Search

With the rapid deployment and wide use of mobile services and applications, more and more sensitive user information is being transmitted wirelessly. Due to the broadcast nature of wireless transmissions, they are exposed to all surrounding entities and thus vulnerable to eavesdropping. To counter this vulnerability, we propose a new physical-layer secure transmission scheme, called DDM-Sec, based on decomposed and distributed modulation (DDM). We show that a high-order modulation can be decomposed into multiple quadrature phase shift keying (QPSK) modulations, each of which can be further represented by two mutually orthogonal binary phase shift keying (BPSK) modulations. Therefore, traditional modulation can be realized by two cooperative transmitters (Txs), each generating a BPSK signal, in a distributed manner. The legitimate receiver (Rx) can decode the desired/intended information from the mixed two received BPSK signals while preventing the eavesdropper from accessing the legitimate user’s information. DDM-Sec can effectively exploit the randomness of wireless channels to secure data transmission, enrich the spatial signatures of the legitimate user’s transmission by employing two cooperative Txs, and then distribute the user’s information to two transmissions so that none of the decomposed signals alone carry the legitimate user’s full information. Moreover, due to random deployment of the two Txs and Rx, delay difference of the two transmissions is introduced. This can be further utilized to make eavesdropping difficult. Our theoretical analysis and simulation have shown that DDM-Sec can effectively prevent the eavesdropping, and hence guarantee the secrecy of the legitimate user’s data transmission.

The second harmonic current (SHC) caused by the input power fluctuation in the single-phase rectifier can ruin the system performance. In this letter, a charge control with nonlinear feedforward is proposed to suppress SHC for a dual-stage single-phase rectifier with an LLC converter. The conventional charge control modulates the switching frequency by regulating the input charge within a single switching cycle to enhance dynamic performance. In the proposed control, the ideal input charge with the input power fluctuation is predicted online and feedforward to the linear compensator, which further enhances the reference tracking performance of the LLC converter. Different from other SHC reduction methods, the proposed control strategy is immune to the line frequency deviation and enjoys a fast dynamic response. Finally, experimental results validate the theoretical analysis.

•What is the primary question addressed by this study?Would Medicare costs be reduced if Black and Hispanic ADRD patients and ADRD patients living in areas with higher social vulnerability enrolled in ACO?•What is the main finding of this study?Significant variations in Medicare costs were observed by patient's race, ethnicity, and social vulnerability of the residing area, with Black and Hispanic patients and those living in the most vulnerable areas experiencing the highest total costs. For ADRD patients residing in the areas with the highest vulnerability, the cost differences by ACO enrollment of the total Medicare costs ranged from $4,403.1 to $6,922.7 and beneficiaries’ savings ranged from $114.5 to $726.6 for a 3-year post-ADRD diagnosis by patient's race and ethnicity.•What is the meaning of the finding?ACOs and emerging CMS Innovation models, such as the Guiding an Improved Dementia Experience (GUIDE) Model, should prioritize strategic resource allocation in socially vulnerable areas while emphasizing culturally competent ADRD care.

The optical quantum control has been successfully applied in modulating biological processes such as energy transfer and bond isomerization. Among the reactions in realizing biological functions, the electron transfer (ET) process is fundamental; hence, the quantum control over such an ET reaction is of far-reaching significance. Here, we realized optical quantum control over ultrafast ET processes in a protein, flavodoxin, by applying various chirped excitation pulses. We observed the wavepacket dynamics within a dephasing time of less than 1 ps. Within this time window, we found that the ultrafast photoinduced ET reaction can be controlled by different chirped excitations with a rate change by a factor of about 2. Furthermore, the control effect is propagated into the subsequent ultrafast back ET reaction, showing a variation of the BET dynamics with different excitation chirps. The underlying mechanism is the initial wavepacket dynamics; the differently prepared wavepackets with chirped excitation evolve along various pathways, resulting in the changes of ET rates. The successful demonstration of optical quantum control of ultrafast biological ET is significant and opens a new avenue to explore the quantum control of real biological ET reactions.

In the field of food processing, enzymes play a pivotal role in improving product quality and flavor, and extending shelf life. However, the exposure of traditional food enzymes to high temperatures during processing often leads to a decrease in activity or even inactivation, limiting the effectiveness of their application under high-temperature conditions. Therefore, the modification of thermostability and activity of enzymes to adapt to extreme conditions through protein engineering has become a key way to improve the efficiency and economic benefits of industrial production. Directed evolution and semirational design strategies in the laboratory have proven to be broadly applicable frameworks for biochemical researchers in the food field, including those who are beginners. In this review, we systematically summarize semirational design strategies and high-throughput screening strategies, and introduce some intuitive computer simulation software to improve the thermostability and enzyme activity of food enzymes. The application of these strategies and techniques provides a comprehensive guide for the optimization of food enzymes. In addition, the latest hot topics of genetically engineered food enzymes in the field of application are discussed.

