Superconducting quantum circuit is one of the leading approaches for realizing universal quantum computation. As the chip integration continues to increase，circuit design faces the challenges of complex wiring and signal crosstalk. In order to suppress the signal crosstalk on the chip，a novel on⁃chip structure covering coplanar waveguides，called tunnel bridge，is proposed based on the electromagnetic shielding principle of the coaxial line. In this paper，we simulate the chip circuit，with the tunnel bridge added，by using the finite element method，and the optimal impedance matching design parameters are obtained. The simulation results of the chip signal crosstalk show that，compared with conventional coplanar waveguides，the addition of the tunnel bridge brings three orders of magnitude shielding effect on the vertical electric field between top and bottom chips in flip⁃chip design. For planar single chip，the crosstalk between microwave drive lines are reduced by approximately 16 dB，and the crosstalk current caused by the DC flux bias lines are reduced by around 70%，yielding a 40% improvement over traditional air bridge schemes. The micro⁃fabrication process of tunnel bridge is also studied in this paper，and a stable sample fabrication process is achieved. The tunnel bridge structure has remarkable crosstalk suppression effect，thus a great prospect of application in large⁃scale superconducting qubit chips.
This paper proposes a parallel multi⁃path transmission design based on the CFDP (CCSDS File Delivery Protocol)，which expands the parallel transmission of files over multiple paths in space network file transfers. This design further enhances the efficiency of file transfer while maintaining compatibility with the original file transfer protocol. The design fully utilizes the characteristics of space network links and achieves the goal of parallel multi⁃path transmission by using load balancing algorithms for data scheduling to allocate file data to transmission paths reasonably. The simulation results demonstrate that compared with traditional CFDP transmission，this parallel multi⁃path transmission design fully utilizes multiple transmission paths，significantly reducing the file delivery time.
In response to the problem of time⁃varying and unknown number of interfering Unmanned Aerial Vehicles (UAVs)，as well as low Signal to Interference Ratio (SIR) and Signal to Noise Ratio (SNR) of Earth station observation signals in the downlink of the Multiple Input Multiple Output (MIMO) communication system of Geostationary Earth Orbit (GEO) satellites facing malicious interference from UAV clusters，this paper proposes a FPN⁃BiLSTM network architecture and its training method that combines Feature Pyramid Network (FPN) and Bidirectional Long Short Term Memory (BiLSTM) networks to achieve end⁃to⁃end underdetermined blind source separation with time⁃varying number of sources. By applying this algorithm，the expected signal bit sequence is directly extracted from underdetermined mixing observation signals without the need for estimating the number of interference sources or traditional signal demodulation. The simulation result shows that compared to other underdetermined blind source separation algorithms based on deep learning networks，the proposed algorithm has a low bit error rate and good interference cancellation performance in scenarios where the number of source signals is time⁃varying and SNR and SIR are as low as
To meet the requirements for high linearity and out⁃of⁃band signal suppression in the receive system of RF communication，a high linearity RF receiver front⁃end chip designed based on SOI （Silicon on Insulator） technology has been proposed. The chip integrates low noise amplifier，RF switch and auxiliary circuit. The low noise amplifier adopts MGTR (Multiple Gated Transistor) structure which is based on transconductance nonlinear compensation technology to eliminate the third⁃order transconductance of transistors，and the resonant network in the matching is applied to suppress the second⁃order nonlinearity in the amplifier，enabling the LNA to achieve good linearity performance. There is a bypass branch in the RF switch，which bypasses the signal when the input signal is too large to protect the low noise amplifier. The measurement results of the chip show that in the operating frequency band of 2.4 to 2.5 GHz，the receiver front⁃end chip is powered by a 2 V voltage，with a current of 16 mA，a gain of more than 15.4 dB，a noise figure of less than 1.76 dB，an input and output return loss of less than -15 dB，an input third order modulation point of 11 dBm at 2.4 GHz，a 1 dB compression point of -1.92 dBm， and an out⁃of⁃band interference rejection in the 5~6 GHz， and the total area of the chip is 700 µm×700 µm.
As one of the important directions in clustering analysis，deep graph clustering is attracting more and more attention from academia and industry. Existing deep graph clustering methods tend to focus on the model⁃level，that is，by adjusting the module structure to achieve better performance. It means that existing methods usually ignore information enhancement at the data⁃level and rely solely on the original data. When there is noise or loss in the original data，the performance of the graph clustering methods that relies heavily on it will decline to varying degrees. To solve this problem，we propose a novel method that utilizes the diffusion strategy to enhance Multi⁃View Graph Clustering，called DMVGC. Specifically，DMVGC applies the idea of diffusion technology to graph data and promotes more effective edges based on the original graph structure to generate multiple views，thereby providing the model with a broader clustering view and richer information propagation. The experimental results show that this DMVGC method achieves performance improvement compared with existing methods without affecting the training speed.
