By sampling perturbed state vectors from each ensemble prediction run at properly selected time levels in the vicinity of the analysis time, the recently proposed time-expanded sampling approach can enlarge the ensemb...By sampling perturbed state vectors from each ensemble prediction run at properly selected time levels in the vicinity of the analysis time, the recently proposed time-expanded sampling approach can enlarge the ensemble size without increasing the number of prediction runs and, hence, can reduce the computational cost of an ensemble-based filter. In this study, this approach is tested for the first time with real radar data from a tornadic thunderstorm. In particular, four assimilation experiments were performed to test the time-expanded sampling method against the conventional ensemble sampling method used by ensemble- based filters. In these experiments, the ensemble square-root filter （EnSRF） was used with 45 ensemble members generated by the time-expanded sampling and conventional sampling from 15 and 45 prediction runs, respectively, and quality-controlled radar data were compressed into super-observations with properly reduced spatial resolutions to improve the EnSRF performances. The results show that the time-expanded sampling approach not only can reduce the computational cost but also can improve the accuracy of the analysis, especially when the ensemble size is severely limited due to computational constraints for real-radar data assimilation. These potential merits are consistent with those previously demonstrated by assimilation experiments with simulated data.展开更多
In the Ensemble Kalman Filter（EnKF） data assimilation-prediction system,most of the computation time is spent on the prediction runs of ensemble members.A limited or small ensemble size does reduce the computational...In the Ensemble Kalman Filter（EnKF） data assimilation-prediction system,most of the computation time is spent on the prediction runs of ensemble members.A limited or small ensemble size does reduce the computational cost,but an excessively small ensemble size usually leads to filter divergence,especially when there are model errors.In order to improve the efficiency of the EnKF data assimilation-prediction system and prevent it against filter divergence,a time-expanded sampling approach for EnKF based on the WRF（Weather Research and Forecasting） model is used to assimilate simulated sounding data.The approach samples a series of perturbed state vectors from Nb member prediction runs not only at the analysis time（as the conventional approach does） but also at equally separated time levels（time interval is △t） before and after the analysis time with M times.All the above sampled state vectors are used to construct the ensemble and compute the background covariance for the analysis,so the ensemble size is increased from Nb to Nb＋2M×Nb=（1＋2M）×Nb） without increasing the number of prediction runs（it is still Nb）.This reduces the computational cost.A series of experiments are conducted to investigate the impact of △t（the time interval of time-expanded sampling） and M（the maximum sampling times） on the analysis.The results show that if t and M are properly selected,the time-expanded sampling approach achieves the similar effect to that from the conventional approach with an ensemble size of（1＋2M）× Nb,but the number of prediction runs is greatly reduced.展开更多
One of the detection objectives of the Chinese Asteroid Exploration mission is to investigate the space environment near the Main-belt Comet(MBC,Active Asteroid)311P/PANSTARRS.This paper outlines the scientific object...One of the detection objectives of the Chinese Asteroid Exploration mission is to investigate the space environment near the Main-belt Comet(MBC,Active Asteroid)311P/PANSTARRS.This paper outlines the scientific objectives,measurement targets,and measurement requirements for the proposed Gas and Ion Analyzer(GIA).The GIA is designed for in-situ mass spectrometry of neutral gases and low-energy ions,such as hydrogen,carbon,and oxygen,in the vicinity of 311P.Ion sampling techniques are essential for the GIA's Time-of-Flight(TOF)mass analysis capabilities.In this paper,we present an enhanced ion sampling technique through the development of an ion attraction model and an ion source model.The ion attraction model demonstrates that adjusting attraction grid voltage can enhance the detection efficiency of low-energy ions and mitigate the repulsive force of ions during sampling,which is influenced by the satellite's surface positive charging.The ion source model simulates the processes of gas ionization and ion multiplication.Simulation results indicate that the GIA can achieve a lower pressure limit below 10-13Pa and possess a dynamic range exceeding 10~9.These performances ensure the generation of ions with stable and consistent current,which is crucial for high-resolution and broad dynamic range mass spectrometer analysis.Preliminary testing experiments have verified GIA's capability to detect gas compositions such as H2O and N2.In-situ measurements near 311P using GIA are expected to significantly contribute to our understanding of asteroid activity mechanisms,the evolution of the atmospheric and ionized environments of main-belt comets,the interactions with solar wind,and the origin of Earth's water.展开更多
Electro-Optic Sampling(EOS)detection technique has been widely used in terahertz science and tech⁃nology,and it also can measure the field time waveform of the few-cycle laser pulse.Its frequency response and band lim...Electro-Optic Sampling(EOS)detection technique has been widely used in terahertz science and tech⁃nology,and it also can measure the field time waveform of the few-cycle laser pulse.Its frequency response and band limitation are determined directly by the electro-optic crystal and duration of the probe laser pulse.Here,we investigate the performance of the EOS with thin GaSe crystal in the measurement of the mid-infrared few-cycle la⁃ser pulse.The shift of the central frequency and change of the bandwidth induced by the EOS detection are calcu⁃lated,and then the pulse distortions induced in this detection process are discussed.It is found that this technique produces a red-shift of the central frequency and narrowing of the bandwidth.These changings decrease when the laser wavelength increases from 2μm to 10μm.This work can help to estimate the performance of the EOS de⁃tection technique in the mid-infrared band and offer a reference for the related experiment as well.展开更多
Critical Height Sampling(CHS)estimates stand volume free from any model and tree form assumptions.Despite its introduction more than four decades ago,CHS has not been widely applied in the field due to perceived chall...Critical Height Sampling(CHS)estimates stand volume free from any model and tree form assumptions.Despite its introduction more than four decades ago,CHS has not been widely applied in the field due to perceived challenges in measurement.The objectives of this study were to compare estimated stand volume between CHS and sampling methods that used volume or taper models,the equivalence of the sampling methods,and their relative efficiency.We established 65 field plots in planted forests of two coniferous tree species.We estimated stand volume for a range of Basal Area Factors(BAFs).Results showed that CHS produced the most similar mean stand volume across BAFs and tree species with maximum differences between BAFs of 5-18m^(3)·ha^(−1).Horizontal Point Sampling(HPS)using volume models produced very large variability in mean stand volume across BAFs with the differences up to 126m^(3)·ha^(−1).However,CHS was less precise and less efficient than HPS.Furthermore,none of the sampling methods were statistically interchangeable with CHS at an allowable tolerance of≤55m^(3)·ha^(−1).About 72%of critical height measurements were below crown base indicating that critical height was more accessible to measurement than expected.Our study suggests that the consistency in the mean estimates of CHS is a major advantage when planning a forest inventory.When checking against CHS,results hint that HPS estimates might contain potential model bias.These strengths of CHS could outweigh its lower precision.Our study also implies serious implications in financial terms when choosing a sampling method.Lastly,CHS could potentially benefit forest management as an alternate option of estimating stand volume when volume or taper models are lacking or are not reliable.展开更多