ABSTRACTBackgroundTralokinumab is a fully human IgG4 monoclonal antibody targeting IL-13, used for treating atopic dermatitis. This study analyzed tralokinumab-related adverse drug events by mining the Food and Drug Administration Adverse Event Reporting System (FAERS) database to provide a safety reference for clinical application.MethodsAdverse drug event reports from Q1 2022 to Q2 2024 were extracted from the FAERS database. After standardizing the data, various signal detection methods were used for analysis, including ROR, PRR, BCPNN, and MGPS.ResultsA total of 1,820 reports of adverse events (AEs) with tralokinumab as the primary suspected drug were identified. 70 preferred terms (PTs) met the criteria across four signal detection methods, involving 11 system organ classes (SOCs). These included known adverse reactions like conjunctivitis and injection site reactions, and signals not previously reported in clinical trials, such as eye pruritus, dry eye, eye swelling, pneumonia pneumococcal, and cutaneous T-cell lymphoma. Most AEs occurred within one month of initiating tralokinumab treatment.ConclusionsBased on the FAERS database, this study comprehensively and systematically analyzed AE signals in tralokinumab treatment. The results enhance the understanding of tralokinumab’s safety and serve as valuable references for reducing the risk of adverse reactions during clinical use.

A typical heterogeneous structure with ultrafine grains embedded in coarse grains is obtained by partial recrystallization annealing, which exhibits excellent cryogenic mechanical properties. Researches reveal that the hard domains, high-density GND pile-ups, and multi-axial stress states of heterogeneous structures provide nucleation positions and critical stresses for triggering twinning, meanwhile, nanotwins can refine the hard domain grains to improve the heterogeneity of heterogeneous structures during cryogenic plastic deformation. Therefore, the heterostructure-twinning interaction is identified as the key factor in realizing the excellent cryogenic strength-ductility synergy of the heterostructured CoCrFeMnNi high-entropy alloy. Besides, the nucleation and propagation of generous microcracks at single-order twin boundaries induced by the excellent twinning activity result in the rapid fracture of the heterostructured samples, whereas the microcracks can hardly be observed nearby high-order nanotwins, which indicates that generating more high-order nanotwins is a promising guiding scheme to prepare the heterostructured materials with superior cryogenic mechanical properties.

This study focuses on the motion planning and risk evaluation of unprotected left turns at occluded intersections for autonomous vehicles. In this paper, we present an interactive motion planning controller that combines Cross-Observability Optimistic-Pessimistic Safe Reinforcement Learning (COOP-SRL) and Nonlinear Model Predictive Control (NMPC), with consideration of the uncertain potential risk of occluded zone, the trade-off between safety and efficiency, and the dynamic interaction between vehicles. The proposed COOP-SRL algorithm integrates fully and partially observable policies through cross-observability soft imitation learning to leverage the expert guidance and improve learning efficiency. Moreover, the optimistic exploration policy and pessimism safe constraint are adopted to provide an adaptive safe strategy without hindering the exploration during learning process. Finally, the evaluations of the proposed controller were conducted in occluded intersection scenarios with various traffic density level, which indicate that the proposed method outperforms both the optimization-based and learning-based baselines in qualitative and quantitative indexes.

Crack detection has attracted extensive attention in an intelligent transportation system (ITS). Despite the substantial progress of deep learning technology on crack recognition tasks, due to the various limitations in traffic, equipment, and time, it is hard to collect copious samples for training deep models. Considering this, implicitly semantic data augmentation (ISDA) tries to augment the training set in the feature space. However, when applying it to crack recognition tasks, our empirical studies reveal that those poor-classified augmented samples have little semantic relevance to the crack class, resulting in a non-negligible negative effect on training deep models. Since the augmented features follow the multivariate normal distribution, it is computationally inefficient to explicitly sample those features and filter out the hard-classified augmented features. To this end, we propose the implicit gradient-modulated semantic data augmentation (IGMSDA) for addressing the above problems. Concretely, this paper first proposes gradient-modulated (GM) loss to dynamically modulate the gradient of those poor-classified augmented samples by reshaping the standard cross-entropy loss. And then, in the feature space, we derive an upper bound of the expected GM loss on the augmented training set to avoid the costly explicit sampling process. Experiments show that IGMSDA improves the generalization performance of the existing deep models on crack recognition datasets.

