Search

Abstract In generating adversarial examples, the conventional black-box attack methods rely on sufficient feedback from the to-be-attacked models by repeatedly querying until the attack is successful, which usually results in thousands of trials during an attack. This may be unacceptable in real applications since Machine Learning as a Service Platform (MLaaS) usually only returns the final result (i.e., hard-label) to the client and a system equipped with certain defense mechanisms could easily detect malicious queries. By contrast, a feasible way is a hard-label attack that simulates an attacked action being permitted to conduct a limited number of queries. To implement this idea, in this paper, we bypass the dependency on the to-be-attacked model and benefit from the characteristics of the distributions of adversarial examples to reformulate the attack problem in a distribution transform manner and propose a distribution transform-based attack (DTA). DTA builds a statistical mapping from the benign example to its adversarial counterparts by tackling the conditional likelihood under the hard-label black-box settings. In this way, it is no longer necessary to query the target model frequently. A well-trained DTA model can directly and efficiently generate a batch of adversarial examples for a certain input, which can be used to attack un-seen models based on the assumed transferability. Furthermore, we surprisingly find that the well-trained DTA model is not sensitive to the semantic spaces of the training dataset, meaning that the model yields acceptable attack performance on other datasets. Extensive experiments validate the effectiveness of the proposed idea and the superiority of DTA over the state-of-the-art.

Abstract The development of semiconducting polymers with good processability in green solvents and competitive electrical performance is essential for realizing sustainable large‐scale manufacturing and commercialization of organic electronics. A major obstacle is the processability‐performance dichotomy that is dictated by the lack of ideal building blocks with balanced polarity, solubility, electronic structures, and molecular conformation. Herein, through the integration of donor, quinoid and acceptor units, an unprecedented building block, namely TQBT, is introduced for constructing a serial of conjugated polymers. The TQBT, distinct in non‐symmetric structure and high dipole moment, imparts enhanced solubility in anisole—a green solvent—to the polymer TQBT‐T. Furthermore, PTQBT‐T possess a highly rigid and planar backbone owing to the nearly coplanar geometry and quinoidal nature of TQBT, resulting in strong aggregation in solution and localized aggregates in film. Remarkably, PTQBT‐T films spuncast from anisole exhibit a hole mobility of 2.30 cm2 V‐1 s‐1, which is record high for green solvent‐processable semiconducting polymers via spin‐coating, together with commendable operational and storage stability. The hybrid building block emerges as a pioneering electroactive unit, shedding light on future design strategies in high‐performance semiconducting polymers compatible with green processing and marking a significant stride towards ecofriendly organic electronics.

Abstract Background Notopterygii Rhizoma et Radix (NRR) is commonly used for the treatment of inflammation-linked diseases. Phenethylferulate (PF) is high content in NRR crude, but its anti-inflammatory effect remains unclear. Therefore, we aimed to investigate the anti-inflammatory properties of PF and its underlying molecular mechanisms in lipopolysaccharide (LPS)-stimulated RAW 264.7 macrophages. Methods The effect of PF on cell viability was measured by MTT assay. The anti-inflammatory properties of PF were studied by detecting the levels of inflammatory mediators and cytokines using enzyme-linked immunosorbent assay (ELISA). Furthermore, the anti-inflammatory mechanisms of PF were determined by Western blot analysis. Results PF was not cytotoxic to RAW 264.7 macrophages at the concentrations of below 48 μM. ELISA showed that PF conspicuously inhibited overproduction of prostaglandin E2 (PGE2), tumor necrosis factor α (TNF-α), interleukin 1β (IL-1β) and interleukin 6 (IL-6). Western blot analysis showed that PF remarkably suppressed overproduction of inducible nitric oxide synthase (iNOS) and cyclooxygenase 2 (COX-2), the phosphorylation of inhibitor of NF-κB kinase α (IκB-α), protein kinase B (Akt), extracellular signal-regulated kinase (ERK), c-Jun N-terminal kinases (JNK) and p38, as well as the degradation and subsequent nuclear translocation of p65. Conclusions PF is a potent inhibitor of inflammation acting on nuclear factor kappa-B (NF-κB), Akt and mitogen-activated protein kinase (MAPK) signaling pathways in LPS-stimulated RAW 264.7 macrophages. This work provides evidence for the suitability of PF as a therapeutic candidate for the management of inflammatory-mediated immune disorders.

