A modified stochastic model for LS+AR hybrid method and its application in polar motion short-term prediction

Short-term(up to 30 days)predictions of Earth Rotation Parameters(ERPs)such as Polar Motion(PM:PMX and PMY)play an essential role in real-time applications related to high-precision reference frame conversion.Currently,least squares(LS)+auto-regressive(AR)hybrid method is one of the main techniques of PM prediction.Besides,the weighted LS+AR hybrid method performs well for PM short-term prediction.However,the corresponding covariance information of LS fitting residuals deserves further exploration in the AR model.In this study,we have derived a modified stochastic model for the LS+AR hybrid method,namely the weighted LS+weighted AR hybrid method.By using the PM data products of IERS EOP 14 C04,the numerical results indicate that for PM short-term forecasting,the proposed weighted LS+weighted AR hybrid method shows an advantage over both the LS+AR hybrid method and the weighted LS+AR hybrid method.Compared to the mean absolute errors(MAEs)of PMX/PMY sho rt-term prediction of the LS+AR hybrid method and the weighted LS+AR hybrid method,the weighted LS+weighted AR hybrid method shows average improvements of 6.61%/12.08%and 0.24%/11.65%,respectively.Besides,for the slopes of the linear regression lines fitted to the errors of each method,the growth of the prediction error of the proposed method is slower than that of the other two methods.

Development and validation of a circulating tumor DNA-based optimization-prediction model for short-term postoperative recurrence of endometrial cancer

BACKGROUND Endometrial cancer(EC)is a common gynecological malignancy that typically requires prompt surgical intervention;however,the advantage of surgical management is limited by the high postoperative recurrence rates and adverse outcomes.Previous studies have highlighted the prognostic potential of circulating tumor DNA(ctDNA)monitoring for minimal residual disease in patients with EC.AIM To develop and validate an optimized ctDNA-based model for predicting shortterm postoperative EC recurrence.METHODS We retrospectively analyzed 294 EC patients treated surgically from 2015-2019 to devise a short-term recurrence prediction model,which was validated on 143 EC patients operated between 2020 and 2021.Prognostic factors were identified using univariate Cox,Lasso,and multivariate Cox regressions.A nomogram was created to predict the 1,1.5,and 2-year recurrence-free survival(RFS).Model performance was assessed via receiver operating characteristic(ROC),calibration,and decision curve analyses(DCA),leading to a recurrence risk stratification system.RESULTS Based on the regression analysis and the nomogram created,patients with postoperative ctDNA-negativity,postoperative carcinoembryonic antigen 125(CA125)levels of<19 U/mL,and grade G1 tumors had improved RFS after surgery.The nomogram’s efficacy for recurrence prediction was confirmed through ROC analysis,calibration curves,and DCA methods,highlighting its high accuracy and clinical utility.Furthermore,using the nomogram,the patients were successfully classified into three risk subgroups.CONCLUSION The nomogram accurately predicted RFS after EC surgery at 1,1.5,and 2 years.This model will help clinicians personalize treatments,stratify risks,and enhance clinical outcomes for patients with EC.

Slope stability prediction based on a long short-term memory neural network:comparisons with convolutional neural networks,support vector machines and random forest models

The numerical simulation and slope stability prediction are the focus of slope disaster research.Recently,machine learning models are commonly used in the slope stability prediction.However,these machine learning models have some problems,such as poor nonlinear performance,local optimum and incomplete factors feature extraction.These issues can affect the accuracy of slope stability prediction.Therefore,a deep learning algorithm called Long short-term memory(LSTM)has been innovatively proposed to predict slope stability.Taking the Ganzhou City in China as the study area,the landslide inventory and their characteristics of geotechnical parameters,slope height and slope angle are analyzed.Based on these characteristics,typical soil slopes are constructed using the Geo-Studio software.Five control factors affecting slope stability,including slope height,slope angle,internal friction angle,cohesion and volumetric weight,are selected to form different slope and construct model input variables.Then,the limit equilibrium method is used to calculate the stability coefficients of these typical soil slopes under different control factors.Each slope stability coefficient and its corresponding control factors is a slope sample.As a result,a total of 2160 training samples and 450 testing samples are constructed.These sample sets are imported into LSTM for modelling and compared with the support vector machine(SVM),random forest(RF)and convo-lutional neural network(CNN).The results show that the LSTM overcomes the problem that the commonly used machine learning models have difficulty extracting global features.Furthermore,LSTM has a better prediction performance for slope stability compared to SVM,RF and CNN models.

