Multi⁃Instance Learning (MIL) uses labeled bags composed of multiple unlabeled instances as training data. Embedding⁃based methods address bag representation issues by embedding bags into single vectors. However，existing methods often focus on individual instances and overlook the relationship between instances and bags，which compromises the representativeness of the prototypes. Additionally，the differences between positive and negative bags are not considered by single⁃angle embedding methods，resulting in weak embedding vector quality. This paper proposes the Cluster Frequency Analysis and Dual⁃Perspective Fusion Embedding for MIL (FADE). The cluster center selection technique utilizes density peak of instances to choose a certain proportion of instances from positive and negative subspaces as cluster centers. The cluster frequency analysis technique clusters instances within subspaces based on the cluster centers，calculates cluster frequency indicators，and selects high⁃frequency cluster centers to form the prototype instance set of subspaces. The dual⁃perspective fusion embedding technique utilizes the prototype instance sets from positive and negative subspaces，along with a difference embedding function，to extract information from both perspectives and fuse the two sets of information to obtain the final embedding vector. The algorithm is tested on 29 datasets and compared with seven MIL algorithms. Experimental results demonstrate that FADE achieves higher overall classification accuracy compared to the seven benchmark algorithms，particularly excelling on image datasets while performing well on text and web datasets.

P53 gene status is an important basis for precise diagnosis and treatment of glioma. To solve the problems of incomplete heterogeneous feature extraction and multiple uncertainties inherent in the current deep learning model for MRI (Magnetic Resonance Imaging)⁃based P53 gene status prediction，we propose the precise prediction model of P53 gene status for glioma，CVT⁃RegNet (Improved RegNet integrating CNN，Vision Transfomer，and Truth Discovery). First，the RegNet network is adopted as the infrastructure of the P53 gene mutation status prediction model，which is adaptively designed to search for the heterogeneous features of the P53 gene. Second，the ViT (Vision Transfomer) module and the CNN (Convolutional Neural Networks) module are fused in the model to improve the RegNet network and further optimize the feature extraction performance and computational efficiency of the model. Finally，the Truth Discovery algorithm is incorporated for iterative optimization to improve the uncertainty of the model output，thus improving the accuracy of the prediction results. The experimental results show that the CVT⁃RegNet model predicts the P53 mutation status with an accuracy of 95.06% and an AUC score of 0.9492，which is better than the existing P53 gene status prediction models. CVT⁃RegNet realizes the non⁃invasive prediction of glioma P53 gene status，and reduces the economic burden and physical and psychological harm to patients，which provides a significant value for the precise clinical diagnosis and treatment of glioma.

Graph sampling obtains graph structures of smaller size compared to the original graph by performing approximation operations on graph data，and thus serves downstream tasks such as graph analysis and graph visualisation. Existing graph sampling algorithms focus on preserving salient structural features in the graph and ignore node attributes，leading to difficulties in achieving the expected results of sampled graphs in many downstream tasks，such as frequent pattern mining. For this reason，this paper proposes Motif⁃Based Node Biased Sampling (MNBS)，an algorithm that redefines the importance of nodes in the graph using frequent Motif substructures，followed by biased node sampling，achieving sampling that fuses node attributes with structural features. In order to quickly identify frequent Motif patterns，Fast Motif⁃Pattern Discovery (FMPD) algorithm with "early termination" is designed to efficiently and accurately discover Motif patterns to support graph sampling. Experiments show that the MNBS sampling algorithm outperforms the other baseline algorithms in a number of metrics. For example，the average reduction of Logarithm Normalized Cumulative Group Relevance is 0.54，and FMPD algorithm with the "early termination" feature reduces the time and memory consumption by 56.1% and 29.8%，respectively.