The rapid advancement of mobile devices and human–machine interaction technologies has ushered in a new era for flexible touch panels as a novel input interface. This article reviews the historical evolution and technical progress of flexible touch panel technologies, from rudimentary single‐point touch to sophisticated grid‐free touch systems. Additionally, the working principles and mechanisms that underpin these advanced systems, including capacitive, resistive, piezoelectric, and triboelectric nanogenerator technologies, are explored. Following this, the integration of machine learning methods into these panels is discussed, offering new avenues for enhancing user experience and expanding functional capabilities. Various machine learning algorithms such as support vector machines, artificial neural networks, convolutional neural networks, and k‐nearest neighbors are examined for their potential applications in touch panel technologies. Finally, the challenges and prospects for the application of flexible touch panels fused with machine learning are discussed.

Abstract A simple and effective method was reported for the rapid synthesis of zeolite X from kaolin by adding a template agent. The results of X‐ray diffraction confirm the good crystallinity of the synthesized X‐zeolite with lamellar structure. Transmission electron microscopy images clearly demonstrate the lamellar structure of the synthesized X‐zeolite. The corresponding selective electron diffraction pattern suggests that the crystal structure of X‐type zeolite is octahedral. The co‐existence and uniform distribution of Si, Al, O, and Na in the layered structure of the synthesized X‐type zeolite are detected by element mapping. In addition, N2 adsorption–desorption measurement indicates that the specific surface area of the lamellar zeolite X is 153.4625 m2 g−1. The micropore surface area is 148.8374 m2 g−1, and the micropore volume is 0.073377 m2 g−1. After exploring the influence of different temperatures on the adsorption of carbon dioxide, the research shows that two‐dimensional zeolite X has the best adsorption capacity for carbon dioxide. The adsorption and separation performance of CO2/CH4 was investigated. The study showed that the adsorption and separation coefficient of two‐dimensional zeolite X was the highest.

Purposes Aiming at the problem of low accuracy of single image depth prediction in dusty environment, a dust image depth prediction network based on sparse input features is proposed. Methods First, by using the relationship between the direct transmission rate of dust image and depth information, a depth prediction network is designed to obtain a depth prediction map. With the prior principle of image color attenuation, the sparse depth features of the dust image are further obtained from the estimated depth map.Then, the sparse depth features and dust images are used as the input of the depth prediction network. The deep prediction network uses an “encoder-decoder” model framework, and the two inputs in the encoder are encoded by different networks. Among them, the dust image is encoded by the residual network, which solves the problem of gradient dispersion and accuracy degradation in the deep network. This prediction method not only ensures the accuracy but also controls the speed. The sparse depth feature adopts a sparse convolutional network with a fusion channel attention mechanism. Coding makes effective feature maps weight more and invalid or less effective feature maps less. Conclusions The resulting output is then channel-fused and fed into the decoder. The decoder uses deconvolution and multi-scale upsampling to design an upsampling module, and uses bicubic upsampling to better reconstruct dense depth information. Minimum absolute value loss and structural similarity loss are adopted as edge preserving loss functions. The experimental results on the NYU-Depth-v2 dataset show that the method can effectively predict the depth information of dust images, the average relative error is reduced to 0.054, the root mean square error is reduced to 0.610, and the accuracy of δ

Abstract Background Recurrence after resection is the main factor for poor survival. The relationship between clinicopathological factors and recurrence after curative distal pancreatectomy for PDAC has rarely been reported separately. Methods Patients with PDAC after left‑sided pancreatectomy between May 2015 and August 2021 were retrospectively identified. Results One hundred forty-one patients were included. Recurrence was observed in 97 patients (68.8%), while 44 (31.2%) patients had no recurrence. The median RFS was 8.8 months. The median OS was 24.9 months. Local recurrence was the predominant first detected recurrence site (n = 36, 37.1%), closely followed by liver recurrence (n = 35, 36.1%). Multiple recurrences occurred in 16 (16.5%) patients, peritoneal recurrence in 6 (6.2%) patients, and lung recurrence in 4 (4.1%) patients. High CA19-9 value after surgery, poor differentiation grade, and positive lymph nodes were found to be independently associated with recurrence. The patients receiving adjuvant chemotherapy had a decreased likelihood of recurrence. In the high CA19-9 value cohort, the median PFS and OS of the patients with or without chemotherapy were 8.0 VS. 5.7 months and 15.6 VS. 13.8 months, respectively. In the normal CA19-9 value cohort, there was no significant difference in PFS with or without chemotherapy (11.7 VS. 10.0 months, P = 0.147). However, OS was significantly longer in the patients with chemotherapy (26.4 VS. 13.8 months, P = 0.019). Conclusions Tumor biologic characteristics, such as T stage, tumor differentiation and positive lymph nodes, affecting CA19-9 value after surgery are associated with patterns and timing of recurrence. Adjuvant chemotherapy significantly reduced recurrence and improved survival. Chemotherapy is strongly recommended in patients with high CA199 after surgery.