Selection of negative samples significantly influences landslide susceptibility assessment,especially when establishing the relationship between landslides and environmental factors in regions with complex geological ...Selection of negative samples significantly influences landslide susceptibility assessment,especially when establishing the relationship between landslides and environmental factors in regions with complex geological conditions.Traditional sampling strategies commonly used in landslide susceptibility models can lead to a misrepresentation of the distribution of negative samples,causing a deviation from actual geological conditions.This,in turn,negatively affects the discriminative ability and generalization performance of the models.To address this issue,we propose a novel approach for selecting negative samples to enhance the quality of machine learning models.We choose the Liangshan Yi Autonomous Prefecture,located in southwestern Sichuan,China,as the case study.This area,characterized by complex terrain,frequent tectonic activities,and steep slope erosion,experiences recurrent landslides,making it an ideal setting for validating our proposed method.We calculate the contribution values of environmental factors using the relief algorithm to construct the feature space,apply the Target Space Exteriorization Sampling(TSES)method to select negative samples,calculate landslide probability values by Random Forest(RF)modeling,and then create regional landslide susceptibility maps.We evaluate the performance of the RF model optimized by the Environmental Factor Selection-based TSES(EFSTSES)method using standard performance metrics.The results indicated that the model achieved an accuracy(ACC)of 0.962,precision(PRE)of 0.961,and an area under the curve(AUC)of 0.962.These findings demonstrate that the EFSTSES-based model effectively mitigates the negative sample imbalance issue,enhances the differentiation between landslide and non-landslide samples,and reduces misclassification,particularly in geologically complex areas.These improvements offer valuable insights for disaster prevention,land use planning,and risk mitigation strategies.展开更多
A comprehensive fishery-independent survey generally incorporates various specialized surveys and integrates different survey objectives to maximize benefits while accounting for cost limitations.It is important to ev...A comprehensive fishery-independent survey generally incorporates various specialized surveys and integrates different survey objectives to maximize benefits while accounting for cost limitations.It is important to evaluate the adaptability of the comprehensive survey for different taxon to get the optimal design.However,the validity and adaptability of ichthyoplankton sampling incorporated in a comprehensive fishery-independent survey program in estimating abundance of ichthyoplankton species is little known.This study included ichthyoplankton sampling in an integrated survey and assessed the appropriateness of survey design.The Kriging interpolation based on Gaussian models was used to estimate the values at unsurveyed locations based on the original ichthyoplankton survey data in the Haizhou Bay as the“true”values.The sampling performances of the ongoing stratified random sampling(StRS),simple random sampling(SRS),cluster sampling(CS),hexagonal systematic sampling(SYS h),and regular systematic sampling(SYS r)with different sample sizes in estimating ichthyoplankton abundance were compared in relative estimation error(REE),relative bias(RB),and coefficient of variation(CV)by computer simulation.The ongoing StRS performed better than CS and SRS,but not as good as the two systematic sampling methods,and the current sample size in StRS design was insufficient to estimate ichthyoplankton abundance.The average REE values(meanREE)were significantly smaller in two systematic sampling designs than those in other three sampling designs,and the two systematic sampling designs could maintain good inter-annual stability of sampling performances.It is suggested that incorporating ichthyoplankton survey directly into stratified random fishery-independent surveys could not achieve the desired level of accuracy for survey objectives,but the accuracy can be improved by setting additional stations.The assessment framework presented in this study serves as a reference for evaluating the adaptability of integrated surveys to different objectives in other waters.展开更多
When dealing with imbalanced datasets,the traditional support vectormachine(SVM)tends to produce a classification hyperplane that is biased towards the majority class,which exhibits poor robustness.This paper proposes...When dealing with imbalanced datasets,the traditional support vectormachine(SVM)tends to produce a classification hyperplane that is biased towards the majority class,which exhibits poor robustness.This paper proposes a high-performance classification algorithm specifically designed for imbalanced datasets.The proposed method first uses a biased second-order cone programming support vectormachine(B-SOCP-SVM)to identify the support vectors(SVs)and non-support vectors(NSVs)in the imbalanced data.Then,it applies the synthetic minority over-sampling technique(SV-SMOTE)to oversample the support vectors of the minority class and uses the random under-sampling technique(NSV-RUS)multiple times to undersample the non-support vectors of the majority class.Combining the above-obtained minority class data set withmultiple majority class datasets can obtainmultiple new balanced data sets.Finally,SOCP-SVM is used to classify each data set,and the final result is obtained through the integrated algorithm.Experimental results demonstrate that the proposed method performs excellently on imbalanced datasets.展开更多
The sap flow method is widely used to estimate forest transpiration.However,at the individual tree level it has spatiotemporal variations due to the impacts of environmental conditions and spatial relationships among ...The sap flow method is widely used to estimate forest transpiration.However,at the individual tree level it has spatiotemporal variations due to the impacts of environmental conditions and spatial relationships among trees.Therefore,an in-depth understanding of the coupling effects of these factors is important for designing sap flow measurement methods and performing accurate assessments of stand scale transpiration.This study is based on observations of sap flux density(SF_(d))of nine sample trees with different Hegyi’s competition indices(HCIs),soil moisture,and meteorological conditions in a pure plantation of Larix gmelinii var.principis-rupprechtii during the 2021 growing season(May to September).A multifactorial model of sap flow was developed and possible errors in the stand scale sap flow estimates associated with sample sizes were determined using model-based predictions of sap flow.Temporal variations are controlled by vapour pressure deficit(VPD),solar radiation(R),and soil moisture,and these relationships can be described by polynomial or saturated exponential functions.Spatial(individual)differences were influenced by the HCI,as shown by the decaying power function.A simple SF_(d)model at the individual tree level was developed to describe the synergistic influences of VPD,R,soil moisture,and HCI.The coefficient of variations(CV)of the sap flow estimates gradually stabilized when the sample size was>10;at least six sample trees were needed if the CV was within 10%.This study improves understanding of the mechanisms of spatiotemporal variations in sap flow at the individual tree level and provides a new methodology for determining the optimal sample size for sap flow measurements.展开更多
Class imbalance can substantially affect classification tasks using traditional classifiers,especially when identifying instances of minority categories.In addition to class imbalance,other challenges can also hinder ...Class imbalance can substantially affect classification tasks using traditional classifiers,especially when identifying instances of minority categories.In addition to class imbalance,other challenges can also hinder accurate classification.Researchers have explored various approaches to mitigate the effects of class imbalance.However,most studies focus only on processing correlations within a single category of samples.This paper introduces an ensemble framework called Inter-and Intra-Class Overlapping Ensemble(llCOE),which incorporates two sampling methods.The first method,which is based on classification hardness undersampling,targets majority category samples by using simple samples as the foundation for classification and improving performance by focusing on samples near classification boundaries.The second method addresses the issue of overfitting minority category samples in undersampling and ensemble learning.To mitigate this,an adaptive augment hybrid sampling method is proposed,which enhances the classification boundary of samples and reduces overfitting.This paper conducts multiple experiments on 15 public datasets and concludes that the IlCOE ensemble framework outperforms other ensemble learning algorithms in classifying imbalanced data.展开更多