To address the limitations of Total Focusing Method (TFM) imaging technology in industrial applications due to its time⁃consuming nature，this paper presents a rapid TFM imaging approach based on a CNN⁃BiLSTM (Convolutional Neural Network⁃Bi⁃directional Long Short⁃Term Memory) network. This method initially employs a CNN to extract key features from the full matrix data，followed by leveraging a BiLSTM network to predict the location of the damage on metal plates. Subsequently，TFM technology is used for precise imaging in the damaged areas. Furthermore，to enhance the accuracy of the damage detection，this paper also introduces a damage size detection method based on the ResNet network to achieve precise measurement of the damage size. To validate the effectiveness of the proposed method，a three⁃dimensional aluminum plate simulation model was established using the finite element analysis software ABAQUS，and a neural network dataset was constructed through model transformation. Experimental results demonstrate that compared to traditional TFM imaging methods，the CNN⁃BiLSTM network exhibits higher region localization precision，with an accuracy rate of 95.26%，and a significant efficiency advantage，with an average positioning speed increased by 46.4%. Additionally，the detection results of the damage size have validated the effectiveness and accuracy of the method based on the ResNet network in damage size assessment，achieving an accuracy rate of 99.26% on the test set.

The P300 speller，a brain⁃computer interface (BCI) system enabling users to input via electroencephalogram (EEG)，plays a vital role in achieving high transmission and accuracy in detecting P300 signals to enhance the performance of P300 speller systems. To address such challenges as low signal⁃to⁃noise ratio and difficulty in feature extraction from P300 EEG signals，this study introduces a convolutional neural network (CNN) model incorporating batch normalization and residual blocks. This model aims to preserve critical features during training while expediting the convergence of the model's loss function.Performance evaluation of the proposed model and comparative algorithms is conducted using metrics such as classification accuracy，AUC，precision，recall and F₁⁃score. Results indicate that compared to traditional CNN algorithms，the proposed model achieves a 6% increase in classification accuracy and exhibites improved convergence speed of the loss function. Furthermore，when compared to traditional machine learning methods，the proposed model outperformes across all evaluation metrics. Thus，this algorithm presents an effective approach for enhancing the performance of P300 speller systems.

In recent years，the multi⁃view clustering problem has received widespread attention both domestically and internationally. The jointly smoothed multi⁃view clustering algorithm utilizes the view⁃consensus grouping effect and the local structure of multiple views to standardize the common representation of views，achieving impressive clustering results. However，this algorithm still has certain limitations in exploring inconsistency，which limits further improvement of the clustering performance. In order to further explore the inconsistency of multiple views，this paper proposes a jointly smoothed multi⁃view subspace clustering algorithm based on feature concatenation. It not only learns the consistency and inconsistency between views simultaneously to enhance view diversity，but also divides the whole inconsistency into cluster⁃specific and sample⁃specific corruptions. It is further associated with low⁃rank representations through kernel norm，and on such a basis of iterates using alternating direction minimization. Experiments conducted on four benchmark datasets have demonstrated the superiority of the proposed algorithm over other excellent algorithms.

The metric⁃based few⁃shot learning models struggle to fully exploit the relationship between the intra⁃class samples and the cross⁃class ones，and treat a single sample feature as an independent item during training,which results in inaccurate prototypes and low⁃quality representation. To handle the issue，a prototype⁃complemented few⁃shot image classification with intra⁃class and cross⁃class information is proposed. Firstly，the features of the support set are fed to the intra⁃class information extraction branch to exploit the intra⁃class feature，which is further processed to obtain the information of the category description to complement the initial prototypes. Then，the query samples from different classes are fused to generate the new samples through the cross⁃class information extraction branch. The labels of the query samples are constructed as soft labels of the new samples. Finally，the complemented prototypes are employed to classify the query and new samples，and the model is optimized by classification loss. In this paper，the comparison experiments are performed on four public few⁃shot learning datasets. The accuracy is improved by 2.03% to 5.48%，2.25% to 8.55%，2.61% to 10.03%，and 5.10% to 8.82% on the MiniImageNet dataset，TieredImageNet dataset，CUB dataset and CIFAR⁃FS dataset，respectively. Experimental results show that the proposed model achieves superior generalization and classification performance than other methods.