Abstract Introduction Intracerebral hemorrhage is a high-risk pathological event that is associated with formidable morality rates. Here, our objective was to perform a retrospective study to determine the best timing for drainage using physiological data on patients who received drainage at different timings. Methods In this retrospective study, we reviewed 198 patients with hypertensive cerebral hemorrhage who underwent stereotactic drainage at the conventional timing (surgery within 12 h of admission; control group) and 216 patients who underwent stereotactic drainage at a customized surgical timing (elective group). Follow-ups were performed at 3 and 6 months after surgery. Results The clinical indicators, including prognosis, hematoma clearance, recurrent hemorrhage, intracerebral infection, pulmonary infection, deep venous thrombosis, gastrointestinal hemorrhage, National Institutes of Health Stroke Scale scores, and matrix metallopeptidase 2 and 9 levels, were compared between the control and elective groups. Our data indicated that the elective group had significantly better prognosis compared to the control group (p = 0.021), with a higher rate of hematoma clearance (p = 0.004) and a lower rate of recurrent hemorrhage (p = 0.018). The total occurrence rate of post-surgery complications was also lower for the elective group (p = 0.026). NIHSS scores and serum MMP2/9 levels of the elective group were lower than those of the control group. Conclusions Customized timing of stereotactic drainage may be superior to conventional fixed timing (within 12 h post-hemorrhage) in reducing post-surgery complications and promoting recovery, which supports the potential use of customized timing of stereotactic minimally invasive drainage as a new convention in clinics.

On the edge side of internet of vehicles (IoV), mobile edge computing (MEC) servers with certain computational resources are deployed to provide computational service for vehicles. However, with the generation of a series of vehicle computing intensive and delay sensitive applications such as virtual reality (VR) navigation and vehicle-mounted games, the quality of service (QoS) for vehicle applications cannot be guaranteed with the limited computational resources in MECs. In this article, adjacent vehicles in IoV are converged into clusters based on topological stability. The idle computational resources of vehicles within the cluster are integrated and scheduled uniformly to serve vehicles within the cluster with computation task requirements. This reduces the dynamic changes in idle computational resources caused by rapid vehicle movement, and effectively ensured reliable feedback of computing results after offloading computing tasks. In order to reduce the task completion time, the task is decomposed into multiple subtasks with time dependencies, which can be processed in parallel. A directed acyclic graph (DAG) is used to characterize the task model. To further minimize latency in completing computing tasks, a federated double deep Q network-based computation task offloading (FDTO) strategy is proposed for vehicle clusters. The optimal computation task offloading decision is obtained based on the collaborative training and updating of double deep Q network (DDQN). Unlike traditional machine learning algorithms that require centralized training, federated learning (FL) allows only the training model parameters to be passed between vehicles, which not only protects data privacy but also reduces communication overhead. Simulation results show that the proposed strategy can effectively reduce task completion latency compared to the benchmark strategies.

Recent adversarial attack research reveals the vulnerability of learning-based deep learning models (DNN) against well-designed perturbations. However, most existing attack methods have inherent limitations in image quality as they rely on a relatively loose noise budget, i.e., limit the perturbations by Lp-norm. Resulting that the perturbations generated by these methods can be easily detected by defense mechanisms and are easily perceptible to the human visual system (HVS). To circumvent the former problem, we propose a novel framework, called DualFlow, to craft adversarial examples by disturbing the image's latent representations with spatial transform techniques. In this way, we are able to fool classifiers with human imperceptible adversarial examples and step forward in exploring the existing DNN's fragility. For imperceptibility, we introduce the flow-based model and spatial transform strategy to ensure the calculated adversarial examples are perceptually distinguishable from the original clean images. Extensive experiments on three computer vision benchmark datasets (CIFAR-10, CIFAR-100 and ImageNet) indicate that our method can yield superior attack performance in most situations. Additionally, the visualization results and quantitative performance (in terms of six different metrics) show that the proposed method can generate more imperceptible adversarial examples than the existing imperceptible attack methods.