The fusion localization algorithm represents a cornerstone of data fusion localization technology. However, traditional algorithms often suffer from target loss in complex environments. We propose a heuristic adaptive particle set-based wireless sensor fusion localization method to address the challenges of inaccurate and inefficient multisensor fusion localization in complex environments. Initially, to enhance fusion localization efficiency, the sampling particle set size is dynamically estimated based on the variance of sensor observational distances. This allows for updating the number of sampling particles at each instance, alongside establishing a particle set threshold. Subsequently, to prevent the reduction of particle diversity and the occurrence of local optima or algorithmic divergence, the rate of change in particle number per unit time serves as a criterion for the update of superfluous particles. Then, the particle swarm optimization (PSO) algorithm is employed to augment particle diversity under this condition. The particle weights corresponding to positional information are computed, filtered, fused, and used to update the posterior probability distribution. A simulation experimental platform was subsequently constructed utilizing the RRESCAN environment and MATLAB/SIMULINK, with the experimental model of the algorithm being established therein. Conclusively, a comparative experiment involving ultra-wideband (UWB) and inertial measurement unit (IMU) fusion positioning was conducted in a real-world setting. Our proposed method demonstrates a significant improvement in tracking efficiency and positioning accuracy for vehicle fusion positioning within complex environments through comparative analysis with conventional fusion positioning algorithms.

This article focuses on the coalition formation (CF) problem in urgent missions, e.g., disaster rescue, where coalition members should reach mission locations quickly. A mathematical model is first constructed to minimize the latest arrival time of coalition members, considering the capability requirements of missions, nonredundant agents in coalitions, etc. Then, incorporating the benefits in both the diversity of random search and the effectiveness of utilizing problem knowledge, a local-search-based heuristic is put forward to solve the CF problem. An initial solution is incrementally constructed by prioritizing agents with shorter movement times for missions with higher-remaining capability requirements. Additionally, two types of neighborhood search operators, namely, the tabu-based one-to-one swap and the destroy and repair operators, are proposed to search the solution space from two perspectives, i.e., “adjustment” and “reconstruction.” To solve the problem effectively and efficiently, the former excludes certain agent-exchange combinations that do not improve the current solution, while the latter consists of multiple heuristic rules extracted from the correlation among different model elements. Experimental results have demonstrated that the proposed method surpasses several advanced methods across various scenarios regarding multiple factors, such as the number of agents, the number of missions, and the demand-supply ratio on capabilities.

Thermal plasma vitrification of municipal solid waste incineration fly ash (MSWI FA) shows potential for effective dioxin degradation and heavy metal immobilization. However, the environmental risks posed by heavy metals released after vitrification remain understudied. This work investigates two types of MSWI FA—circulated fluidized bed fly ash (CFA) and mechanical grate furnace fly ash (GFA)—blended with waste glass and vitrified via thermal plasma treatment. The physical and chemical properties of the original FA and vitrified glasses (CFA-glass and GFA-glass) were analyzed. Environmental risks were assessed using the risk assessment code (RAC) and the overall pollution toxicity index (OPTI). Results showed that CFA, with higher Si and Al content, exhibited better vitrification than GFA. Plasma vitrification reduced chlorine content from 5.79 % and 10.59 % to 0.22 % and 0.16% due to Cl volatilization. Heavy metals with low boiling points, such as Cd and Pb, were significantly reduced, while high-boiling-point Cr, Ni remained and the acid resistance of vitrificated FA significantly deteriorated, which increased the potential leaching of heavy metals. In alkaline conditions, the OPTI value of CFA-glass dropped to 2, indicating a strong immobilization effect. This study provides critical insights for managing heavy metals in fly ash through thermal plasma vitrification.