With the deep integration of agriculture and modern information technologies such as big data，artificial intelligence，the Internet of Things，and cloud computing，the field of modern agriculture is gradually embracing intelligence. Knowledge engineering plays a crucial role in integrating，managing，mining，and utilizing agricultural knowledge，providing robust support for personalized and precise agricultural cognitive intelligent services. This article discusses the primary challenges in agricultural knowledge engineering and cognitive intelligent services，reviews the current research status of both domestic and international agricultural cognitive intelligent services，and proposes a foundational research framework that integrates the layers of data，algorithms，and cognitive services. Building upon this framework，a novel framework for agricultural cognitive intelligent services utilizing active meta⁃learning is introduced to achieve data integrating and knowledge modeling，extraction，fusion，and reasoning for agricultural big data through bidirectional coupling with software intelligent agents and scientific big data. Key technologies and service applications involved in each step are outlined. Finally，the article concludes by summarizing future trends and offering recommendations for the development of agricultural cognitive intelligent services.

With the widespread adoption of 5G mobile video applications and the rapid surge of domestic video streaming demands，there exists an imperative for a novel data transmission protocol that ensures reliable security while satisfying the escalating requirements for increased access connections and reduced latency. The MPQUIC （Multipath Quick UDP Internet Connection） protocol is widely regarded as a pivotal element in the future mobile internet data transmission due to its unique capability to leverage multiple link bandwidth resources，exhibit robust connection fault tolerance，and deliver high reliability. However，despite its promising potential，research on the MPQUIC protocol is still in its nascent stages，and there is currently a dearth of universally adopted or open⁃source simulation platforms dedicated to studying this protocol. To address this pressing need，we have developed the MPQUIC simulation platform on the widely acclaimed NS⁃3 network simulator (ns3⁃mpquic). This open⁃source and freely available platform establishes a solid foundation for investigating the intricacies of the MPQUIC protocol，empowering global experts and scholars to simulate and explore its deployment and optimization. By providing this essential resource，we aim to facilitate in⁃depth studies and foster the advancement of the MPQUIC protocol.

A compact analog baseband for the millimeter⁃wave radar receiver is implemented on a 28 nm CMOS (Complementary Metal Oxide Semiconductor) process. The circuit consists of a three⁃stage programmable gain amplifier which is embedded with DCOC （DC Offset Cancellation） loops and a sixth⁃order Butterworth low⁃pass filter，enabling reconfigurable gain and bandwidth. Reusable resistor and capacitor arrays are used in the analog baseband，and transistors operating in the subthreshold region are introduced as active resistors into the DCOC loops，which significantly reduce the chip area. The measurement results show that the analog baseband with an area of 0.01 mm² provides a gain range of -0.6 to 68.4 dB with a gain step of 5.8 dB，and a bandwidth adjustment range of 500 kHz to 17 MHz. It achieves an output third⁃order intercept point of 22.4 dBm and consumes a power of 12 mW from a 1.8 V supply voltage.

There are many profound and unknown questions about the Universe. Among them,the research on the origin of the mass of visible matter is particularly fundamental and of far⁃reaching significance. The mass of visible matter is mostly described by a single mass scale,that is,the mass of the proton m_{\mathrm{p}}. With the discovery of the Higgs particle,we have been able to fully explain the origin of the mass of "fundamental" particles less than or equal to 1% in visible matter,that is,the mass obtained through the Higgs mechanism. The remaining mass,which is greater than or approximately equal to 99%,according to the present theory,is the consequence of gluon self interactions,which is a unique feature of quantum chromodynamics (QCD)—the theory of strong interactions in the standard model,that is,the mass generation mechanism of hadrons⁃EHM (Emergence of Hadronic Mass). In this work,we study the influences of the current quark mass,effective gluon mass,QCD running coupling,infrared as well as ultraviolet cutoffs on the constituent quark mass,by using a contact interaction model,thereby discuss the influences of changing the size of these quantities on the mass of visible matter or the presentations of the visible Universe and describe more forms of the present Universe,which strengthens our understanding of why the Universe takes these specific values on these quantities.