DPAL-BERT:A Faster and Lighter Question Answering Model

Recent advancements in natural language processing have given rise to numerous pre-training language models in question-answering systems.However,with the constant evolution of algorithms,data,and computing power,the increasing size and complexity of these models have led to increased training costs and reduced efficiency.This study aims to minimize the inference time of such models while maintaining computational performance.It also proposes a novel Distillation model for PAL-BERT(DPAL-BERT),specifically,employs knowledge distillation,using the PAL-BERT model as the teacher model to train two student models:DPAL-BERT-Bi and DPAL-BERTC.This research enhances the dataset through techniques such as masking,replacement,and n-gram sampling to optimize knowledge transfer.The experimental results showed that the distilled models greatly outperform models trained from scratch.In addition,although the distilled models exhibit a slight decrease in performance compared to PAL-BERT,they significantly reduce inference time to just 0.25%of the original.This demonstrates the effectiveness of the proposed approach in balancing model performance and efficiency.

Long Short-Term Memory Recurrent Neural Network-Based Acoustic Model Using Connectionist Temporal Classification on a Large-Scale Training Corpus

A Long Short-Term Memory(LSTM) Recurrent Neural Network(RNN) has driven tremendous improvements on an acoustic model based on Gaussian Mixture Model(GMM). However, these models based on a hybrid method require a forced aligned Hidden Markov Model(HMM) state sequence obtained from the GMM-based acoustic model. Therefore, it requires a long computation time for training both the GMM-based acoustic model and a deep learning-based acoustic model. In order to solve this problem, an acoustic model using CTC algorithm is proposed. CTC algorithm does not require the GMM-based acoustic model because it does not use the forced aligned HMM state sequence. However, previous works on a LSTM RNN-based acoustic model using CTC used a small-scale training corpus. In this paper, the LSTM RNN-based acoustic model using CTC is trained on a large-scale training corpus and its performance is evaluated. The implemented acoustic model has a performance of 6.18% and 15.01% in terms of Word Error Rate(WER) for clean speech and noisy speech, respectively. This is similar to a performance of the acoustic model based on the hybrid method.

A mathematical model of synaptotagmin 7 revealing functional importance of short-term synaptic plasticity

Synaptotagmin 7(Syt7), a presynaptic calcium sensor, has a significant role in the facilitation in shortterm synaptic plasticity: Syt7 knock out mice show a significant reduction in the facilitation. The functional importance of short-term synaptic plasticity such as facilitation is not well understood. In this study, we attempt to investigate the potential functional relationship between the short-term synaptic plasticity and postsynaptic response by developing a mathematical model that captures the responses of both wild-type and Syt7 knock-out mice. We then studied the model behaviours of wild-type and Syt7 knock-out mice in response to multiple input action potentials. These behaviors could establish functional importance of short-term plasticity in regulating the postsynaptic response and related synaptic properties. In agreement with previous modeling studies, we show that release sites are governed by non-uniform release probabilities of neurotransmitters. The structure of non-uniform release of neurotransmitters makes shortterm synaptic plasticity to act as a high-pass filter. We also propose that Syt7 may be a modulator for the long-term changes of postsynaptic response that helps to train the target frequency of the filter. We have developed a mathematical model of short-term plasticity which explains the experimental data.

A STUDY ON THE RELIABILITY OF SHORT-TERM AFB_(1)-INDUCED HEPATOCARCINOGENESIS TEST MODEL

In order to test the reliability of γ-GT foci(γ-glutamyltranspeptidase positive hepatocyticfoci) as a preneoplastic marker in AFB-inducedhepatocarcinogenesis, this experiment was car-ried out for a long period after a short-term invivo test model of AFB-induced hepatocarcino-

Short-Term and Long-Term Price Forecasting Models for the Future Exchange of Mongolian Natural Sea Buckthorn Market

Sea buckthorn market floated uncertainly within a narrow range. The market situation provided upward pressure on prices, and producer and consumer interest were poor, coupled with weak prices in the regional markets. The objectives of the study are: 1) to estimate the relationship between wild Sea buckthorn (SB) price and Supply, Demand, while some other factors of crude oil price and exchange rate by using simultaneous Supply-Demand and Price system equation and Vector Error Correction Method (VECM);2) to forecast the short-term and long-term SB price;3) to compare and evaluate the price forecasting models. Firstly, the data was analyzed by Ferris and Engle-Granger’s procedure;secondly, both price forecasting methodologies were tested by Pindyck-Rubinfeld and Makridakis’s procedure. The result shows that the VECM model is more efficient using yearly data;a short-term price forecast decreases, and a long-term price forecast is predicted to increase the Mongolian Sea buckthorn market.