In this paper, we use sample average approximation with adaptive multiple importance sampling to explore moderate deviations for the optimal values. Utilizing the moderate deviation principle for martingale difference...In this paper, we use sample average approximation with adaptive multiple importance sampling to explore moderate deviations for the optimal values. Utilizing the moderate deviation principle for martingale differences and an appropriate Delta method, we establish a moderate deviation principle for the optimal value. Moreover, for a functional form of stochastic programming, we obtain a functional moderate deviation principle for its optimal value.展开更多
Optical solitons,as self-sustaining waveforms in a nonlinear medium where dispersion and nonlinear effects are balanced,have key applications in ultrafast laser systems and optical communications.Physics-informed neur...Optical solitons,as self-sustaining waveforms in a nonlinear medium where dispersion and nonlinear effects are balanced,have key applications in ultrafast laser systems and optical communications.Physics-informed neural networks(PINN)provide a new way to solve the nonlinear Schrodinger equation describing the soliton evolution by fusing data-driven and physical constraints.However,the grid point sampling strategy of traditional PINN suffers from high computational complexity and unstable gradient flow,which makes it difficult to capture the physical details efficiently.In this paper,we propose a residual-based adaptive multi-distribution(RAMD)sampling method to optimize the PINN training process by dynamically constructing a multi-modal loss distribution.With a 50%reduction in the number of grid points,RAMD significantly reduces the relative error of PINN and,in particular,optimizes the solution error of the(2+1)Ginzburg–Landau equation from 4.55%to 1.98%.RAMD breaks through the lack of physical constraints in the purely data-driven model by the innovative combination of multi-modal distribution modeling and autonomous sampling control for the design of all-optical communication devices.RAMD provides a high-precision numerical simulation tool for the design of all-optical communication devices,optimization of nonlinear laser devices,and other studies.展开更多
Neutron time-of-flight(ToF)measurement is a highly accurate method for obtaining the kinetic energy of a neutron by measuring its velocity,but requires precise acquisition of the neutron signal arrival time.However,th...Neutron time-of-flight(ToF)measurement is a highly accurate method for obtaining the kinetic energy of a neutron by measuring its velocity,but requires precise acquisition of the neutron signal arrival time.However,the high hardware costs and data burden associated with the acquisition of neutron ToF signals pose significant challenges.Higher sampling rates increase the data volume,data processing,and storage hardware costs.Compressed sampling can address these challenges,but it faces issues regarding optimal sampling efficiency and high-quality reconstructed signals.This paper proposes a revolutionary deep learning-based compressed sampling(DL-CS)algorithm for reconstructing neutron ToF signals that outperform traditional compressed sampling methods.This approach comprises four modules:random projection,rising dimensions,initial reconstruction,and final reconstruction.Initially,the technique adaptively compresses neutron ToF signals sequentially using three convolutional layers,replacing random measurement matrices in traditional compressed sampling theory.Subsequently,the signals are reconstructed using a modified inception module,long short-term memory,and self-attention.The performance of this deep compressed sampling method was quantified using the percentage root-mean-square difference,correlation coefficient,and reconstruction time.Experimental results showed that our proposed DL-CS approach can significantly enhance signal quality compared with other compressed sampling methods.This is evidenced by a percentage root-mean-square difference,correlation coefficient,and reconstruction time results of 5%,0.9988,and 0.0108 s,respectively,obtained for sampling rates below 10%for the neutron ToF signal generated using an electron-beam-driven photoneutron source.The results showed that the proposed DL-CS approach significantly improves the signal quality compared with other compressed sampling methods,exhibiting excellent reconstruction accuracy and speed.展开更多
Large dynamic range and ultra-wideband receiving abilities are significant for many receivers. With these abilities, receivers can obtain signals with different power in ultra-wideband frequency space without informat...Large dynamic range and ultra-wideband receiving abilities are significant for many receivers. With these abilities, receivers can obtain signals with different power in ultra-wideband frequency space without information loss. However, conventional receiving scheme is hard to have large dynamic range and ultra-wideband receiving simultaneously because of the analog-to-digital converter(ADC) dynamic range and sample rate limitations. In this paper, based on the modulated sampling and unlimited sampling, a novel receiving scheme is proposed to achieve large dynamic range and ultra-wideband receiving. Focusing on the single carrier signals, the proposed scheme only uses a single self-rest ADC(SR-ADC) with low sample rate, and it achieves large dynamic range and ultra-wideband receiving simultaneously. Two receiving scenarios are considered, and they are cooperative strong signal receiving and non-cooperative strong/weak signals receiving. In the cooperative receiving scenario, an improved fast recovery method is proposed to obtain the modulated sampling output. In the non-cooperative receiving scenario, the strong and weak signals with different carrier frequencies are considered, and the signal processing method can recover and estimate each signal. Simulation results show that the proposed scheme can realize large dynamic range and ultra-wideband receiving simultaneously when the input signal-to-noise(SNR) ratio is high.展开更多
With the widespread use of blockchain technology for smart contracts and decentralized applications on the Ethereum platform, the blockchain has become a cornerstone of trust in the modern financial system. However, i...With the widespread use of blockchain technology for smart contracts and decentralized applications on the Ethereum platform, the blockchain has become a cornerstone of trust in the modern financial system. However, its anonymity has provided new ways for Ponzi schemes to commit fraud, posing significant risks to investors. Current research still has some limitations, for example, Ponzi schemes are difficult to detect in the early stages of smart contract deployment, and data imbalance is not considered. In addition, there is room for improving the detection accuracy. To address the above issues, this paper proposes LT-SPSD (LSTM-Transformer smart Ponzi schemes detection), which is a Ponzi scheme detection method that combines Long Short-Term Memory (LSTM) and Transformer considering the time-series transaction information of smart contracts as well as the global information. Based on the verified smart contract addresses, account features, and code features are extracted to construct a feature dataset, and the SMOTE-Tomek algorithm is used to deal with the imbalanced data classification problem. By comparing our method with the other four typical detection methods in the experiment, the LT-SPSD method shows significant performance improvement in precision, recall, and F1-score. The results of the experiment confirm the efficacy of the model, which has some application value in Ethereum Ponzi scheme smart contract detection.展开更多