Neuron tracing, alternately referred to as neuron reconstruction, is the procedure for extracting the digital representation of the three-dimensional neuronal morphology from stacks of microscopic images. Achieving accurate neuron tracing is critical for profiling the neuroanatomical structure at single-cell level and analyzing the neuronal circuits and projections at whole-brain scale. However, the process often demands substantial human involvement and represents a nontrivial task. Conventional solutions towards neuron tracing often contend with challenges such as non-intuitive user interactions, suboptimal data generation throughput, and ambiguous visualization. In this paper, we introduce a novel method that leverages the power of extended reality (XR) for intuitive and progressive semi-automatic neuron tracing in real time. In our method, we have defined a set of interactors for controllable and efficient interactions for neuron tracing in an immersive environment. We have also developed a GPU-accelerated automatic tracing algorithm that can generate updated neuron reconstruction in real time. In addition, we have built a visualizer for fast and improved visual experience, particularly when working with both volumetric images and 3D objects. Our method has been successfully implemented with one virtual reality (VR) headset and one augmented reality (AR) headset with satisfying results achieved. We also conducted two user studies and proved the effectiveness of the interactors and the efficiency of our method in comparison with other approaches for neuron tracing.

Multi-view subspace clustering employs learned self-representation from multiple tensor decompositions to exploit the low-rank information. However, the data structures embedded with self-representation tensors may vary in different multi-view datasets. Therefore, a pre-defined decomposition may not fully exploit low-rank information from various data, resulting in sub-optimal multi-view clustering performance. To alleviate this, we proposed the adaptively topological tensor network (ATTN). ATTN can learn a suitable decomposition structure that can represent the low-rank structure and high-order correlation of the self-representation tensors better in a data-driven way, which can capture the intra-view and inter-view information better. Firstly, instead of connecting the tensor network blindly, ATTN utilizes the correlation between adjacent factors to prune redundant connections from the fully connected tensor network, making the tensor network more expressive. Furthermore, a greedy adaptive rank-increasing strategy is applied to optimize the pruned tensor network structure, which improves the capacity of capturing low-rank structure. We apply ATTN on a multi-view subspace clustering task and utilize the alternating direction method of multipliers(ADMM) method to optimize it. Experiments show that multi-view subspace clustering based on ATTN has better performance on nine multi-view datasets.

Fast and safe start-up is crucial for the dual-active-bridge (DAB) converters. Conventional closed-loop start-up controls fail to achieve theoretical maximum power output due to manually predefined phase-shift modes, limiting start-up speed and safe operational range. To address these issues, this letter proposes a novel variable frequency plus extended phase-shift (EPS) control. This approach enables ultrafast and wide-safe-range start-up of DAB converters by maximizing power-per-current-stress capability within safety limitations. In addition, sole EPS modulation is required in the proposed control, thus eliminating complex mode transitions and streamlining the start-up process. Experimental results indicate that the proposed control achieves a broader safe start-up range and improves start-up speed by up to 43% under voltage-match condition compared to the state-of-the-art methods.

Dual active bridge converters are extensively adopted in more electric aircrafts, wherein maintaining a sufficient hold-up time is crucial for air safety. As the current stress in conventional phase-shift controls couples with voltage gain, high-degree voltage gain occurred in the hold-up process would lead to terrible current stress higher than safe limit, which not only heightens the risk of device failure but also shortens the hold-up time due to the potential execution of protective measures. To address such issues, this letter proposes a novel hybrid phase-frequency (HPF) control method, particularly suitable for operations under low input voltage with high inductor currents. The HPF control facilitates the decoupling of the current stress from operation gain, offering advantages such as further optimized and fully regulated current stress, an extended hold-up time and enhanced operational safety. Experimental results show that current stress at full-load can be reduced 56% with HPF control, thereby extending hold-up time by 3 times. Additionally, the reduced current stress under certain power ranges contributes to improved steady-state efficiency

The inherent power flows coupling caused by the integrated transformer among multiple ports is challenging for triple active bridge (TAB) converter. To address this issue, a simple but effective power decoupling control for TAB converters is proposed in this article, which universally involves the influence of internal phase shift and can be seamlessly imbedded into other TAB converter optimal control schemes. The TAB converter is modeled in polar coordinates of frequency domain, and the concept of zero-power vector is introduced. By feedforwarding the phase of zero-power vector to the corresponding port, power flows decoupling can be achieved. To further simplify the algorithm, Taylor expansion is performed, and the proposed algorithm is elegantly simplified into a basic arithmetic expression. An 1.5 kW + 0.5 kW prototype is built up for verification, and the proposed control is compared with the state-of-the-art methods quantitatively. The prototype achieved a rate efficiency of 95% with the proposed control. The results show that the proposed control can reduce the computational load by 20 times compared with the conventional power decoupling control and improves the overall suppression performance by 10 dB.