Abstract Background Intracerebral hemorrhage (ICH), a common cerebrovascular disease, seriously threatens human health and has severe secondary injuries, while existing treatment methods have many limitations. α2δ‐1 is a subunit of voltage‐gated Ca2+ channels (VGCCs) and can act on glutamate receptor N‐methyl‐D‐aspartate receptors (NMDARs) to relieve neuropathic pain. Methods We first performed ICH modeling on WT mice and Cacna2d1 knockout (KO) mice. The expression levels of GluN1 and α2δ‐1 were measured by Western blot and q‐PCR, and the interaction between the two proteins was evaluated by co‐precipitation. The neuronal apoptosis was detected by the TUNEL assay, and the expression levels of inflammatory factors were assessed by ELISA. The nerve functions of mice were evaluated using behavioral experiments including corner turn test and forelimb use asymmetry. Cerebral hematoma was indicated by brain water content and lesion volume. Results ICH up‐regulated the expression levels of α2δ‐1 and GluN1. KO of Cacna2d1 significantly reduced the ICH‐induced apoptosis. The treatment of gabapentin on α2δ‐1 also significantly reduced the occurrence of apoptosis. KO of Cacna2d1 also reduced the ICH‐induced levels of inflammatory factors. Furthermore, neural functions were also significantly improved. Conclusion Cacna2d1 KO alleviates cerebral hematoma in ICH mice, exhibits a significant regulating effect on its secondary injuries such as neuronal apoptosis and inflammation, and restores the nerve functions of ICH mice. Loss of Cacna2d1 can provide useful therapeutic clues for ICH treatment.

Ferroptosis is an iron-dependent cell death process characterized by excessive accumulation of reactive oxygen species and lipid peroxidation. The elucidation of ferroptosis pathways may lead to novel cancer therapies. Current evidence suggests that the mechanism of ferroptosis can be summarized as oxidative stress and antioxidant defense mechanisms. During this process, ferrous ions play a crucial role in cellular oxidation, plasma membrane damage, reactive oxygen species removal imbalance and lipid peroxidation. Although, disregulation of intracellular cations (Fe2+, Ca2+, Zn2+, etc.) and anions (Cl-, etc.) have been widely reported to be involved in ferroptosis, their specific regulatory mechanisms have not been established. To further understand the crosstalk effect between ferrous and other ions in ferroptosis, we reviewed the ferroptosis process from the perspective of ions metabolism. In addition, the role of ferrous and other ions in tumor therapy is briefly summarized.

Aviation components play an important role in national defense and aviation development. Bolt connections are widely used in the assembly of aviation components, due to their simple structure and convenient disassembly. In addition to the impact of elastic interaction, the gap between the tightened parts also makes it very difficult to obtain a uniform bolt load, to achieve the required tightness during the tightening process. However, the impact of elastic interaction can be reduced by selecting the best tightening sequence, and the optimal tightening sequence of aviation components under different gaps can be predicted by constructing a neural network surrogate model. Based on the predicted optimal sequence, the elastic interaction matrix corresponding to the sequence can be obtained. In order to obtain a uniform preload, the initial load of each bolt is calculated according to an elastic interaction matrix. This research has improved the tightness of aviation components and the real-life efficiency of tightening process planning.

Solid rocket motors (SRMs) are widely used as propulsion devices in the aerospace industry. The SRM nozzle and combustion chamber are connected with a plugged-in structure, which makes it difficult to use the existing technology to investigate the internal conditions of the SRM during docking and assembly. The unknown deformation of the O-ring inside the groove caused by different assembly conditions will prevent the engine assembly quality from being accurately predicted. Algorithms such as machine learning can be used to fit mechanical simulation data to create a model that can be used to make predictions during assembly. In this paper, the prediction method uses the sampled parameters as boundary conditions and applies the finite element method (FEM) to calculate the stresses and strains of the O-ring under different assembly conditions. The simulation data are fitted using the gradient-enhanced Kriging (GEK) model, which is more suitable for high-dimensional data than the ordinary Kriging model. A genetic algorithm (GA) and conditional tabular generative adversarial networks (CTGAN) are used to optimize the prediction model and improve its accuracy as new data are incorporated. The proposed method is not only accurate but also efficient, allowing for a significant reduction in assembly time. The use of the surrogate model and FEM makes it possible to predict the stresses and strains of the O-ring in real-time, making the assembly process smoother and more efficient. In conclusion, the proposed method provides a promising solution to the challenges associated with the assembly process of SRM in the aerospace industry.