Joint On-Demand Pruning and Online Distillation in Automatic Speech Recognition Language Model Optimization

Automatic speech recognition(ASR)systems have emerged as indispensable tools across a wide spectrum of applications,ranging from transcription services to voice-activated assistants.To enhance the performance of these systems,it is important to deploy efficient models capable of adapting to diverse deployment conditions.In recent years,on-demand pruning methods have obtained significant attention within the ASR domain due to their adaptability in various deployment scenarios.However,these methods often confront substantial trade-offs,particularly in terms of unstable accuracy when reducing the model size.To address challenges,this study introduces two crucial empirical findings.Firstly,it proposes the incorporation of an online distillation mechanism during on-demand pruning training,which holds the promise of maintaining more consistent accuracy levels.Secondly,it proposes the utilization of the Mogrifier long short-term memory(LSTM)language model(LM),an advanced iteration of the conventional LSTM LM,as an effective alternative for pruning targets within the ASR framework.Through rigorous experimentation on the ASR system,employing the Mogrifier LSTM LM and training it using the suggested joint on-demand pruning and online distillation method,this study provides compelling evidence.The results exhibit that the proposed methods significantly outperform a benchmark model trained solely with on-demand pruning methods.Impressively,the proposed strategic configuration successfully reduces the parameter count by approximately 39%,all the while minimizing trade-offs.

Research on Welding Quality Traceability Model of Offshore Platform Block Construction Process

Quality traceability plays an essential role in assembling and welding offshore platform blocks.The improvement of the welding quality traceability system is conducive to improving the durability of the offshore platform and the process level of the offshore industry.Currently,qualitymanagement remains in the era of primary information,and there is a lack of effective tracking and recording of welding quality data.When welding defects are encountered,it is difficult to rapidly and accurately determine the root cause of the problem from various complexities and scattered quality data.In this paper,a composite welding quality traceability model for offshore platform block construction process is proposed,it contains the quality early-warning method based on long short-term memory and quality data backtracking query optimization algorithm.By fulfilling the training of the early-warning model and the implementation of the query optimization algorithm,the quality traceability model has the ability to assist enterprises in realizing the rapid identification and positioning of quality problems.Furthermore,the model and the quality traceability algorithm are checked by cases in actual working conditions.Verification analyses suggest that the proposed early-warningmodel for welding quality and the algorithmfor optimizing backtracking requests are effective and can be applied to the actual construction process.

Short-Term Mindfulness Intervention on Adolescents’ Negative Emotion under Global Pandemic

Objective:In this research,we tried to explore how short-term mindfulness(STM)intervention affects adoles-cents’anxiety,depression,and negative and positive emotion during the COVID-19 pandemic.Design:10 classes were divided into experiment groups(5 classes;n=238)and control(5 classes;n=244)randomly.Hospital Anxi-ety and Depression Scale(HADS)and Positive and Negative Affect Schedule(PANAS)were used to measure par-ticipants’dependent variables.In the experiment group,we conducted STM practice interventions every morning in theirfirst class from March to November 2020.No interventions were conducted in the control group.Methods:Paired-sample t-tests were used to identify if a significant difference exists between every time point of the experimental and control groups.Repeated ANOVA and Growth Mixture Model(GMM)were used to analyze the tendency of positive and negative emotions,anxiety,and depression in the experimental group.Results and Conclusions:(1)With the intervention of STM,there was a significant decrease in negative emotions and an increase in positive emotions in the experimental group,whereas there were non-significant differences in the control group.(2)To explore the heterogeneity trajectories of dependent variables,we built a GMM and found there were two latent growth classes in the trajectories.(3)The results of the models showed their trajec-tories were downward,which meant that the levels of anxiety,depression,and negative emotions of participants decreased during the STM training period.Nonetheless,the score of positive affect showed upward in three loops of intervention,which indicated that the level of the participants’positive affect increased through the STM inter-vention.(4)This research indicated that STM should be given increasing consideration to enhance mental health during the worldwide outbreak of COVID-19.

Convolution Neural Network-based Load Model Parameter Selection Considering Short-term Voltage Stability

The recently proposed ambient signal-based load modeling approach offers an important and effective idea to study the time-varying and distributed characteristics of power loads.Meanwhile,it also brings new problems.Since the load model parameters of power loads can be obtained in real-time for each load bus,the numerous identified parameters make parameter application difficult.In order to obtain the parameters suitable for off-line applications,load model parameter selection(LMPS)is first introduced in this paper.Meanwhile,the convolution neural network(CNN)is adopted to achieve the selection purpose from the perspective of short-term voltage stability.To begin with,the field phasor measurement unit(PMU)data from China Southern Power Grid are obtained for load model parameter identification,and the identification results of different substations during different times indicate the necessity of LMPS.Meanwhile,the simulation case of Guangdong Power Grid shows the process of LMPS,and the results from the CNNbased LMPS confirm its effectiveness.