Effective implementation of carbon capture and storage (CCS) hinges on the proper matching of the sources and sinks. In this study, we have investigated a mixed integer linear programming model for source-sink matching optimization considering hydrothermal properties of CO2pipeline transportation. To incorporate the hydrothermal conditions into the optimization model, pressure and temperature control coefficients (KPand KT) are proposed and coupled. Subsequently, various operating conditions for transporting CO2by pipeline are discussed with different combinations of pressure and temperature control coefficients, exploring the influence on the optimal source-sink matching. It is evident from the conclusions that the CO2pipeline transportation constraints will inevitably affect optimal matching. However, considering CO2dense phase transportation at the back-end of the supercritical CO2pipeline transportation allows the transport temperature to drop below the critical temperature, which increases the number of effective matches, consequently increasing the amount of CO2injected. For instance, when KP= 1.1 and KT= −0.15, the CO2injection amount in the base planning region can be increased by 3.4%. It is important to implement a source-sink matching optimization analysis in combination with CO2pipeline transportation conditions.

We report a novel halogen-bond-assisted NHC-catalyzed (dynamic) kinetic resolution strategy for the synthesis of axially chiral heterobiaryls. A class of axially chiral quinolines are prepared efficiently in excellent enantioselectivities (≤98% ee) employing 3–5 mol % NHC catalyst. Mechanistic studies reveal the indispensability of 5-bromo-2-iodobenzaldehyde in this reaction, in which a pivotal halogen bonding interaction plays a crucial role in the process.

Copper (Cu)-based nanocatalysts play crucial roles in the electrochemical CO2reduction reaction (ECO2RR) for sustainable energy resources. Particularly, Cu-based nanostructures incorporating Au and Ag are promising, offering enhanced activity, selectivity, and stability. However, precise control over the structure and composition of heterostructures remains challenging, hindering the development of highly efficient catalysts. Herein, we present a silver (Ag) transition-layer-mediated approach to synthesize ternary heterostructures with two specific morphologies, namely, Au/Ag–Cu-side and Au/Ag–Cu-tip, which exhibit different Ag–Cu interface epitaxial patterns. The two heterostructures achieve high C2 product selectivity in ECO2RR. Especially, the Au/Ag–Cu-side structure achieves 50.3% C2 selectivity with 35.5% ethanol, while the tip structure shows higher ethylene selectivity. Our study reveals the impact of the Ag layer in directing deposition sites on heterostructure growth and further facilitating the design of multicomponent Cu-based catalysts with enhanced structural integrity and ECO2RR performance.

In recent years, the prediction of antimicrobial peptides (AMPs) has gained prominence due to their high antibacterial activity and reduced susceptibility to drug resistance, making them potential antibiotic substitutes. To advance the field of AMP recognition, an increasing number of natural language processing methods are being applied. These methods exhibit diversity in terms of pretraining models, pretraining data sets, word vector embeddings, feature encoding methods, and downstream classification models. Here, we provide a comprehensive survey of current BERT-based methods for AMP prediction. An independent benchmark test data set is constructed to evaluate the predictive capabilities of the surveyed tools. Furthermore, we compared the predictive performance of these computational methods based on six different AMP public databases. LM_pred (BFD) outperformed all other surveyed tools due to abundant pretraining data set and the unique vector embedding approach. To avoid the impact of varying training data sets used by different methods on prediction performance, we performed the 5-fold cross-validation experiments using the same data set, involving retraining. Additionally, to explore the applicability and generalization ability of the models, we constructed a short peptide data set and an external data set to test the retrained models. Although these prediction methods based on BERT can achieve good prediction performance, there is still room for improvement in recognition accuracy. With the continuous enhancement of protein language model, we proposed an AMP prediction method based on the ESM-2 pretrained model called iAMP-bert. Experimental results demonstrate that iAMP-bert outperforms other approaches. iAMP-bert is freely accessible to the public at http://iamp.aielab.cc/.

In low-illumination environments, visible light images significantly degrade, and the simultaneous localization and mapping (SLAM) system based on visible cameras fails. In contrast, thermal sensors are not affected by scene lighting conditions and have become a promising alternative. However, due to the problems of low contrast, high noise, and data interruption in thermal images, there are still many challenges in implementing thermal sensor SLAM. Therefore, this article proposes an edge-aided monocular thermal sensor SLAM system, called EAMT-SLAM, suitable for low illumination environments. First, the edge provides more feature associations to assist the SLAM system in initializing quickly and accurately. Subsequently, the saliency map is calculated based on the strong edge in the thermal image, and the weights are reasonably assigned to the feature points in the optimization process of pose estimation to obtain more accurate results. Finally, this article proposes an edge-guided loop detection method, which achieves higher accuracy and real-time performance compared with the appearance-based method, effectively ensuring the global consistency of SLAM. In addition, this article proposes a new thermal image processing method that can enhance image details and remove complex noise in real time, significantly improve the quality of thermal images, and broaden the application scope of the EAMT-SLAM system. The experimental results show that our method can obtain the highest quality thermal image under quantitative and qualitative evaluation, complete accurate loop detection, and finally realize accurate pose estimation in low-illumination environments.