The existence and uniqueness of the maximum likelihood estimator(MLE)of parameter for the exponential-Poisson distribution is discussed by Ku s[2007.A new lifetime distribution.Computational Statistics and Data Analys...The existence and uniqueness of the maximum likelihood estimator(MLE)of parameter for the exponential-Poisson distribution is discussed by Ku s[2007.A new lifetime distribution.Computational Statistics and Data Analysis 51(9):4497-4509]in simple random sampling(SRS).As an alternative to the MLEs in SRS,Joukar et al.[2021.Parameter estimation for the exponential-poisson distribution based on ranked set samples.Communication in Statistics-Theory and Methods 50(3):560-581]discussed the MLE of parameter for this distribution in ranked set sampling(RSS).However,they did not discuss the existence and uniqueness of the MLE in RSS and did not provide explicit expressions for the Fisher information in RSS.In this article,we discuss the existence and uniqueness of the MLE of parameter in RSS and give explicit expressions for the Fisher information in RSS.The MLEs will be compared in terms of asymptotic efficiencies.Numerical studies and a real data application show that these MLEs in RSS can be real competitors for those in SRS.展开更多
Landslide susceptibility mapping(LSM) is crucial for reducing disaster risk in complex mountainous regions. This study evaluated the impact of various sampling methods on the accuracy of LSM over the next decade in Bi...Landslide susceptibility mapping(LSM) is crucial for reducing disaster risk in complex mountainous regions. This study evaluated the impact of various sampling methods on the accuracy of LSM over the next decade in Bijie City, Guizhou Province, China. Datasets were collected from 614 landslides and 500 non-landslides, and four sampling methods were proposed. Recurrent Neural Network(RNN), Gated Recurrent Unit(GRU), K-Nearest Neighbor(KNN), and Extreme Gradient Boosting(XGB) models were assessed utilising 15 metrics(Elevation, Slope, Aspect, Plan curvature, Profile curvature, Stream Power Index, Sediment Transport Index, Vector Ruggedness Measurement, Topographic Roughness Index, Lithology, Land use, Normalized Difference Vegetation Index(NDVI), Rainfall, Distance from Road, Distance from River). The results demonstrated that the GRU model combined with a 5-m sample boundary from the interior of the landslide and non-landslide areas exhibited superior performance with F1, Accuracy, and Area Under Curve(AUC) scores of 0.9700, 0.9450, and 0.8925, respectively. LSM will be projected for the next decade by coupling the Geophysical Fluid Dynamics Laboratory Earth System Model version 4(GFDLESM4) with the Shared Socioeconomic Pathway(SSP119). This study provides valuable insights into landslide risk management in landslide-prone areas.展开更多
To realize dynamic statistical publishing and protection of location-based data privacy,this paper proposes a differential privacy publishing algorithm based on adaptive sampling and grid clustering and adjustment.The...To realize dynamic statistical publishing and protection of location-based data privacy,this paper proposes a differential privacy publishing algorithm based on adaptive sampling and grid clustering and adjustment.The PID control strategy is combined with the difference in data variation to realize the dynamic adjustment of the data publishing intervals.The spatial-temporal correlations of the adjacent snapshots are utilized to design the grid clustering and adjustment algorithm,which facilitates saving the execution time of the publishing process.The budget distribution and budget absorption strategies are improved to form the sliding window-based differential privacy statistical publishing algorithm,which realizes continuous statistical publishing and privacy protection and improves the accuracy of published data.Experiments and analysis on large datasets of actual locations show that the privacy protection algorithm proposed in this paper is superior to other existing algorithms in terms of the accuracy of adaptive sampling time,the availability of published data,and the execution efficiency of data publishing methods.展开更多
The small punch test technique facilitates the convenient acquisition of the mechanical properties of in-service equipment materials and the assessment of their remaining service life through sampling.However,the weld...The small punch test technique facilitates the convenient acquisition of the mechanical properties of in-service equipment materials and the assessment of their remaining service life through sampling.However,the weldability of components with thin walls after small punch sampling,such as ethylene cracking furnace tubes,requires further investigation.Therefore,the weldability of in-service ethylene cracking furnace tubes following small punch sampling was investigated through nondestructive testing,microstructural characterization,and mechanical testing.Additionally,the impact of small punch sampling size and residual stress on the creep performance of the specimens was studied using an improved ductility exhaustion model.The results indicate that both the surface and interior of the weld repair areas on new furnace tubes and service-exposed furnace tubes after small-punch sampling are defect-free,exhibiting good weld quality.The strength of the specimens after weld repair was higher than that before sampling,whereas toughness decreased.Weld repair following small punch sampling of furnace tubes is both feasible and necessary.Furthermore,a linear relationship was observed between specimen thickness,diameter,and creep fracture time.The residual stress of welding affects the creep performance of the specimen under different stresses.展开更多
Bilingual lexicon induction focuses on learning word translation pairs,also known as bitexts,from monolingual corpora by establishing a mapping between the source and target embedding spaces.Despite recent advancement...Bilingual lexicon induction focuses on learning word translation pairs,also known as bitexts,from monolingual corpora by establishing a mapping between the source and target embedding spaces.Despite recent advancements,bilingual lexicon induction is limited to inducing bitexts consisting of individual words,lacking the ability to handle semantics-rich phrases.To bridge this gap and support downstream cross-lingual tasks,it is practical to develop a method for bilingual phrase induction that extracts bilingual phrase pairs from monolingual corpora without relying on cross-lingual knowledge.In this paper,the authors propose a novel phrase embedding training method based on the skip-gram structure.Specifically,a local hard negative sampling strategy that utilises negative samples of central tokens in sliding windows to enhance phrase embedding learning is introduced.The proposed method achieves competitive or superior performance compared to baseline approaches,with exceptional results recorded for distant languages.Additionally,we develop a phrase representation learning method that leverages multilingual pre-trained language models.These mPLMs-based representations can be combined with the above-mentioned static phrase embeddings to further improve the accuracy of the bilingual phrase induction task.We manually construct a dataset of bilingual phrase pairs and integrate it with MUSE to facilitate the bilingual phrase induction task.展开更多
基金supported by ONR Grants N000140410312 and N000141010778 to CIMMS,the University of Oklahomaby the radar data assimilation projects No. 2008LASW-A01 and No.GYHY200806003 at the Institute of Atmospheric Physics,Chinese Academy of SciencesProvided to CIMMS by NOAA/Office of Oceanic and Atmospheric Research under NOAA-University of Oklahoma Coopera-tive Agreement #NA17RJ1227,U.S. Department of Commerce
文摘By sampling perturbed state vectors from each ensemble prediction run at properly selected time levels in the vicinity of the analysis time, the recently proposed time-expanded sampling approach can enlarge the ensemble size without increasing the number of prediction runs and, hence, can reduce the computational cost of an ensemble-based filter. In this study, this approach is tested for the first time with real radar data from a tornadic thunderstorm. In particular, four assimilation experiments were performed to test the time-expanded sampling method against the conventional ensemble sampling method used by ensemble- based filters. In these experiments, the ensemble square-root filter （EnSRF） was used with 45 ensemble members generated by the time-expanded sampling and conventional sampling from 15 and 45 prediction runs, respectively, and quality-controlled radar data were compressed into super-observations with properly reduced spatial resolutions to improve the EnSRF performances. The results show that the time-expanded sampling approach not only can reduce the computational cost but also can improve the accuracy of the analysis, especially when the ensemble size is severely limited due to computational constraints for real-radar data assimilation. These potential merits are consistent with those previously demonstrated by assimilation experiments with simulated data.
