Explore the vast range of titles available.

Loess slopes with steep gradients are particularly prone to vertical tension cracks at the crest, resulting from unloading and other factors. These cracks significantly affect the spatiotemporal distribution of moisture infiltration during rainfall, potentially leading to slope instability. This study investigates the impact of crest-tension cracks on moisture infiltration in loess slopes under extreme rainfall conditions, focusing on crack position, depth, and width. Soil moisture content and the dynamics of wetting fronts were monitored to assess how these tension cracks influence infiltration patterns. The results indicate that tension cracks at the slope crest act as preferential infiltration pathways, causing water retention within the cracks and forming a “U-shaped” preferential infiltration zone. The extent of this “U-shaped” wetting front is influenced by the crack’s width, depth, and proximity to the slope shoulder; wider, deeper cracks closer to the shoulder result in a more pronounced wetting front. Over time, as rainfall persists, the influence of preferential infiltration decreases, and the infiltration patterns of slopes with crest cracks begin to resemble those of homogeneous slopes. In both cases, wetting fronts exhibit intersecting patterns: one parallel to the slope crest and the other parallel to the slope surface. During the initial stages of rainfall, the migration speed of wetting fronts in slopes with crest-tension cracks was significantly higher than in homogeneous slopes. However, after prolonged rainfall, the migration speeds of wetting fronts in both scenarios converged. A strong linear correlation was observed between the average migration depth of the horizontal wetting front at the slope crest and the parallel wetting front on the slope surface, for both slope types. These findings deepen our understanding of moisture migration dynamics in loess slopes with crest-tension cracks, providing insights for developing effective slope hazard mitigation strategies.

Plant roots are important in ecological slope protection and reinforcement, significantly affecting soil’s water-holding characteristics and shear strength. The typical herb Festuca Arundinacea root-loess composite in the Loess Plateau was taken as the research object in this paper. The matrix suction test, unsaturated shear strength test, nuclear magnetic resonance (NMR) test, and scanning electron microscope (SEM) test were used to systematically study the root–soil composite’s suction state and strength characteristics under different water content conditions and reveal its internal physical mechanism. The main research results are as follows: (i) The incorporation of roots can increase the air entry value (AEV), reduce the residual water content, and significantly enhance the matrix suction within a specific water content range, thereby increasing the unsaturated shear strength and the enhancement effect of shear strength gradually decreases with the increase of water content. (ii) The microscopic test results show that the root system accelerates the water loss of the soil under low suction conditions by changing the pore structure, increasing the total porosity and the proportion of medium and large pores. At the same time, the increase of tiny pores enhances the capillary action, increasing matrix suction. (iii) Roots can also improve the soil stress state, enhance the adsorption strength of matrix suction, and effectively improve soil shear strength. The results of this study are helpful to understanding the evolution mechanism of unsaturated strength characteristics of root–soil composite and provide the experimental basis and theoretical reference for vegetation slope protection engineering in the loess area.

The Alternating Direction Method of Multipliers (ADMM) is widely used for linearly constrained convex problems. It is proven to have an o ( 1 / K ) nonergodic convergence rate and a faster O (1 /  K ) ergodic rate after ergodic averaging, where K is the number of iterations. Such nonergodic convergence rate is not optimal. Moreover, the ergodic averaging may destroy the sparseness and low-rankness in sparse and low-rank learning. In this paper, we modify the accelerated ADMM proposed in Ouyang et al. (SIAM J. Imaging Sci. 7(3):1588–1623, 2015 ) and give an O (1 /  K ) nonergodic convergence rate analysis, which satisfies | F ( x K ) - F ( x ∗ ) | ≤ O ( 1 / K ) , ‖ A x K - b ‖ ≤ O ( 1 / K ) and x K has a more favorable sparseness and low-rankness than the ergodic peer, where F ( x ) is the objective function and A x = b is the linear constraint. As far as we know, this is the first O (1 /  K ) nonergodic convergent ADMM type method for the general linearly constrained convex problems. Moreover, we show that the lower complexity bound of ADMM type methods for the separable linearly constrained nonsmooth convex problems is O (1 /  K ), which means that our method is optimal.