Dichloromethane (DCM) dehalogenase stands as a crucial enzyme implicated in the degradation of methylene chloride across diverse environmental and biological contexts. However,the unbinding pathways of ligands from DCM dehalogenase remain unexplored. In order to gain a deeper understanding of the binding sites and dissociation pathways of dichloromethane (DCM) and glutathione (GSH) from the DCM dehalogenase,random accelerated molecular dynamics (RAMD) simulations were performed,in which DCM and GSH were forced to leave the active site. The protein structure was predicted using Alphafold2,and the conformations of GSH and DCM in the binding pocket were predicted by docking. A long equilibrium simulation was conducted to validate the structure of the complex. The results show that GSH is most commonly observed in three main pathways,one of which is more important than the other two. In addition,DCM was observed to escape along a unique pathway. The key residues and protein helices of each pathway were identified. The results can provide a theoretical foundation for the subsequent dissociation mechanism of DCM dehalogenase.

Lithium⁃ion batteries (LIBs) are widely used in emerging energy applications due to their advantages such as high energy density and output power，as well as long service life. Polymer binders used in the cathode play a crucial role in LIBs as they provide adhesion between active material particles and the electrode current collector，thereby facilitating the diffusion of Li⁺ ions and improving the cycle performance of batteries. Commercial polyvinylidene fluoride (PVDF) is the main binder used in cathode materials at present，but it still has the problems of low bonding performance and poor ion conductivity. Polyimide (PI)，on the other hand，offers excellent mechanical properties，adhesive strength，and favorable design flexibility，making it an attractive alternative binder. Therefore，we designed and synthesized a sulfonimide lithium structure⁃based diamine monomer (BAPSI⁃Li)，taking the advantage of excellent mechanical properties and structural tunability of PI. The BAPSI⁃Li monomer was copolymerized with 3,3',4,4'⁃Biphenyltetracarboxylic dianhydride (s⁃BPDA) and 4,4'⁃oxydianiline (ODA) to form an ion⁃conductive PI binder. The adhesive performance of PI binder was found to be improved compared to PVDF. The LFP electrodes prepared with this ion⁃conductive PI binder exhibited outstanding electrochemical performance，with a discharge specific capacity of 158.19 mA·h·g^-1 after 250 cycles at a current density of 0.2 C. Compared with commercial PVDF binder，the ion⁃conductive PI binder demonstrated superior cycling stability. This was due to the excellent adhesive performance of PI，which enabled the electrode to maintain structural integrity during long cycling processes. Meanwhile，the introduction of the ion⁃conductive diamine monomer was beneficial for promoting the transport of lithium ions within the electrode.