Aerospace engine is the source of power for the development of national defense and aerospace. Engine bolt tightening is an important part of the aerospace engine production process. At present, most of the bolt tightening methods of aerospace engines are manual tightening or traditional rigid tightening, which lead to the problem of low efficiency and large engine deformation. This paper proposes a process planning method for bolt tightening of aerospace engine based on digital twin, including flexible tightening mechanism design, tightening sequence optimization, the establishment of the digital twin system of the tightening mechanism, etc. The flexible tightening mechanism is suitable for different types of aerospace engines. Based on bolt tightening sequence optimization, the deformation and stress of the cabin section during the tightening process are reduced. The bolt tightening work is completed in the digital twin system, the real-time communication of physical and digital space synchronizes the operations of them. After tightening, the residual pre-tightening force of bolts are detected based on the ultrasonic method. The results show that the proposed method can solve the problems of low efficiency of aerospace engine bolt tightening and the problem of the cabin’s large deformation. Furthermore, it provides technical reference and theoretical basis for the bolt tightening technology of aerospace engine.

Heart failure is a major cause of morbidity and mortality worldwide. LongShengZhi capsule (LSZ), a traditional Chinese medicine, is used for treatment of patients with vascular diseases. Herein we investigated the effect of LSZ treatment on doxorubicin (DOX)-induced heart failure in mice. C57BL/6 mice randomly in 3 groups received following treatment: Control group, mice were fed normal chow; DOX group, mice were intraperitoneally injected DOX to induce heart failure and fed normal chow; and LSZ group, mice were injected DOX and fed normal chow containing LSZ. DOX induced heart failure as evidenced by increased serum creatine kinase, lactic dehydrogenase and α-hydroxybutyrate dehydrogenase, and cardiac fibrosis. However, LSZ treatment substantially inhibited DOX-induced heart failure parameters. Mechanistically, LSZ reduced collagen content and fibrosis by inhibiting expression of collagen type I α1 (COL1α1), COL1α2, α-smooth muscle actin and transforming growth factor β1. In addition, DOX-induced cell apoptosis was inhibited by LSZ, coupled with reduced caspase 3 activity and mRNA expression. LSZ decreased inflammatory cytokine levels. More importantly, LSZ decreased oxidative stress by inducing expression of anti-oxidative stress enzymes including superoxide dismutase 1 (SOD1), SOD2, catalase and glutathione peroxidase 1 through activation of forkhead box O3A and sirtuin 3. In conclusion, our study demonstrates that LSZ reduces heart failure by reducing production of reactive oxygen species and inhibiting inflammation/apoptosis. Our study also suggests the potential application of LSZ for heart failure treatment.

Glioma stem cell (GSC)–derived extracellular vesicles (EVs) can mediate the communication between GSCs and microglia. Metastasis-associated lung adenocarcinoma transcript 1 (MALAT1) expression in GSCs, EVs, and supernatant was detected by real-time PCR. The direct targeting between MALAT1 and miR-129-5p, miR-129-5p, and HMGB1 were tested with luciferase reporter analysis. The expression and secretion of interleukin (IL)-6, IL-8, and tumor necrosis factor (TNF)-α were determined in lipopolysaccharide-stimulated microglia or miR-129-5p inhibitor transferred to microglia exposed to GSC EVs or EVs derived from siMALAT1 pre-transferred GSCs. MALAT1 was enriched in GSC EVs compared with GSCs, and up-regulated MALAT1 was also observed in microglia upon GSC EVs incubation. The relative expression and secretion of IL-6, IL-8, and TNF-α in lipopolysaccharide-stimulated microglia were up-regulated in the GSC supernatant group, which could be reversed by dimethyl amiloride (DMA) (EV secretion inhibitor) co-administration or si-MALAT1 pre-transfection of GSCs. Luciferase reporter assay testified the direct binding of MALAT1 and miR-129-5p, miR-129-5p, and HMGB1, and si-MALAT1 could up-regulate miR-129-5p expression and down-regulate HMGB1 expression in microglia cells. The concentration of IL-6, IL-8, and TNF-α in lipopolysaccharide-stimulated microglia exposed to EVs from siMALAT1 transfected GSCs could be up-regulated by miR-129-5p inhibition. EVs lncRNA MALAT1 released from GSCs could modulate the inflammatory response of microglia after lipopolysaccharide stimulation through regulating the miR-129-5p/HMGB1 axis.