicators, such as fasting plasma glucose and glycosylated hemoglobin (P < 0.05). The odds ratios (ORs) [95% confidence interval (95% CI), P] for T2DM across the GGT categories (14–16, 17–20, 21–25, 26–35, ≥36) were 1.14 [(0.88-1.47), P = 0.330], 1.55 [(1.22-1.98), P < 0.001], 1.87 [(1.47-2.28), P < 0.001], 1.97 [(1.55-2.52), P < 0.001], and 2.29 [(1.78-2.94), P < 0.001] versus GGT ≤ 13 category after adjusting for potential confounding factors. A generalized additive model identified a non-linear correlation between GGT and T2DM and indicated that the risk of T2DM almost levelled out when GGT exceeded 34 IU/L. The MR analysis showed that the odds of having T2DM for a one-time increase in genetically determined GGT was 0.998 [(0.995-1.002), P = 0.34].ConclusionsOur analysis of observational study suggested that GGT, its increment, within a certain range, is indicative of the development of T2DM. However, MR analysis provided no evidence that GGT is a linear causal factor of T2DM. Further investigation is required to determine if GGT exerts a non-linear causal effect on T2DM.

Immune response in plants is tightly regulated by the coordination of the cell surface and intracellular receptors. In animals, the membrane attack complex/perforin-like (MACPF) protein superfamily creates oligomeric pore structures on the cell surface during pathogen infection. However, the function and molecular mechanism of MACPF proteins in plant pathogen responses remain largely unclear. In this study, we identified an Arabidopsis MACP2 and investigated the responsiveness of this protein during both bacterial and fungal pathogens. We suggest that MACP2 induces programmed cell death, bacterial pathogen resistance, and necrotrophic fungal pathogen sensitivity by activating the biosynthesis of tryptophan-derived indole glucosinolates and the salicylic acid signaling pathway dependent on the activity of enhanced disease susceptibility 1 (EDS1). Moreover, the response of MACP2 mRNA isoforms upon pathogen attack is differentially regulated by a posttranscriptional mechanism: alternative splicing. In comparison to previously reported MACPFs in Arabidopsis, MACP2 shares a redundant but nonoverlapping role in plant immunity. Thus, our findings provide novel insights and genetic tools for the MACPF family in maintaining SA accumulation in response to pathogens in Arabidopsis.

The ring laser gyro (RLG) light intensity mode-scanning curve is the basis of the frequency stabilization, and also is the basic characterization of the oscillation characteristics for RLG ring laser and is one of the important signs of the gyro performance. By introducing a new parameter which called mode height, other characteristics of the light intensity mode-scanning curve were described. The mode height of the light intensity mode-scanning curve was the maximum variation of the corresponded light intensity under conditions of one 0.632 8 μm cavity length change, which was a characteristic quantity reflecting the sharpness of the light intensity mode-scanning curve. This characteristic quantity was directly related to the gyro frequency stabilization accuracy and the frequency stabilization response time. The higher the mode height, the higher the corresponding frequency stabilization accuracy, and the shorter the frequency stabilization response time. The physical mechanism of the mode height in RLG scanning process was analyzed theoretically, and the main factors affecting the mode height were determined. Through the theoretical analysis and the numerical simulation, it is concluded that: by increasing the curvature radius of spherical mirror, increasing the cavity length, reducing the loss and increasing the gain, the mode height of the RLG light intensity mode-scanning curve can be increased more than 3 times, which has important guiding significance and practical engineering value for improving the frequency stabilization accuracy, shortening the frequency stabilization response time, reducing the nonlinear error of gyro proportional factor, and improving the fast stability of RLG.