Sucralose is a typical recalcitrant organic compound. However，the reducibility of sucralose，the resultant reduction product of sucralose during chemical reduction water treatment processes and the biodegradability of the reduction product are still unknown. In order to address the above issues，the reduction product of sucralose during a typical advanced reduction process of ultraviolet/sulfite (UV/\mathrm{S}{\mathrm{O}}_{\mathrm{3}}^{\mathrm{2}-}) and the biodegradability of the reduction product were investigated. The results showed that sucralose could be almost completely dechlorinated (>99.5%) by UV/\mathrm{S}{\mathrm{O}}_{\mathrm{3}}^{\mathrm{2}-} conversion while the total organic carbon (TOC) remained basically unchanged，and the rate of dechlorination improved with the increase of pH (5~9) and the increase of \mathrm{S}{\mathrm{O}}_{\mathrm{3}}^{\mathrm{2}-} concentration (0~5 mmol·L^-1). The dominant reduction product was identified as 4，1'，6'⁃trideoxysucrose through the analysis of high⁃performance liquid chromatography coupled with high resolution mass spectrometry (HPLC⁃HRMS). The evolution kinetics of sucralose and 4，1'，6'⁃trideoxysucrose revealed that the transformation of sucralose during the UV/\mathrm{S}{\mathrm{O}}_{\mathrm{3}}^{\mathrm{2}-} process was mainly a hydrodechlorination process. The evolution of the five⁃day biochemical oxygen demand (BOD₅) and chemical oxygen demand (COD) during the conversion of sucralose in UV/\mathrm{S}{\mathrm{O}}_{\mathrm{3}}^{\mathrm{2}-} process were monitored，and the biodegradability index，\mathrm{B}\mathrm{O}{\mathrm{D}}_{\mathrm{5}}/\mathrm{C}\mathrm{O}\mathrm{D} ratio (B/C)，increased dramatically from <0.05 to >0.4，and BOD₅ significantly increased to be comparable to the reference value of glucose under the identical testing conditions (99.4%)，indicating that the dominant product 4，1'，6'⁃trideoxysucrose is a well biodegradable microbial carbon source. This study provides new insights into the potential transformation pathways and environmental fate of sucralose，and paves the way for establishing standard methods employing sucralose as a model compound to evaluate the performances of various chemical reduction processes.

In order to remediate chlorinated hydrocarbons contaminated groundwater，a bioactive composite functional material was developed with coconut shell biochar and zero⁃valent iron as main raw materials. A field pilot test was performed to treat chlorinated hydrocarbons contaminated groundwater at a site in Jiangsu province. The field monitoring data revealed that the composite functional material could continuously release organic carbon and iron，facilitating the growth of dechlorinating bacteria. The 16S rRNA gene sequencing showed that after 1140 days of PRB operation，the relative abundance of dechlorinating bacteria in the composite functional materials (1.9%~8.3%) was higher than that in the initial PRB material (0.5%). During 1140 days of PRB operation，the surface morphology of the composite functional material remained stable. The removal efficiencies of various chlorinated hydrocarbons compounds by PRB ranged from 13.7% to 100% during the 2000⁃day operation period. The composite functional material can effectively remove chlorinated hydrocarbons from groundwater through the combined mechanism of adsorption，iron reduction and microbial degradation. The research could provide important technical support for the in⁃situ remediation of groundwater contaminated with chlorinated hydrocarbons.

Thioarsenates have been proved to be important species of arsenic in the soil⁃rice system，their transformations playing a crucial role in the biogeochemical cycle of arsenic and posing a potential threat to the environment and human health. This article provides a comprehensive overview of the latest research progress on quantitative detection methods，species，influence factors，biogeochemical processes and rice uptake mechanisms of thioarsenates in paddy soil，in an effort to provide methodological references and theoretical foundation for future research. To minimize the transformations of thioarsenates before analysis，diethylenetriamine⁃pentaacetic acid (DTPA) can be added to the sample，which can be flash⁃frozen and stored in the dark at low temperature. Ion chromatography coupled to inductively coupled plasma–mass spectrometry (IC⁃ICP⁃MS) is suitable for the separation and quantitative determination of thioarsenates in environmental samples. There are significant differences between inorganic and methylated thioarsenates in their formation conditions，environmental behaviors and rice uptake. Concentrations of thioarsenates in paddy soil are affected by many factors such as pH，redox environment and reduced sulfur. Changes of environmental conditions can facilitate transformations between various thioarsenates or between thioarsenates and non⁃thioarsenic species. Notably，the adsorption capacities of thioarsenates on iron minerals are usually weaker compared with arsenite and arsenate，and thioarsenates show higher instability with the changes of the soil redox conditions. Thioarsenates can be absorbed by rice and accumulated in grains. Dimethymonothioarsenate (DMMTA) exhibits high uptake，transport，and toxicity in rice，and it exist in rice grains and commercial rice globally. Currently，the mechanisms of uptake，transport and detoxification of thioarsenates in rice remain unclear，which is worthy of further research to accurately assess the potential risks of thioarsenates on rice growth，food safety，and human health.