Short-term prognostic factors for hepatitis B virus-related acute-onchronic liver failure

BACKGROUND Acute-on-chronic liver failure(ACLF)is the abrupt exacerbation of declined hepatic function in patients with chronic liver disease.AIM To explore the independent predictors of short-term prognosis in patients with hepatitis B virus(HBV)-related ACLF and to establish a predictive short-term prognosis model for HBV-related ACLF.METHODS From January 2016 to December 2019,207 patients with HBV-related ACLF attending the 910^(th) Hospital of Chinese People's Liberation Army were continuously included in this retrospective study.Patients were stratified based on their survival status 3 mo after diagnosis.Information was collected regarding gender and age;coagulation function in terms of prothrombin time and international normalized ratio(INR);hematological profile in terms of neutrophil-tolymphocyte ratio(NLR)and platelet count(PLT);blood biochemistry in terms of alanine aminotransferase,aspartate aminotransferase,total bilirubin(Tbil),albumin,cholinesterase,blood urea nitrogen(BUN),creatinine,blood glucose,and sodium(Na);tumor markers including alpha-fetoprotein(AFP)and Golgi protein 73(GP73);virological indicators including HBV-DNA,HBsAg,HBeAg,Anti-HBe,and Anti-HBc;and complications including hepatic encephalopathy,hepatorenal syndrome,spontaneous peritonitis,gastrointestinal bleeding,and pulmonary infection.RESULTS There were 157 and 50 patients in the survival and death categories,respectively.Univariate analysis revealed significant differences in age,PLT,Tbil,BUN,NLR,HBsAg,AFP,GP73,INR,stage of liver failure,classification of liver failure,and incidence of complications(pulmonary infection,hepatic encephalopathy,spontaneous bacterial peritonitis,and upper gastrointestinal bleeding)between the two groups(P<0.05).GP73[hazard ratio(HR):1.009,95%confidence interval(CI):1.005-1.013,P=0.000],middle stage of liver failure(HR:5.056,95%CI:1.792-14.269,P=0.002),late stage of liver failure(HR:22.335,95%CI:8.544-58.388,P=0.000),pulmonary infection(HR:2.056,95%CI:1.145-3.690,P=0.016),hepatorenal syndrome(HR:6.847,95%CI:1.930-24.291,P=0.003),and HBsAg(HR:0.690,95%CI:0.524-0.908,P=0.008)were independent risk factors for short-term prognosis in patients with HBV-related ACLF.Following binary logistics regression analysis,we arrived at the following formula for predicting short-term prognosis:Logit(P)=Ln(P/1-P)=0.013×(GP73 ng/mL)+1.907×(middle stage of liver failure)+4.146×(late stage of liver failure)+0.734×(pulmonary infection)+22.320×(hepatorenal syndrome)-0.529×(HBsAg)-5.224.The predictive efficacy of the GP73-ACLF score was significantly better than that of the Model for End-Stage Liver Disease(MELD)and MELD-Na score models(P<0.05).CONCLUSION The stage of liver failure,presence of GP73,pulmonary infection,hepatorenal syndrome,and HBsAg are independent predictors of short-term prognosis in patients with HBV-related ACLF,and the GP73-ACLF model has good predictive value among these patients.

Multiproduct and multistage integrated production planning model and algorithm based on an available production capacity network

This research attempts to devise a multistage and multiproduct short-term integrative production plan that can dynamically change based on the order priority and virtual occupancy for application in steel plants. Considering factors such as the delivery time, varietal compatibility between different products, production capacity of variety per hour, minimum or maximum batch size, and transfer time, we propose an available production capacity network with varietal compatibility and virtual occupancy for enhancing production plan implementation and quick adjustment in the case of dynamic production changes. Here available means the remaining production capacity after virtual occupancy.To quickly build an available production capacity network and increase the speed of algorithm solving, constraint selection and cutting methods with order priority were used for model solving. Finally, the genetic algorithm improved with local search was used to optimize the proposed production plan and significantly reduce the order delay rate. The validity of the proposed model and algorithm was numerically verified by simulating actual production practices. The simulation results demonstrate that the model and improved algorithm result in an effective production plan.