Objective: To analyze the role of serum sortilin in coronary artery calcification (CAC) and cardiovascular and cerebrovascular events (CCE) in maintenance hemodialysis (MHD) patients. Methods: One hundred eleven patients with MHD ≥3 months were included in this study. The general data, clinical features, hematological data, and medication history of the patients were recorded. Eighty-five cases were examined by vascular color Doppler ultrasound, cardiac color Doppler ultrasound, lateral lumbar radiography, and coronary artery calcification score. The patients were followed up for a median time of 45 months. The primary endpoint was CCE or death from a vascular event, and the role of sortilin in this process was analyzed. Results: Among 85 MHD patients, 51 cases (60.00%) had different degrees of CAC. There were significant differences in diabetes, dialysis time, serum phosphorus, calcium-phosphorus product, medical history of phosphate binders, sortilin, and carotid artery plaque between 4 different degrees of calcification groups (p < 0.05). Logistic regression analysis showed that diabetes (OR = 5.475; 95% CI: 1.794–16.71, p = 0.003), calcium-phosphorus product (OR = 2.953; 95% CI: 1.198–7.279, p = 0.019), and sortilin (OR = 1.475 per 100 pg/mL; 95% CI: 1.170–1.858, p = 0.001) were independent risk factors for CAC. During the follow-up, 28 cases of 111 patients (25.23%) suffered from CCE. There were significant differences in CCE between mild, moderate, and severe CAC groups and noncalcification groups (p < 0.05). Cox regression analysis showed that diabetes mellitus (HR 3.424; 95% CI: 1.348–8.701, p = 0.010), CAC (HR 5.210; 95% CI: 1.093–24.83, p = 0.038), and serum sortilin (HR = 8.588; 95% CI: 1.919–38.43, p = 0.005) were independent risk factors for CCE. Besides, we proposed a cutoff value of 418 pg/mL for serum sortilin level, which was able to predict the occurrence of CCE with 75.0% sensitivity and 71.9% specificity. The area under the curve was 0.778 (95% CI: 0.673–0.883). Conclusion: Sortilin is newly found to be independently associated with CAC and CCE in MHD patients.

Precipitation is mostly formed via the ice phase in mixed phase clouds, and ice clouds are very relevant for Earths’ climate. Freezing or prevention of freezing is common to everyday life, e.g. for food and drug storage, icing and de-icing, etc. However, the ice nucleation process is not well understood, since it occurs on the size scale of clusters of molecules and time scales of molecular fluctuations. In this study, we have taken a step toward nanoscale observation of particles that nucleate ice by developing a new ice nucleation instrument, referred as the INXcell, which couples an ice nucleation environmental cell to the scanning transmission X-ray microscope (STXM) at the Swiss Light Source. We employ near-edge X-ray absorption fine-structure spectroscopy (NEXAFS) to map in situ chemical composition of ice nucleating particles with 35 × 35 nm2 spatial resolution. The main technical challenge was control of temperature, T, and thus relative humidity, RH, while maintaining X-ray transparency. In the INXcell, X-rays are focused onto a sample through a temperature-controlled aperture, which was modified to host a jet of nitrogen cooled down to 170 K. The cold jet impinges on the back surface of a sample exposed to water vapor to control sample temperature and thus RH. We used our unique spectroscopic and ice nucleation capability and investigated the heterogeneous freezing ability of ferrihydrite particles with and without coatings of citric acid. Ferrihydrite is an amorphous or poorly crystalline iron oxyhydroxide abundant in mineral dust and is difficult to identify with conventional XRD analysis. We confirmed that ferrihydrite could nucleate ice via immersion freezing and deposition ice nucleation, depending on whether or not the particles first take up water, respectively. When coating ferrihydrite with citric acid, mimicking organic coatings that aerosol particles obtain throughout their atmospheric lifetime, we observed a reduction in the efficiency to nucleate ice following freezing point depression. Spectroscopic identification of the coated ferrihydrite structure emplyed the iron and carbon X-ray absorption L-edges and K-edge, respectively. We also investigated feldspar particles coated with xanthan gum, a surrogate for a highly ice active mineral with a highly viscous organic coating. We observed that deposition ice nucleation occurred only below the RH dependent glass transition of xanthan gum. Using a newly developed stochastic freezing model (SFM) based on solution water activity, we reproduced average conditions and data scatter of the RH and T at which ice formed. Additionally, we ran our model with atmospheric idealized air parcel trajectories and found overall that deposition ice nucleation was the dominant heterogeneous freezing mechanism. Homogeneous ice nucleation subsequent to water uptake out-performed immersion freezing.