Optimization over low rank matrices has broad applications in machine learning. For large-scale problems, an attractive heuristic is to factorize the low rank matrix to a product of two much smaller matrices. In this paper, we study the nonconvex problem minU∈Rn×rg(U)=f(UUT) under the assumptions that f(X) is restricted μ-strongly convex and L-smooth on the set X:X⪰0,rank(X)≤r. We propose an accelerated gradient method with alternating constraint that operates directly on the U factors and show that the method has local linear convergence rate with the optimal dependence on the condition number of L/μ. Globally, our method converges to the critical point with zero gradient from any initializer. Our method also applies to the problem with the asymmetric factorization of X=U~V~T and the same convergence result can be obtained. Extensive experimental results verify the advantage of our method.

Medication-related adverse events are common in pregnant women, and most are due to misunderstanding medication information. The identification of appropriate medication information sources requires adequate medical information literacy (MIL). It is important for pregnant women to comprehensively evaluate the risk of medication treatment, self-monitor their medication response, and actively participate in decision-making to reduce medication-related adverse events. This study aims to examine the effectiveness of a medication education course on a web-based platform in improving pregnant women's MIL and decision self-efficacy. A randomized controlled trial was conducted. Pregnant women were recruited from January to June 2021 in the Department of Obstetrics and Gynecology of a large hospital in a major city in central China. A total of 108 participants were randomly divided into a control group (CG), which received routine prenatal care from nurses and physicians, and an intervention group (IG), which received an additional 3-week web-based medication education course based on the theory of planned behavior as part of routine prenatal care. Participants completed a Medication Information Literacy Scale and a decision self-efficacy questionnaire at baseline, upon completion of the intervention, and at a 4-week follow-up. Generalized estimation equations (GEE) were used to analyze the main effect (time and grouping) and interaction effect (grouping×time) of the 2 outcomes. The CONSORT-EHEALTH (V 1.6.1) checklist was used to guide the reporting of this randomized controlled trial. A total of 91 pregnant women (48 in the IG and 43 in the CG) completed the questionnaires at the 3 time points. The results of GEE indicated that there was no statistically significant difference in time×group interactions of MIL between the 2 groups (F =3.12; P=.21). The results of the main effect analysis showed that there were statistically significant differences in MIL between the 2 groups at T1 and T2 (F =17.79; P<.001). Moreover, the results of GEE indicated that there was a significant difference in decision self-efficacy regarding the time factor, grouping factor, and time×group interactions (F =21.98; P<.001). The results of the simple effect analysis indicated a statistically significant difference in decision self-efficacy between the 2 groups at T1 (F =36.29; P<.001) and T2 (F =36.27; P<.001) compared to T0. Results showed that MIL and decision self-efficacy in the IG were found to be significantly higher than those in the CG (d=0.81; P<.001 and d=1.26; P<.001, respectively), and they remained significantly improved at the 4-week follow-up (d=0.59; P<.001 and d=1.27; P<.001, respectively). Web-based medication education courses based on the theory of planned behavior can effectively improve pregnant women's MIL and decision self-efficacy, and they can be used as supplementary education during routine prenatal care. Chinese Clinical Trial Registry ChiCTR2100041817; https://www.chictr.org.cn/showproj.html?proj=66685.

Many problems in machine learning and other fields can be (re)formulated as linearly constrained separable convex programs. In most of the cases, there are multiple blocks of variables. However, the traditional alternating direction method (ADM) and its linearized version (LADM, obtained by linearizing the quadratic penalty term) are for the two-block case and cannot be naively generalized to solve the multi-block case. So there is great demand on extending the ADM based methods for the multi-block case. In this paper, we propose LADM with parallel splitting and adaptive penalty (LADMPSAP) to solve multi-block separable convex programs efficiently. When all the component objective functions have bounded subgradients, we obtain convergence results that are stronger than those of ADM and LADM, e.g., allowing the penalty parameter to be unbounded and proving the sufficient and necessary conditions for global convergence. We further propose a simple optimality measure and reveal the convergence rate of LADMPSAP in an ergodic sense. For programs with extra convex set constraints, with refined parameter estimation we devise a practical version of LADMPSAP for faster convergence. Finally, we generalize LADMPSAP to handle programs with more difficult objective functions by linearizing part of the objective function as well. LADMPSAP is particularly suitable for sparse representation and low-rank recovery problems because its subproblems have closed form solutions and the sparsity and low-rankness of the iterates can be preserved during the iteration. It is also highly parallelizable and hence fits for parallel or distributed computing. Numerical experiments testify to the advantages of LADMPSAP in speed and numerical accuracy.