Causal inference helps people make more rational decision⁃making plans and has wide applications in fields such as e⁃commerce and precision medicine，and its performance relies critically on the accurate estimation of Individual Treatment Effect (ITE). The selection bias problem and the sample imbalance problem in the observational data affect the accuracy of the individual treatment effect estimation. For the selection bias problem，existing deep learning methods mainly mitigate it by balancing all the covariates，but balancing the processing⁃independent noise variables in the covariates can lead to inaccurate estimation of individual treatment effect. For the sample imbalance problem，these methods mainly mitigate it by adding sample weights to the loss function. However，this practice does not effectively improve the accuracy of neural network prediction model. In this paper，we propose a method based on deep representation learning，which jointly induces neural networks to obtain balanced shared representations of non⁃noise variables in the covariates through gnn and IPM (Integral Probability Metric) networks，and then introduces the X⁃Net to alleviate the sample imbalance problem. The experimental results on semi⁃synthetic and real datasets respectively show that our algorithm can improve the accuracy of the model individual treatment effect estimation by mitigating the sample selection bias problem and sample imbalance problem.
Random noise denoising is an important step in seismic data processing. Many methods based on convolutional neural networks only consider single⁃scale features and cannot adaptively linearly aggregate seismic data features，resulting in difficulties in removing complex noise and protecting weak signals. In this paper，a multi⁃scale convolutional residual seismic denoising network fusing quadruple attention mechanisms (MARN) is proposed. It consists of three main parts: a single⁃scale feature extraction layer，a multi⁃scale feature fetching layer，and a feature recovery layer. The single⁃scale feature extraction layer uses a single identical convolutional kernel to extract global coarse features. The multiscale feature extraction layer contains multiple residual multi⁃scale attention feature extraction blocks (RMSAB)，each consisting of multiple multiaxial attention multi⁃scale feature fusion blocks (MAFB).The MAFBs contain three structures: the feature extraction structure extracts the local fine features through the four attentional mechanisms，the feature fusion structure fuses the features extracted by the four attentional mechanisms，and the feature transfer structure delivers the features to the feature recovery layer. The feature recovery layer fuses the extracted single⁃scale and multi⁃scale features to obtain denoised seismic data. The experimental results show that MARN can not only remove random noise in a more targeted way，but also retain the weak signals better.
Track defect detection is a critical task for ensuring the safe operation and maintenance of rail transportation. Existing machine vision⁃based detection methods mainly focus on segmentation of railway images，which have high model time complexity，serious background noise interference and poor segmentation effects. This paper proposes an improved multi⁃task image segmentation model based on Segment Anything Model (Multi⁃Task Advanced SAM，MASAM) to effectively improve training efficiency and defect segmentation accuracy. First，the defect scope is determined and the boundary coordinates are obtained through the target detection module; then the boundary coordinates are converted into sparse embeddings; finally，the sparse embeddings and the image feature vectors processed by the Image Encoder module in the SAM model are inputted to the Mask Decoder part to obtain the defect mask prediction result. The experimental results show that the prediction efficiency and accuracy of the MASAM model are superior to other models in multi⁃task railway defect detection.
Graph Neural Networks (GNNs) have achieved notable success in node classification tasks. However，current GNN models tend to focus on majority classes with a large amount of labeled data，paying little attention to minority classes with fewer labels. Traditional methods often address this issue through oversampling，which may lead to overfitting. Some recent studies suggest synthesizing additional nodes for minority classes from labeled nodes，yet there's no clear guarantee that these generated nodes truly represent the corresponding minority classes. In fact，incorrect synthetic nodes may undermine the generalization ability of the algorithm. To address this issue，this paper introduces a simple，self⁃supervised data augmentation method based on adversarial training，GraphA2，which enhances the data by adding perturbations at the farthest gradient space around minority classes while using contrastive learning to ensure consistency after augmentation. This approach not only increases the diversity of data but also ensures smoothness and coherence across the entire space，thereby enhancing generalization capability. The experiments show that this method outperforms the current state⁃of⁃the⁃art baseline models on various imbalanced datasets.