基金Supported by the National Natural Science Foundation of China (40805044)Natural Science Foundation of Gansu Province(1010RJZA118)Fundmental Research Fund for Central Universities Science and Technology Development Program of China(lzujbky-2010-12)
文摘In the Ensemble Kalman Filter（EnKF） data assimilation-prediction system,most of the computation time is spent on the prediction runs of ensemble members.A limited or small ensemble size does reduce the computational cost,but an excessively small ensemble size usually leads to filter divergence,especially when there are model errors.In order to improve the efficiency of the EnKF data assimilation-prediction system and prevent it against filter divergence,a time-expanded sampling approach for EnKF based on the WRF（Weather Research and Forecasting） model is used to assimilate simulated sounding data.The approach samples a series of perturbed state vectors from Nb member prediction runs not only at the analysis time（as the conventional approach does） but also at equally separated time levels（time interval is △t） before and after the analysis time with M times.All the above sampled state vectors are used to construct the ensemble and compute the background covariance for the analysis,so the ensemble size is increased from Nb to Nb＋2M×Nb=（1＋2M）×Nb） without increasing the number of prediction runs（it is still Nb）.This reduces the computational cost.A series of experiments are conducted to investigate the impact of △t（the time interval of time-expanded sampling） and M（the maximum sampling times） on the analysis.The results show that if t and M are properly selected,the time-expanded sampling approach achieves the similar effect to that from the conventional approach with an ensemble size of（1＋2M）× Nb,but the number of prediction runs is greatly reduced.
基金Supported by the National Natural Science Foundation of China(42474239,41204128)China National Space Administration(Pre-research project on Civil Aerospace Technologies No.D010301)Strategic Priority Research Program of the Chinese Academy of Sciences(XDA17010303)。
文摘One of the detection objectives of the Chinese Asteroid Exploration mission is to investigate the space environment near the Main-belt Comet(MBC,Active Asteroid)311P/PANSTARRS.This paper outlines the scientific objectives,measurement targets,and measurement requirements for the proposed Gas and Ion Analyzer(GIA).The GIA is designed for in-situ mass spectrometry of neutral gases and low-energy ions,such as hydrogen,carbon,and oxygen,in the vicinity of 311P.Ion sampling techniques are essential for the GIA's Time-of-Flight(TOF)mass analysis capabilities.In this paper,we present an enhanced ion sampling technique through the development of an ion attraction model and an ion source model.The ion attraction model demonstrates that adjusting attraction grid voltage can enhance the detection efficiency of low-energy ions and mitigate the repulsive force of ions during sampling,which is influenced by the satellite's surface positive charging.The ion source model simulates the processes of gas ionization and ion multiplication.Simulation results indicate that the GIA can achieve a lower pressure limit below 10-13Pa and possess a dynamic range exceeding 10~9.These performances ensure the generation of ions with stable and consistent current,which is crucial for high-resolution and broad dynamic range mass spectrometer analysis.Preliminary testing experiments have verified GIA's capability to detect gas compositions such as H2O and N2.In-situ measurements near 311P using GIA are expected to significantly contribute to our understanding of asteroid activity mechanisms,the evolution of the atmospheric and ionized environments of main-belt comets,the interactions with solar wind,and the origin of Earth's water.
基金Supported by the National Natural Science Foundation of China(12064028)Jiangxi Provincial Natural Science Foundation(20232BAB201045).
文摘Electro-Optic Sampling(EOS)detection technique has been widely used in terahertz science and tech⁃nology,and it also can measure the field time waveform of the few-cycle laser pulse.Its frequency response and band limitation are determined directly by the electro-optic crystal and duration of the probe laser pulse.Here,we investigate the performance of the EOS with thin GaSe crystal in the measurement of the mid-infrared few-cycle la⁃ser pulse.The shift of the central frequency and change of the bandwidth induced by the EOS detection are calcu⁃lated,and then the pulse distortions induced in this detection process are discussed.It is found that this technique produces a red-shift of the central frequency and narrowing of the bandwidth.These changings decrease when the laser wavelength increases from 2μm to 10μm.This work can help to estimate the performance of the EOS de⁃tection technique in the mid-infrared band and offer a reference for the related experiment as well.
文摘Critical Height Sampling(CHS)estimates stand volume free from any model and tree form assumptions.Despite its introduction more than four decades ago,CHS has not been widely applied in the field due to perceived challenges in measurement.The objectives of this study were to compare estimated stand volume between CHS and sampling methods that used volume or taper models,the equivalence of the sampling methods,and their relative efficiency.We established 65 field plots in planted forests of two coniferous tree species.We estimated stand volume for a range of Basal Area Factors(BAFs).Results showed that CHS produced the most similar mean stand volume across BAFs and tree species with maximum differences between BAFs of 5-18m^(3)·ha^(−1).Horizontal Point Sampling(HPS)using volume models produced very large variability in mean stand volume across BAFs with the differences up to 126m^(3)·ha^(−1).However,CHS was less precise and less efficient than HPS.Furthermore,none of the sampling methods were statistically interchangeable with CHS at an allowable tolerance of≤55m^(3)·ha^(−1).About 72%of critical height measurements were below crown base indicating that critical height was more accessible to measurement than expected.Our study suggests that the consistency in the mean estimates of CHS is a major advantage when planning a forest inventory.When checking against CHS,results hint that HPS estimates might contain potential model bias.These strengths of CHS could outweigh its lower precision.Our study also implies serious implications in financial terms when choosing a sampling method.Lastly,CHS could potentially benefit forest management as an alternate option of estimating stand volume when volume or taper models are lacking or are not reliable.