A Recurrent Attention and Interaction Model for Pedestrian Trajectory Prediction

The movement of pedestrians involves temporal continuity,spatial interactivity,and random diversity.As a result,pedestrian trajectory prediction is rather challenging.Most existing trajectory prediction methods tend to focus on just one aspect of these challenges,ignoring the temporal information of the trajectory and making too many assumptions.In this paper,we propose a recurrent attention and interaction(RAI)model to predict pedestrian trajectories.The RAI model consists of a temporal attention module,spatial pooling module,and randomness modeling module.The temporal attention module is proposed to assign different weights to the input sequence of a target,and reduce the speed deviation of different pedestrians.The spatial pooling module is proposed to model not only the social information of neighbors in historical frames,but also the intention of neighbors in the current time.The randomness modeling module is proposed to model the uncertainty and diversity of trajectories by introducing random noise.We conduct extensive experiments on several public datasets.The results demonstrate that our method outperforms many that are state-ofthe-art.

Recent Progresses in Deep Learning Based Acoustic Models

In this paper,we summarize recent progresses made in deep learning based acoustic models and the motivation and insights behind the surveyed techniques.We first discuss models such as recurrent neural networks(RNNs) and convolutional neural networks(CNNs) that can effectively exploit variablelength contextual information,and their various combination with other models.We then describe models that are optimized end-to-end and emphasize on feature representations learned jointly with the rest of the system,the connectionist temporal classification(CTC) criterion,and the attention-based sequenceto-sequence translation model.We further illustrate robustness issues in speech recognition systems,and discuss acoustic model adaptation,speech enhancement and separation,and robust training strategies.We also cover modeling techniques that lead to more efficient decoding and discuss possible future directions in acoustic model research.

Probability Theory Predicts That Chunking into Groups of Three or Four Items Increases the Short-Term Memory Capacity

Short-term memory allows individuals to recall stimuli, such as numbers or words, for several seconds to several minutes without rehearsal. Although the capacity of short-term memory is considered to be 7 &#177?2 items, this can be increased through a process called chunking. For example, in Japan, 11-digit cellular phone numbers and 10-digit toll free numbers are chunked into three groups of three or four digits: 090-XXXX-XXXX and 0120-XXX-XXX, respectively. We use probability theory to predict that the most effective chunking involves groups of three or four items, such as in phone numbers. However, a 16-digit credit card number exceeds the capacity of short-term memory, even when chunked into groups of four digits, such as XXXX-XXXX-XXXX-XXXX. Based on these data, 16-digit credit card numbers should be sufficient for security purposes.

Forecast on Price of Agricultural Futures in China Based on ARIMA Model

The forecast on price of agricultural futures is studied in this paper. We use the ARIMA model to estimate the price trends of agricultural futures,which can help the investors to optimize their investing plans. The soybean future contracts are taken as an example to simulate the forecast based on the auto-regression coefficient(p),differential times(d) and moving average coefficient(q). The results show that ARIMA model is better to simulate and forecast the trend of closing prices of soybean futures contract,and it is applicable to forecasting the price of agricultural futures.

Chinese Named Entity Recognition with Character-Level BLSTM and Soft Attention Model

Unlike named entity recognition(NER)for English,the absence of word boundaries reduces the final accuracy for Chinese NER.To avoid accumulated error introduced by word segmentation,a deep model extracting character-level features is carefully built and becomes a basis for a new Chinese NER method,which is proposed in this paper.This method converts the raw text to a character vector sequence,extracts global text features with a bidirectional long short-term memory and extracts local text features with a soft attention model.A linear chain conditional random field is also used to label all the characters with the help of the global and local text features.Experiments based on the Microsoft Research Asia(MSRA)dataset are designed and implemented.Results show that the proposed method has good performance compared to other methods,which proves that the global and local text features extracted have a positive influence on Chinese NER.For more variety in the test domains,a resume dataset from Sina Finance is also used to prove the effectiveness of the proposed method.

Comparison of ARIMA and ANN Models Used in Electricity Price Forecasting for Power Market

In power market, electricity price forecasting provides significant information which can help the electricity market participants to prepare corresponding bidding strategies to maximize their profits. This paper introduces the models of autoregressive integrated moving average (ARIMA) and artificial neural network (ANN) which are applied to the price forecasts for up to 3 steps 8 weeks ahead in the UK electricity market. The half hourly data of historical prices are obtained from UK Reference Price Data from March 22nd to July 14th 2010 and the predictions are derived from a sliding training window with a length of 8 weeks. The ARIMA with various AR and MA orders and the ANN with different numbers of delays and neurons have been established and compared in terms of the root mean square errors (RMSEs) of price forecasts. The experimental results illustrate that the ARIMA (4,1,2) model gives greater improvement over persistence than the ANN (20 neurons, 4 delays) model.