The catalytic reduction of NO with NH3 (NH3-SCR) on phosphorus-doped carbon aerogels (P-CAs) was studied in the temperature range of 100–200 °C. The P-CAs were prepared by a one-pot sol–gel method by using phosphoric acid as a phosphorus source followed by carbonization at 600–900 °C. A correlation between catalytic activity and surface P content is observed. The P-CA-800vac sample obtained via carbonization at 800 °C and vacuum treatment at 380 °C shows the highest NO conversion of 45.6–76.8% at 100–200 °C under a gas hourly space velocity of 500 h−1 for the inlet gas mixture of 500 ppm NO, 500 ppm NH3 and 5.0 vol% O2. The coexistence of NH3 and O2 is essential for the high conversion of NO on the P-CA carbon catalysts, which can decrease the spillover of NO2 and N2O. The main Bronsted acid sites derived from P-doping and contributed by the C–OH group at edges of carbon sheets are beneficial for NH3 adsorption. In addition, the C3–PO configuration seems to have the most active sites for favorable adsorption and dissociation of O2 and facilitates the formation of NO2. Therefore, the simultaneous presence of acidic groups for NH3 adsorption and the C3–PO active sites for NO2 generation due to the activation of O2 molecules is likely responsible for the significant increase in the NH3-SCR activity over the P–CAs. The transformation of C3–PO to C–O–P functional groups after the reaction is found, which could be assigned to the oxidation of C3–PO by the dissociated O*, resulting in an apparent decrease of catalytic activity for P-CAs. The C–O–P based functional groups are also active in the NH3-SCR reaction.

Supplemental material, sj-docx-1-npx-10.1177_1934578X221123698 for Tanshinone IIA Inhibits Liver Fibrosis by Regulating COL1A1 Expression Through H19/let-7a in Mice by Chao Lin, Jianming Xing, Ziping Jiang, Liqun Sun, Yongjian Gao, Shuo Yang, Dongxu Wang and Ning Yin in Natural Product Communications

The hyperparameter optimization of neural network can be expressed as a bilevel optimization problem. The bilevel optimization is used to automatically update the hyperparameter, and the gradient of the hyperparameter is the approximate gradient based on the best response function. Finding the best response function is very time consuming. In this paper we propose CPMLHO, a new hyperparameter optimization method using cutting plane method and mixed-level objective function.The cutting plane is added to the inner layer to constrain the space of the response function. To obtain more accurate hypergradient,the mixed-level can flexibly adjust the loss function by using the loss of the training set and the verification set. Compared to existing methods, the experimental results show that our method can automatically update the hyperparameters in the training process, and can find more superior hyperparameters with higher accuracy and faster convergence.

Language-driven action localization in videos is a challenging task that involves not only visual-linguistic matching but also action boundary prediction. Recent progress has been achieved through aligning language query to video segments, but estimating precise boundaries is still under-explored. In this paper, we propose entity-aware and motion-aware Transformers that progressively localizes actions in videos by first coarsely locating clips with entity queries and then finely predicting exact boundaries in a shrunken temporal region with motion queries. The entity-aware Transformer incorporates the textual entities into visual representation learning via cross-modal and cross-frame attentions to facilitate attending action-related video clips. The motion-aware Transformer captures fine-grained motion changes at multiple temporal scales via integrating long short-term memory into the self-attention module to further improve the precision of action boundary prediction. Extensive experiments on the Charades-STA and TACoS datasets demonstrate that our method achieves better performance than existing methods., Comment: accepted by IJCAI-22, Codes are available at https://github.com/shuoyang129/EAMAT

With the shift towards on-device deep learning, ensuring a consistent behavior of an AI service across diverse compute platforms becomes tremendously important. Our work tackles the emergent problem of reducing predictive inconsistencies arising as negative flips: test samples that are correctly predicted by a less accurate model, but incorrectly by a more accurate one. We introduce REGression constrained Neural Architecture Search (REG-NAS) to design a family of highly accurate models that engender fewer negative flips. REG-NAS consists of two components: (1) A novel architecture constraint that enables a larger model to contain all the weights of the smaller one thus maximizing weight sharing. This idea stems from our observation that larger weight sharing among networks leads to similar sample-wise predictions and results in fewer negative flips; (2) A novel search reward that incorporates both Top-1 accuracy and negative flips in the architecture search metric. We demonstrate that \regnas can successfully find desirable architectures with few negative flips in three popular architecture search spaces. Compared to the existing state-of-the-art approach, REG-NAS enables 33-48% relative reduction of negative flips., Comment: To be presented at ECCV 2022