基金supported by Natural Science Research Project of Anhui Educational Committee(2023AH030041)National Natural Science Foundation of China(42277136)Anhui Province Young and Middle-aged Teacher Training Action Project(DTR2023018).
文摘Selection of negative samples significantly influences landslide susceptibility assessment,especially when establishing the relationship between landslides and environmental factors in regions with complex geological conditions.Traditional sampling strategies commonly used in landslide susceptibility models can lead to a misrepresentation of the distribution of negative samples,causing a deviation from actual geological conditions.This,in turn,negatively affects the discriminative ability and generalization performance of the models.To address this issue,we propose a novel approach for selecting negative samples to enhance the quality of machine learning models.We choose the Liangshan Yi Autonomous Prefecture,located in southwestern Sichuan,China,as the case study.This area,characterized by complex terrain,frequent tectonic activities,and steep slope erosion,experiences recurrent landslides,making it an ideal setting for validating our proposed method.We calculate the contribution values of environmental factors using the relief algorithm to construct the feature space,apply the Target Space Exteriorization Sampling(TSES)method to select negative samples,calculate landslide probability values by Random Forest(RF)modeling,and then create regional landslide susceptibility maps.We evaluate the performance of the RF model optimized by the Environmental Factor Selection-based TSES(EFSTSES)method using standard performance metrics.The results indicated that the model achieved an accuracy(ACC)of 0.962,precision(PRE)of 0.961,and an area under the curve(AUC)of 0.962.These findings demonstrate that the EFSTSES-based model effectively mitigates the negative sample imbalance issue,enhances the differentiation between landslide and non-landslide samples,and reduces misclassification,particularly in geologically complex areas.These improvements offer valuable insights for disaster prevention,land use planning,and risk mitigation strategies.
基金Supported by the National Key R&D Program of China(No.2022YFD2401301)the Special Financial Fund of Spawning Ground Survey in the Bohai Sea and the Yellow Sea from the Ministry of Agriculture and Rural Affairs,China(No.125C0505)。
文摘A comprehensive fishery-independent survey generally incorporates various specialized surveys and integrates different survey objectives to maximize benefits while accounting for cost limitations.It is important to evaluate the adaptability of the comprehensive survey for different taxon to get the optimal design.However,the validity and adaptability of ichthyoplankton sampling incorporated in a comprehensive fishery-independent survey program in estimating abundance of ichthyoplankton species is little known.This study included ichthyoplankton sampling in an integrated survey and assessed the appropriateness of survey design.The Kriging interpolation based on Gaussian models was used to estimate the values at unsurveyed locations based on the original ichthyoplankton survey data in the Haizhou Bay as the“true”values.The sampling performances of the ongoing stratified random sampling(StRS),simple random sampling(SRS),cluster sampling(CS),hexagonal systematic sampling(SYS h),and regular systematic sampling(SYS r)with different sample sizes in estimating ichthyoplankton abundance were compared in relative estimation error(REE),relative bias(RB),and coefficient of variation(CV)by computer simulation.The ongoing StRS performed better than CS and SRS,but not as good as the two systematic sampling methods,and the current sample size in StRS design was insufficient to estimate ichthyoplankton abundance.The average REE values(meanREE)were significantly smaller in two systematic sampling designs than those in other three sampling designs,and the two systematic sampling designs could maintain good inter-annual stability of sampling performances.It is suggested that incorporating ichthyoplankton survey directly into stratified random fishery-independent surveys could not achieve the desired level of accuracy for survey objectives,but the accuracy can be improved by setting additional stations.The assessment framework presented in this study serves as a reference for evaluating the adaptability of integrated surveys to different objectives in other waters.
基金supported by the Natural Science Basic Research Program of Shaanxi(Program No.2024JC-YBMS-026).
文摘When dealing with imbalanced datasets,the traditional support vectormachine(SVM)tends to produce a classification hyperplane that is biased towards the majority class,which exhibits poor robustness.This paper proposes a high-performance classification algorithm specifically designed for imbalanced datasets.The proposed method first uses a biased second-order cone programming support vectormachine(B-SOCP-SVM)to identify the support vectors(SVs)and non-support vectors(NSVs)in the imbalanced data.Then,it applies the synthetic minority over-sampling technique(SV-SMOTE)to oversample the support vectors of the minority class and uses the random under-sampling technique(NSV-RUS)multiple times to undersample the non-support vectors of the majority class.Combining the above-obtained minority class data set withmultiple majority class datasets can obtainmultiple new balanced data sets.Finally,SOCP-SVM is used to classify each data set,and the final result is obtained through the integrated algorithm.Experimental results demonstrate that the proposed method performs excellently on imbalanced datasets.
基金supported by the Fundamental Research Funds of the Chinese Academy of Forestry(CAFYBB2020QB004)the National Natural Science Foundation of China(41971038,32171559,U20A2085,and U21A2005).
文摘The sap flow method is widely used to estimate forest transpiration.However,at the individual tree level it has spatiotemporal variations due to the impacts of environmental conditions and spatial relationships among trees.Therefore,an in-depth understanding of the coupling effects of these factors is important for designing sap flow measurement methods and performing accurate assessments of stand scale transpiration.This study is based on observations of sap flux density(SF_(d))of nine sample trees with different Hegyi’s competition indices(HCIs),soil moisture,and meteorological conditions in a pure plantation of Larix gmelinii var.principis-rupprechtii during the 2021 growing season(May to September).A multifactorial model of sap flow was developed and possible errors in the stand scale sap flow estimates associated with sample sizes were determined using model-based predictions of sap flow.Temporal variations are controlled by vapour pressure deficit(VPD),solar radiation(R),and soil moisture,and these relationships can be described by polynomial or saturated exponential functions.Spatial(individual)differences were influenced by the HCI,as shown by the decaying power function.A simple SF_(d)model at the individual tree level was developed to describe the synergistic influences of VPD,R,soil moisture,and HCI.The coefficient of variations(CV)of the sap flow estimates gradually stabilized when the sample size was>10;at least six sample trees were needed if the CV was within 10%.This study improves understanding of the mechanisms of spatiotemporal variations in sap flow at the individual tree level and provides a new methodology for determining the optimal sample size for sap flow measurements.
基金supported by the National Natural Science Foundation of China(No.62173158)the National Key Research and Development Program of China(No.2019YFC0119600)the Major Science and Technology Program of Hainan Province(No.ZDKJ202004).