Owing to the excellent mobility character, unmanned aerial vehicles (UAVs) show great potentials as a means of data collector in a wireless sensor network (WSN). However, limitations in communication resources require UAVs to collect data from large-scale WSNs with high spectral efficiency. In this work, we investigate a UAV-assisted data collection strategy for a WSN enabled by hybrid non-orthogonal multiple access (NOMA). Specifically, the WSN is organized into clusters and each cluster is further subdivided into multiple groups. The sensor nodes in each group transmit data to the UAV using NOMA scheme. To ensure fairness between sensor nodes, we aim to maximize the minimum throughput by jointly optimizing the sensor node pairing, decoding order, transmit power and the UAV waypoints, subject to transmit power and UAV trajectory length constraints. To address the formulated mixed-integer nonlinear programming problem, we first propose an optimization-based algorithm by applying block coordinate descent method where a closed-form solution of the transmit power is derived, and then design a low-complexity heuristic algorithm. In addition, we present an initialization scheme for the proposed algorithms and an implementation strategy for our system. Extensive simulations demonstrate that our proposed UAV-assisted data collection scheme with hybrid NOMA can enhance system performance efficiently.

Due to the homophily assumption in graph convolution networks (GCNs), a common consensus in the graph node classification task is that graph neural networks (GNNs) perform well on homophilic graphs but may fail on heterophilic graphs with many interclass edges. However, the previous interclass edges’ perspective and related homo-ratio metrics cannot well explain the GNNs’ performance under some heterophilic datasets, which implies that not all the interclass edges are harmful to GNNs. In this work, we propose a new metric based on the von Neumann entropy to reexamine the heterophily problem of GNNs and investigate the feature aggregation of interclass edges from an entire neighbor identifiable perspective. Moreover, we propose a simple yet effective Conv-Agnostic GNN framework (CAGNNs) to enhance the performance of most GNNs on the heterophily datasets by learning the neighbor effect for each node. Specifically, we first decouple the feature of each node into the discriminative feature for downstream tasks and the aggregation feature for graph convolution (GC). Then, we propose a shared mixer module to adaptively evaluate the neighbor effect of each node to incorporate the neighbor information. The proposed framework can be regarded as a plug-in component and is compatible with most GNNs. The experimental results over nine well-known benchmark datasets indicate that our framework can significantly improve performance, especially for the heterophily graphs. The average performance gain is 9.81%, 25.81%, and 20.61% compared with graph isomorphism network (GIN), graph attention network (GAT), and GCN, respectively. Extensive ablation studies and robustness analysis further verify the effectiveness, robustness, and interpretability of our framework. Code is available at https://github.com/JC-202/CAGNN.

Alzheimer’s disease (AD) is one of the most common forms of neurodegenerative dementia. The etiology of AD is multifactorial, and its complex pathophysiology involves tau and amyloid-β deposition, increased oxidative stress, neuroinflammation, metabolic disorders, and massive neuronal loss. Due to its complex pathology, no effective cure for AD has been found to date. Therefore, there is an unmet clinical need for the development of new drugs against AD. Natural products are known to be good sources of compounds with pharmacological activity and have potential for the development of new therapeutic agents. Naringin, a naturally occurring flavanone glycoside, is predominantly found in citrus fruits and Chinese medicinal herbs. Mounting evidence shows that naringin and its aglycone, naringenin, have direct neuroprotective effects on AD, such as anti-amyloidogenic, antioxidant, anti-acetylcholinesterase, and anti-neuroinflammatory effects, as well as metal chelation. Furthermore, they are known to improve disordered glucose/lipid metabolism, which is a high risk factor for AD. In this review, we summarize the latest data on the impact of naringin and naringenin on the molecular mechanisms involved in AD pathophysiology. Additionally, we provide an overview of the current clinical applications of naringin and naringenin. The novel delivery systems for naringin and naringenin, which can address their widespread pharmacokinetic limitations, are also discussed. The literature indicates that naringin and naringenin could be multilevel, multitargeted, and multifaceted for preventing and treating AD.