The great success of deep learning heavily relies on increasingly larger training data, which comes at a price of huge computational and infrastructural costs. This poses crucial questions that, do all training data contribute to model's performance? How much does each individual training sample or a sub-training-set affect the model's generalization, and how to construct the smallest subset from the entire training data as a proxy training set without significantly sacrificing the model's performance? To answer these, we propose dataset pruning, an optimization-based sample selection method that can (1) examine the influence of removing a particular set of training samples on model's generalization ability with theoretical guarantee, and (2) construct the smallest subset of training data that yields strictly constrained generalization gap. The empirically observed generalization gap of dataset pruning is substantially consistent with our theoretical expectations. Furthermore, the proposed method prunes 40% training examples on the CIFAR-10 dataset, halves the convergence time with only 1.3% test accuracy decrease, which is superior to previous score-based sample selection methods.

Safety-critical Autonomous Systems require trustworthy and transparent decision-making process to be deployable in the real world. The advancement of Machine Learning introduces high performance but largely through black-box algorithms. We focus the discussion of explainability specifically with Autonomous Vehicles (AVs). As a safety-critical system, AVs provide the unique opportunity to utilize cutting-edge Machine Learning techniques while requiring transparency in decision making. Interpretability in every action the AV takes becomes crucial in post-hoc analysis where blame assignment might be necessary. In this paper, we provide positioning on how researchers could consider incorporating explainability and interpretability into design and optimization of separate Autonomous Vehicle modules including Perception, Planning, and Control.

E-commerce has gone a long way in empowering merchants through the internet. In order to store the goods efficiently and arrange the marketing resource properly, it is important for them to make the accurate gross merchandise value (GMV) prediction. However, it's nontrivial to make accurate prediction with the deficiency of digitized data. In this article, we present a solution to better forecast GMV inside Alipay app. Thanks to graph neural networks (GNN) which has great ability to correlate different entities to enrich information, we propose Gaia, a graph neural network (GNN) model with temporal shift aware attention. Gaia leverages the relevant e-seller' sales information and learn neighbor correlation based on temporal dependencies. By testing on Alipay's real dataset and comparing with other baselines, Gaia has shown the best performance. And Gaia is deployed in the simulated online environment, which also achieves great improvement compared with baselines., Comment: Accepted by ICDE 2022

Negative flips are errors introduced in a classification system when a legacy model is replaced with a new one. Existing methods to reduce the negative flip rate (NFR) either do so at the expense of overall accuracy using model distillation, or use ensembles, which multiply inference cost prohibitively. We present a method to train a classification system that achieves paragon performance in both error rate and NFR, at the inference cost of a single model. Our method introduces a generalized distillation objective, Logit Difference Inhibition (LDI), that penalizes changes in the logits between the new and old model, without forcing them to coincide as in ordinary distillation. LDI affords the model flexibility to reduce error rate along with NFR. The method uses a homogeneous ensemble as the reference model for LDI, hence the name Ensemble LDI, or ELODI. The reference model can then be substituted with a single model at inference time. The method leverages the observation that negative flips are typically not close to the decision boundary, but often exhibit large deviations in the distance among their logits, which are reduced by ELODI., Comment: Tech report

Additional file 3: Table S1. Primer sequences for genes in our article.

In this paper, we investigate the problem of synthesizing controllers for temporal logic specifications under security constraint. We assume that there exists a passive intruder (eavesdropper) that can partially observe the behavior of the system. For the purpose of security, we require that the system's behaviors are unpredictable in the sense that the intruder cannot determine for sure that the system will exactly accomplish the task in $K$ steps ahead. This problem is particularly challenging since future information is involved in the synthesis process. We propose a novel information structure that predicts the effect of control in the future. A sound and complete algorithm is developed to synthesize a controller which ensures both task completion and security guarantee. The proposed approach is illustrated by a case study of robot task planning.