Adversarial robustness, which primarily comprises sensitivity-based robustness and spatial robustness, plays an integral part in achieving robust generalization. In this paper, we endeavor to design strategies to achieve universal adversarial robustness. To achieve this, we first investigate the relatively less-explored realm of spatial robustness. Then, we integrate the existing spatial robustness methods by incorporating both local and global spatial vulnerability into a unified spatial attack and adversarial training approach. Furthermore, we present a comprehensive relationship between natural accuracy, sensitivity-based robustness, and spatial robustness, supported by strong evidence from the perspective of robust representation. Crucially, to reconcile the interplay between the mutual impacts of various robustness components into one unified framework, we incorporate the Pareto criterion into the adversarial robustness analysis, yielding a novel strategy called Pareto adversarial training for achieving universal robustness. The resulting Pareto front, which delineates the set of optimal solutions, provides an optimal balance between natural accuracy and various adversarial robustness. This sheds light on solutions for achieving universal robustness in the future. To the best of our knowledge, we are the first to consider universal adversarial robustness via multi-objective optimization.

Switchable wetting and optical properties on a surface is synergistically realized by mechanical or temperature stimulus. Unfortunately, in situ controllable wettability together with programmable transparency on 2D/3D surfaces is rarely explored. Herein, Joule‐heat‐responsive paraffin‐impregnated slippery surface (JR‐PISS) is reported by the incorporation of lubricant paraffin, superhydrophobic micropillar‐arrayed elastomeric membrane, and embedded transparent silver nanowire thin‐film heater. Owing to its good flexibility, in situ controllable locomotion for diverse liquids on planar/curved JR‐PISS is unfolded by alternately applying/discharging low electric‐trigger of 6 V. Simultaneously, optical visibility can be reversibly converted between opaque and transparent modes. The switching principle is that in the presence of Joule‐heat, solid paraffin would be melt and swell within 20 s to enable a slippery surface for decreasing light scattering and frictional force derived from contact angle hysteresis (FCAH). Once Joule‐heat is discharged, undulating rough surface would reconfigure by cold‐shrinkage of paraffin within 8 s to render light blockage and high FCAH. Upon its portable merit, in situ thermal management, programmable visibility, as well as steering functionalized droplets by electric‐activated JR‐PISSs are successfully deployed. Compared with previous Nepenthes‐inspired slippery surfaces, the current JR‐PISS is more competent for in situ harnessing optical and wetting properties on‐demand. Switchable optical and wetting properties are highly desirable for up‐to‐date smart surfaces. By embedding a portable, low‐voltage‐driven and transparent silver nanowires heater, the sandwich‐structured bioinspired Joule‐heat‐responsive paraffin‐impregnated slippery surface (JR‐PISS) is competent for in situ tuning these two features in synergy. Insights into the rational design of electric‐induced actuator offer a platform for developing and optimizing next‐generation smart windows.

Tomato, as the vegetable queen, is cultivated worldwide due to its rich nutrient content and unique flavor. Nondestructive technology provides efficient and noninvasive solutions for the quality assessment of tomatoes. However, processing the substantial datasets to achieve a robust model and enhance detection performance for nondestructive technology is a great challenge until deep learning is developed. The aim of this paper is to provide a systematical overview of the principles and application for three categories of nondestructive detection techniques based on mechanical characterization, electromagnetic characterization, as well as electrochemical sensors. Tomato quality assessment is analyzed, and the characteristics of different nondestructive techniques are compared. Various data analysis methods based on deep learning are explored and the applications in tomato assessment using nondestructive techniques with deep learning are also summarized. Limitations and future expectations for the quality assessment of the tomato industry by nondestructive techniques along with deep learning are discussed. The ongoing advancements in optical equipment and deep learning methods lead to a promising outlook for the application in the tomato industry and agricultural engineering.

[This corrects the article DOI: .].