It is of great clinical significance to extract the specific neuroimaging features of patients diagnosed with seismic post⁃traumatic stress disorder (PTSD) and design its categorical model. Compared with the traditional neuroimaging classification model,this study used prior knowledge to extract ROIs to reduce a large number of noisy signals. Meanwhile,the research also changed the functional connection mode and proposed a classification model named RFE⁃MLP,which effectively improved the classification accuracy of post⁃earthquake PTSD and post⁃earthquake non⁃PTSD control groups. Firstly,in view of a large number of noisy signals in the whole brain,the PTSD⁃related brain regions were extracted by using prior knowledge,focusing on the feature construction of PTSD⁃related brain regions after the earthquake. Secondly,the traditional node function connection was replaced by the edge⁃centric function connection to obtain the high⁃order feature information of the brain function connection. Finally,the RFE⁃MLP classification model was proposed,which learned the model weight information adaptively and improved the accuracy of PTSD after earthquake. The results showed that the classification accuracy （ACC）,sensitivity （SEN）,specificity (SPE) and area under curve (AUC) values of post⁃earthquake PTSD patients and post⁃earthquake health control groups reached 92.5%,93.5%,92.1% and 94.8%,and it was found that the left middle frontal gyrus (Frontal_Mid_L) had the highest correlation with post⁃earthquake PTSD severity. The experimental results show that this method extracts more key feature information from fMRI,improves the accuracy of classification and diagnosis of PTSD after earthquake,and facilitates the precise localization of brain regions related to PTSD after earthquake and other trauma types.
Image⁃text matching aims to achieve high⁃quality semantic alignment between images and texts，which is an important task in the cross⁃disciplinary field of computer vision and natural language processing. Images and texts are two distinct mediums for conveying information. However，their differences in the content and distribution lead to uncertainty and ambiguity in fine⁃grained cross⁃modal information correlation. To address the challenges and enhance fine⁃grained alignment between images and texts，BSEM⁃Net (Bidirectional Semantic Embedding for Fine⁃Grained Image⁃Text Matching) is proposed. Firstly，in order to reduce redundancy in image information，this paper introduces IE (Image Semantic Embedding Module) that utilizes text words as supervisory signals to guide the model in constraining the expression of irrelevant image regions. Secondly，to reduce the distribution differences between modalities and establish fine⁃grained semantic alignment，this paper introduces TE (Text Semantic Embedding Module) that utilizes image regions to select words and transform these words into phrases that exhibit a similar information distribution to the image regions. In addition，the two modules utilize region relationship connectivity graphs and phrase relationship connectivity graphs to mine contextual information between intra modal features，reducing semantic divergence. Experimental comparisons are conducted on publicly available cross⁃modal retrieval datasets Flickr30k and MSCOCO，and the results demonstrate that the proposed method has significant superiority over existing methods in image⁃text matching tasks.
It is very important to develop non⁃precious metal catalysts of low cost，high activity and high stability for oxygen evolution of electrolytic water (OER). Fe⁃doped (Ni,Mn)Co2O4 nanorod array was prepared using bimetallic MnCo as precursor and K3[Fe(CN)6] as iron source. The morphology and structure of the as⁃synthesized nanomaterials were analyzed by scanning electron microscope (SEM)，X⁃ray diffraction (XRD)，transmission electron microscopy (TEM) and the oxygen evolution performance was investigated by electrochemical workstation. The results showed that the iron content of (Ni,Mn)Co2O4 nanorods could be controlled by adjusting the concentration of the iron source. Fe⁃doped (Ni,Mn)Co2O4⁃6 catalyst had better oxygen evolution performance in 1.0 mol·L-1 KOH electrolyte，and the overpotential was only 242 mV when the current density was 10 mA·cm-2. At the same time，the catalyst had extraordinary stability. After continuously working at the current density of 100 mA·cm-2 for 155 h，the overpotential hardly fluctuated.
Non⁃Hermitian metamaterials provide novel phenomena that cannot be easily obtained in traditional Hermitian systems by regulating the imaginary part of material parameters. Among them，acoustic conjugate materials are designed on the basis of the non⁃Hermitian acoustic principle，in which the equivalent acoustic parameters of the material，the effective mass density and volume compression coefficient，are conjugate to each other. Acoustic conjugate materials provide unique acoustic phenomena，acoustic laser mode and coherent perfection absorption，by parameter regulation. In realization of acoustic conjugate materials，a simplified design of acoustic conjugate materials is proposed in this study and special acoustic properties are investigated by theoretical analysis and numerical simulation. It is observed that the simplified design can realize acoustic laser mode and coherent perferct absorption. Meanwhile，it is shown that the phase difference of the incident wave and the asymmetric condition of the beamwidth exert influences on the performance of the materials. This study will bring new possibilities and prospects in designing various acoustic devices，such as acoustic absorbers，acoustic excitation，and small acoustic wave modulation.