As the next-generation display technology, the mini-LED (MLED) emissive display adopts a current type thin-film transistor (TFT) device to drive light emission, possessing advantages, such as high contrast ratio and high brightness. Among them, package on glass (POG) is to bond the encapsulated MLED to the TFT glass substrates. Compared to printed circuit board (PCB) substrate, the usage of rigid glass substrates allows for smaller linewidth and line space to meet the needs of high pixel per inch (PPI) products while maintaining product lightness and thinness. In this study, we used parylene C and barrier film packaging to explore different processing routes. By comparing different packaging methods, it was demonstrated that the P1.2 (lamp bead spacing 1.2 mm) MLED product could pass the 500-h reliability test at 60 °C and 90% relative humidity (RH) using parylene. This advanced packaging solution is of great significance for improving the reliability and panel lifetime of MLED emissive displays.

In this article, we consider a distributed learning problem in a subnetwork zero-sum game, where agents are competing in different subnetworks. These agents are connected through time-varying graphs where each agent has its own cost function and can receive information from its neighbors. We propose a distributed mirror descent algorithm for computing a Nash equilibrium and establish a sublinear regret bound on the sequence of iterates when the graphs are uniformly strongly connected and the cost functions are convex–concave. Moreover, we prove its convergence with suitably selected diminishing step sizes for a strictly convex–concave cost function. We also consider a constant step-size variant of the algorithm and establish an asymptotic error bound between the cost function values of running average actions and a Nash equilibrium. In addition, we apply the algorithm to compute a mixed-strategy Nash equilibrium in subnetwork zero-sum finite-strategy games, which have merely convex–concave (to be specific, multilinear) cost functions, and obtain a final-iteration convergence result and an ergodic convergence result, respectively, under different assumptions.

ABSTRACTAs a deep learning method, Convolutional Neural Network (CNN) can be used in image recognition, fault diagnosis and so on. Due to the internal parameter optimisation problem, the Particle Swarm Optimisation (PSO) has been introduced, but PSO is easy to fall into local optimal solution. In this paper, an adaptive CNN based on Quantum Particle Swarm Optimisation (QPSO-CNN) is proposed and applied to rolling bearings fault diagnosis, which increases the richness of particles and makes it easy to find the global optimal solution. Firstly, the one-dimensional time-series data is compressed by piecewise aggregate approximation algorithm and converted into the heat map by the Gramian angular field algorithm; Secondly, QPSO algorithm is used to search the best CNN model to fit the data set; Finally, the training and validation set are used to search the best network architecture, and the test set is used to test the diagnostic accuracy of the best CNN model, which show that the proposed method has high accuracy.

Selective electrocatalytic transformation of alcohols to aldehydes offers an efficient and environmentally friendly platform for the simultaneous production of fine chemicals and pure hydrogen gas. However, traditional alcohol oxidation reactions (AORs) in aqueous electrolyte unavoidably face competitive reactions (e.g., water oxidation and overoxidations reactions) for the presence of active oxygen species from water oxidation, causing an unwanted decrease in final efficiency and selectivity. Here, we developed an integrated all-solid proton generator-transfer electrolyzer to trigger the pure alcohol splitting reaction (ASR). In this splitting process, only O–H and C–H bonds can be cleaved at the proton generator (Pt nanoparticles), thereby completely avoiding all competitive reactions involving oxygen active species to give a > 99% selectivity to aldehydes. The as-generated protons are transported to the cathode by a three-dimensional (3D) conducting network (assemblies of ionomers and carbon spheres) for efficient hydrogen production. Unlike the poor selectivity (<22%) and durability (<3 h) of a conventional AOR electrolyzer, this ASR electrolyzer could be continuously operated at a low cell voltage of 1.2 V for at least 10 days to give a high Faradaic efficiency of 80–93% for aldehyde production.