文摘Class imbalance can substantially affect classification tasks using traditional classifiers,especially when identifying instances of minority categories.In addition to class imbalance,other challenges can also hinder accurate classification.Researchers have explored various approaches to mitigate the effects of class imbalance.However,most studies focus only on processing correlations within a single category of samples.This paper introduces an ensemble framework called Inter-and Intra-Class Overlapping Ensemble(llCOE),which incorporates two sampling methods.The first method,which is based on classification hardness undersampling,targets majority category samples by using simple samples as the foundation for classification and improving performance by focusing on samples near classification boundaries.The second method addresses the issue of overfitting minority category samples in undersampling and ensemble learning.To mitigate this,an adaptive augment hybrid sampling method is proposed,which enhances the classification boundary of samples and reduces overfitting.This paper conducts multiple experiments on 15 public datasets and concludes that the IlCOE ensemble framework outperforms other ensemble learning algorithms in classifying imbalanced data.
基金Supported by the National Natural Science Foundation of China(Grant No.12071175)。
文摘In this paper, we use sample average approximation with adaptive multiple importance sampling to explore moderate deviations for the optimal values. Utilizing the moderate deviation principle for martingale differences and an appropriate Delta method, we establish a moderate deviation principle for the optimal value. Moreover, for a functional form of stochastic programming, we obtain a functional moderate deviation principle for its optimal value.
基金supported by the National Key R&D Program of China(Grant No.2022YFA1604200)National Natural Science Foundation of China(Grant No.12261131495)+1 种基金Beijing Municipal Science and Technology Commission,Adminitrative Commission of Zhongguancun Science Park(Grant No.Z231100006623006)Institute of Systems Science,Beijing Wuzi University(Grant No.BWUISS21)。
文摘Optical solitons,as self-sustaining waveforms in a nonlinear medium where dispersion and nonlinear effects are balanced,have key applications in ultrafast laser systems and optical communications.Physics-informed neural networks(PINN)provide a new way to solve the nonlinear Schrodinger equation describing the soliton evolution by fusing data-driven and physical constraints.However,the grid point sampling strategy of traditional PINN suffers from high computational complexity and unstable gradient flow,which makes it difficult to capture the physical details efficiently.In this paper,we propose a residual-based adaptive multi-distribution(RAMD)sampling method to optimize the PINN training process by dynamically constructing a multi-modal loss distribution.With a 50%reduction in the number of grid points,RAMD significantly reduces the relative error of PINN and,in particular,optimizes the solution error of the(2+1)Ginzburg–Landau equation from 4.55%to 1.98%.RAMD breaks through the lack of physical constraints in the purely data-driven model by the innovative combination of multi-modal distribution modeling and autonomous sampling control for the design of all-optical communication devices.RAMD provides a high-precision numerical simulation tool for the design of all-optical communication devices,optimization of nonlinear laser devices,and other studies.
基金supported by the National Defense Technology Foundation Program of China(No.JSJT2022209A001-3)Sichuan Science and Technology Program(No.2021JDRC0011)+1 种基金Nuclear Energy Development Research Program of China(Research on High Energy X-ray Imaging of Nuclear Fuel)Scientific Research and Innovation Team Program of Sichuan University of Science and Engineering(No.SUSE652A001).
文摘Neutron time-of-flight(ToF)measurement is a highly accurate method for obtaining the kinetic energy of a neutron by measuring its velocity,but requires precise acquisition of the neutron signal arrival time.However,the high hardware costs and data burden associated with the acquisition of neutron ToF signals pose significant challenges.Higher sampling rates increase the data volume,data processing,and storage hardware costs.Compressed sampling can address these challenges,but it faces issues regarding optimal sampling efficiency and high-quality reconstructed signals.This paper proposes a revolutionary deep learning-based compressed sampling(DL-CS)algorithm for reconstructing neutron ToF signals that outperform traditional compressed sampling methods.This approach comprises four modules:random projection,rising dimensions,initial reconstruction,and final reconstruction.Initially,the technique adaptively compresses neutron ToF signals sequentially using three convolutional layers,replacing random measurement matrices in traditional compressed sampling theory.Subsequently,the signals are reconstructed using a modified inception module,long short-term memory,and self-attention.The performance of this deep compressed sampling method was quantified using the percentage root-mean-square difference,correlation coefficient,and reconstruction time.Experimental results showed that our proposed DL-CS approach can significantly enhance signal quality compared with other compressed sampling methods.This is evidenced by a percentage root-mean-square difference,correlation coefficient,and reconstruction time results of 5%,0.9988,and 0.0108 s,respectively,obtained for sampling rates below 10%for the neutron ToF signal generated using an electron-beam-driven photoneutron source.The results showed that the proposed DL-CS approach significantly improves the signal quality compared with other compressed sampling methods,exhibiting excellent reconstruction accuracy and speed.
文摘Large dynamic range and ultra-wideband receiving abilities are significant for many receivers. With these abilities, receivers can obtain signals with different power in ultra-wideband frequency space without information loss. However, conventional receiving scheme is hard to have large dynamic range and ultra-wideband receiving simultaneously because of the analog-to-digital converter(ADC) dynamic range and sample rate limitations. In this paper, based on the modulated sampling and unlimited sampling, a novel receiving scheme is proposed to achieve large dynamic range and ultra-wideband receiving. Focusing on the single carrier signals, the proposed scheme only uses a single self-rest ADC(SR-ADC) with low sample rate, and it achieves large dynamic range and ultra-wideband receiving simultaneously. Two receiving scenarios are considered, and they are cooperative strong signal receiving and non-cooperative strong/weak signals receiving. In the cooperative receiving scenario, an improved fast recovery method is proposed to obtain the modulated sampling output. In the non-cooperative receiving scenario, the strong and weak signals with different carrier frequencies are considered, and the signal processing method can recover and estimate each signal. Simulation results show that the proposed scheme can realize large dynamic range and ultra-wideband receiving simultaneously when the input signal-to-noise(SNR) ratio is high.
基金This work was granted by Qin Xin Talents Cultivation Program(No.QXTCP C202115)Beijing Information Science and Technology University+1 种基金the Beijing Advanced Innovation Center for Future Blockchain and Privacy Computing Fund(No.GJJ-23)National Social Science Foundation,China(No.21BTQ079).