The high⁃energy density，high⁃nickel terpolymer lithium⁃ion battery system puts forward higher requirements for the key positive electrode binder. The use of green and clean biomass materials instead of traditional petroleum⁃based materials is an important measure to achieve the dual⁃carbon energy conservation and emission reduction. In order to solve the problems of single positive electrode binder and low cycle stability of lithium⁃ion battery，a lysine⁃based soluble polyimide binder (PMDA⁃Lys) was designed and synthesized by the chemical imide method，which can meet the performance requirements of high⁃nickel terpolymer LiNi0.8Co0.1Mn0.1O2 (NCM811) system. The side chain carboxyl group of the binder can form a strong interaction force with the surface of the electrode particles，and the peel strength can reach 0.213 N·mm-1. The electrode prepared based on this can maintain the stability of the structure in the electrolyte environment with the continuous process of charge and discharge. Compared with traditional polyvinylidene fluoride (PVDF) adhesives，PMDA⁃Lys binders show better electrochemical performance and cycle stability in cyclic tests with lithium sheets (capacity retention greater than 85% after 100 cycles at 0.5 C). This bio⁃based polyimide binder provides new insights into the design of positive binders for lithium⁃ion batteries with green and high energy density.
Research on active noise control in the cabin of vehicles has generally focused on scenarios where the vehicle speed is below 80 km·h-1 and road noise predominates. However，the active control of cabin noise in high⁃speed driving vehicles，which mainly comprises road noise and wind noise，has received limited attention. Therefore，this study investigates the issue through offline simulations based on measured noise data and transfer functions. Firstly，experiments were conducted in a wind tunnel laboratory，a semi⁃anechoic chamber with a chassis dynamometer and a high⁃speed circular road to measure the road noise，the wind noise，and the total noise levels at various vehicle speeds. The proportion of the road noise and wind noise in the total noise，along with the requirements for noise reduction，were analyzed. Subsequently，decoupling analyses of the active noise control performance for the road noise and the wind noise were performed based on the measured noise data and the transfer functions from the aforementioned scenarios. The influence of the quantity and placement of reference sensors on control performance was investigated. The results indicated that employing 24 vibration references could reduce the road noise by over 10 dBA. However，even with more than 100 vibration references，the reduction in the wind noise remained below 5 dBA，primarily due to the poor coherence between vibration reference signals and error signals under wind noise excitation. Finally，the optimal placement of 24 vibration reference sensors was analyzed using a genetic algorithm. It was found that nearly half of the reference sensors should be primarily placed on the vehicle chassis and the interior floor，ensuring a significant reduction in the noise transmitted through this path. Additional reference sensors should be distributed across other locations on the vehicle body to achieve higher noise reduction levels.
In order to study the effect of basalt on crop growth and dry weight of grains,we investigated the effects of applying basalt powder with different grain sizes (coarse particle size <1000 μm,fine particle size <100 μm,applied amount 5 kg·m-2) to agricultural soils under greenhouse conditions on the dry weight of wheat grains,and on the content of the major elements (K,Ca,Mg) in wheat organs at different growth stages and potential soil exchangeable cations. The results showed that the application of basalt powder could promote the growth of wheat plants (especially leaves and spikes) and increase the dry weight of grains. During the entire stage of wheat growth,there was no significant change in the content of single major elements in wheat organs,but the content of K,Ca,Mg in wheat plants increased with wheat growth,and the effect of applying basalt was more significant（P<0.01）. The application of basalt did not increase the total amount of potential exchangeable ions in the soil（P<0.01）. This study shows that the application of basalt powder increases the dry weight of wheat grains,promotes wheat growth,and improves wheat quality.
With the development of high⁃performance computing，CFD (Computational Fluid Dynamics) methods are widely used in atmospheric research，especially in the fields of micro scale processes and atmospheric boundary layer problems. The investigations benefit from the features of fine meshing，well⁃representing of complex terrain or buildings，and multi⁃optional turbulence closure schemes for the numerical simulations of urban micro⁃climate/meteorology，atmospheric boundary layer flow over complex terrain and improvement of parameterizations in mesoscale models. A literature review is given in this paper. In atmospheric science，there are three challenges to further apply CFD methods. As the simulation range expands，the number of computational grids will rapidly increase，and the computational efficiency of CFD will decrease. At present，CFD methods lack effective descriptions of water and heat exchange between soil and atmosphere，atmospheric radiation process of particulate matter in urban pollution，and cloud microphysical processes. CFD methods have not yet effectively described boundary forcing. All these issues require continuous research.