Transformers have shown remarkable performance, however, their architecture design is a time-consuming process that demands expertise and trial-and-error. Thus, it is worthwhile to investigate efficient methods for automatically searching high-performance Transformers via Transformer Architecture Search (TAS). In order to improve the search efficiency, training-free proxy based methods have been widely adopted in Neural Architecture Search (NAS). Whereas, these proxies have been found to be inadequate in generalizing well to Transformer search spaces, as confirmed by several studies and our own experiments. This paper presents an effective scheme for TAS called TRansformer Architecture search with ZerO-cost pRoxy guided evolution (T-Razor) that achieves exceptional efficiency. First, through theoretical analysis, we discover that the synaptic diversity of multi-head self-attention (MSA) and the saliency of multi-layer perceptron (MLP) are correlated with the performance of corresponding Transformers. The properties of synaptic diversity and synaptic saliency motivate us to introduce the ranks of synaptic diversity and saliency that denoted as DSS++ for evaluating and ranking Transformers. DSS++ incorporates correlation information among sampled Transformers to provide unified scores for both synaptic diversity and synaptic saliency. We then propose a block-wise evolution search guided by DSS++ to find optimal Transformers. DSS++ determines the positions for mutation and crossover, enhancing the exploration ability. Experimental results demonstrate that our T-Razor performs competitively against the state-of-the-art manually or automatically designed Transformer architectures across four popular Transformer search spaces. Significantly, T-Razor improves the searching efficiency across different Transformer search spaces, e.g., reducing required GPU days from more than 24 to less than 0.4 and outperforming existing zero-cost approaches. We also apply T-Razor to the BERT search space and find that the searched Transformers achieve competitive GLUE results on several Neural Language Processing (NLP) datasets. This work provides insights into training-free TAS, revealing the usefulness of evaluating Transformers based on the properties of their different blocks.

Digital reconstruction of the intricate 3D morphology of individual neurons from microscopic images is a crucial challenge in both individual laboratories and large-scale projects focusing on cell types and brain anatomy. This task often fails in both conventional manual reconstruction and state-of-the-art artificial intelligence (AI)-based automatic reconstruction algorithms. It is also challenging to organize multiple neuroanatomists to generate and cross-validate biologically relevant and mutually agreed upon reconstructions in large-scale data production. Based on collaborative group intelligence augmented by AI, we developed a collaborative augmented reconstruction (CAR) platform for neuron reconstruction at scale. This platform allows for immersive interaction and efficient collaborative editing of neuron anatomy using a variety of devices, such as desktop workstations, virtual reality headsets and mobile phones, enabling users to contribute anytime and anywhere and to take advantage of several AI-based automation tools. We tested CAR’s applicability for challenging mouse and human neurons toward scaled and faithful data production.

The preparation of calcium carbonate (CaCO3) oligomers in municipal solid waste incineration fly ash (MSWI FA) is an efficient and low-carbon method for solidifying FA. This study investigated the effects and mechanisms of thermal induction on heavy metal solidification in CaCO3oligomer fly ash (COFA). Ca in FA formed CaCO3chains that crystallized into calcite after thermal induction, effectively solidifying heavy metals. The thermal induction temperature rather than induction time, heating rate, or cooling rate was found to be the most significant factor influencing solidification. As temperatures increased (200 °C∼400 °C), CaCO3in COFA was transformed into calcite, improving the solidification efficiencies of Cu, Pb, and Zn from 70.02 %, 99.91 %, and 71.68 % to above 99.99 %. However, Cr’s efficiency decreased from 96.04 % to −19.66 % due to oxidation. Switching the thermal induction atmosphere from air to nitrogen reduced Cr leaching by 62.9 %. Optimal results were achieved at 300 °C for 30 minutes, with leaching concentrations below China’s reuse standard (HJ 1134–2020). Zn in FA formed oligomer networks with Ca, and then decomposed into stable ZnO during thermal induction. Additionally, Zn promoted the conversion of CaCO3to calcite. This study provides scientific and theoretical guidance for disposing of MSWI FA using CaCO3oligomers.

Irregular bone defects, characterized by unpredictable size, shape, and depth, pose a major challenge to clinical treatment. Although various bone grafts are available, none can fully meet the repair needs of the defective area. Here, this study fabricates a dough-type hydrogel (DR-Net), in which the first dynamic network is generated by coordination between thiol groups and silver ions, thereby possessing kneadability to adapt to various irregular bone defects. The second rigid covalent network is formed through photocrosslinking, maintaining the osteogenic space under external forces and achieving a better match with the bone regeneration process. In vitro, an irregular alveolar bone defect is established in the fresh porcine mandible, and the dough-type hydrogel exhibits outstanding shape adaptability, perfectly matching the morphology of the bone defect. After photocuring, the storage modulus of the hydrogel increases 8.6 times, from 3.7 kPa (before irradiation) to 32 kPa (after irradiation). Furthermore, this hydrogel enables effective loading of P24 peptide, which potently accelerates bone repair in Sprague–Dawley (SD) rats with critical calvarial defects. Overall, the dough-type hydrogel with kneadability, space-maintaining capability, and osteogenic activity exhibits exceptional potential for clinical translation in treating irregular bone defects.