文摘With the widespread use of blockchain technology for smart contracts and decentralized applications on the Ethereum platform, the blockchain has become a cornerstone of trust in the modern financial system. However, its anonymity has provided new ways for Ponzi schemes to commit fraud, posing significant risks to investors. Current research still has some limitations, for example, Ponzi schemes are difficult to detect in the early stages of smart contract deployment, and data imbalance is not considered. In addition, there is room for improving the detection accuracy. To address the above issues, this paper proposes LT-SPSD (LSTM-Transformer smart Ponzi schemes detection), which is a Ponzi scheme detection method that combines Long Short-Term Memory (LSTM) and Transformer considering the time-series transaction information of smart contracts as well as the global information. Based on the verified smart contract addresses, account features, and code features are extracted to construct a feature dataset, and the SMOTE-Tomek algorithm is used to deal with the imbalanced data classification problem. By comparing our method with the other four typical detection methods in the experiment, the LT-SPSD method shows significant performance improvement in precision, recall, and F1-score. The results of the experiment confirm the efficacy of the model, which has some application value in Ethereum Ponzi scheme smart contract detection.
基金Supported by the National Natural Science Foundation of China(11901236,12261036)Scientific Research Fund of Hunan Provincial Education Department(21A0328)Young Core Teacher Foundation of Hunan Province([2020]43).
文摘The existence and uniqueness of the maximum likelihood estimator(MLE)of parameter for the exponential-Poisson distribution is discussed by Ku s[2007.A new lifetime distribution.Computational Statistics and Data Analysis 51(9):4497-4509]in simple random sampling(SRS).As an alternative to the MLEs in SRS,Joukar et al.[2021.Parameter estimation for the exponential-poisson distribution based on ranked set samples.Communication in Statistics-Theory and Methods 50(3):560-581]discussed the MLE of parameter for this distribution in ranked set sampling(RSS).However,they did not discuss the existence and uniqueness of the MLE in RSS and did not provide explicit expressions for the Fisher information in RSS.In this article,we discuss the existence and uniqueness of the MLE of parameter in RSS and give explicit expressions for the Fisher information in RSS.The MLEs will be compared in terms of asymptotic efficiencies.Numerical studies and a real data application show that these MLEs in RSS can be real competitors for those in SRS.
基金supported by the Technological Innovation Center of Geological Disaster Prevention and ecological Restoration in the western region of the Ministry of Natural Resources opened the fund (Grant No.TICGP2023K003)the Project of Sichuan Disaster Economic Research Center (Grant No.ZHJJ2024YJS004)。
文摘Landslide susceptibility mapping(LSM) is crucial for reducing disaster risk in complex mountainous regions. This study evaluated the impact of various sampling methods on the accuracy of LSM over the next decade in Bijie City, Guizhou Province, China. Datasets were collected from 614 landslides and 500 non-landslides, and four sampling methods were proposed. Recurrent Neural Network(RNN), Gated Recurrent Unit(GRU), K-Nearest Neighbor(KNN), and Extreme Gradient Boosting(XGB) models were assessed utilising 15 metrics(Elevation, Slope, Aspect, Plan curvature, Profile curvature, Stream Power Index, Sediment Transport Index, Vector Ruggedness Measurement, Topographic Roughness Index, Lithology, Land use, Normalized Difference Vegetation Index(NDVI), Rainfall, Distance from Road, Distance from River). The results demonstrated that the GRU model combined with a 5-m sample boundary from the interior of the landslide and non-landslide areas exhibited superior performance with F1, Accuracy, and Area Under Curve(AUC) scores of 0.9700, 0.9450, and 0.8925, respectively. LSM will be projected for the next decade by coupling the Geophysical Fluid Dynamics Laboratory Earth System Model version 4(GFDLESM4) with the Shared Socioeconomic Pathway(SSP119). This study provides valuable insights into landslide risk management in landslide-prone areas.
基金supported by National Nature Science Foundation of China(No.62361036)Nature Science Foundation of Gansu Province(No.22JR5RA279).
文摘To realize dynamic statistical publishing and protection of location-based data privacy,this paper proposes a differential privacy publishing algorithm based on adaptive sampling and grid clustering and adjustment.The PID control strategy is combined with the difference in data variation to realize the dynamic adjustment of the data publishing intervals.The spatial-temporal correlations of the adjacent snapshots are utilized to design the grid clustering and adjustment algorithm,which facilitates saving the execution time of the publishing process.The budget distribution and budget absorption strategies are improved to form the sliding window-based differential privacy statistical publishing algorithm,which realizes continuous statistical publishing and privacy protection and improves the accuracy of published data.Experiments and analysis on large datasets of actual locations show that the privacy protection algorithm proposed in this paper is superior to other existing algorithms in terms of the accuracy of adaptive sampling time,the availability of published data,and the execution efficiency of data publishing methods.
基金supports provided by the National Natural Science Foundation of China(No.52372330).
文摘The small punch test technique facilitates the convenient acquisition of the mechanical properties of in-service equipment materials and the assessment of their remaining service life through sampling.However,the weldability of components with thin walls after small punch sampling,such as ethylene cracking furnace tubes,requires further investigation.Therefore,the weldability of in-service ethylene cracking furnace tubes following small punch sampling was investigated through nondestructive testing,microstructural characterization,and mechanical testing.Additionally,the impact of small punch sampling size and residual stress on the creep performance of the specimens was studied using an improved ductility exhaustion model.The results indicate that both the surface and interior of the weld repair areas on new furnace tubes and service-exposed furnace tubes after small-punch sampling are defect-free,exhibiting good weld quality.The strength of the specimens after weld repair was higher than that before sampling,whereas toughness decreased.Weld repair following small punch sampling of furnace tubes is both feasible and necessary.Furthermore,a linear relationship was observed between specimen thickness,diameter,and creep fracture time.The residual stress of welding affects the creep performance of the specimen under different stresses.
基金National Key Research and Development Program of China,Grant/Award Number:2023YFC3305003National Natural Science Foundation of China,Grant/Award Number:62376076。
文摘Bilingual lexicon induction focuses on learning word translation pairs,also known as bitexts,from monolingual corpora by establishing a mapping between the source and target embedding spaces.Despite recent advancements,bilingual lexicon induction is limited to inducing bitexts consisting of individual words,lacking the ability to handle semantics-rich phrases.To bridge this gap and support downstream cross-lingual tasks,it is practical to develop a method for bilingual phrase induction that extracts bilingual phrase pairs from monolingual corpora without relying on cross-lingual knowledge.In this paper,the authors propose a novel phrase embedding training method based on the skip-gram structure.Specifically,a local hard negative sampling strategy that utilises negative samples of central tokens in sliding windows to enhance phrase embedding learning is introduced.The proposed method achieves competitive or superior performance compared to baseline approaches,with exceptional results recorded for distant languages.Additionally,we develop a phrase representation learning method that leverages multilingual pre-trained language models.These mPLMs-based representations can be combined with the above-mentioned static phrase embeddings to further improve the accuracy of the bilingual phrase induction task.We manually construct a dataset of bilingual phrase pairs and integrate it with MUSE to facilitate the bilingual phrase induction task.
