Question-Answering Pair Matching Based on Question Classification and Ensemble Sentence Embedding

Question-answering(QA)models find answers to a given question.The necessity of automatically finding answers is increasing because it is very important and challenging from the large-scale QA data sets.In this paper,we deal with the QA pair matching approach in QA models,which finds the most relevant question and its recommended answer for a given question.Existing studies for the approach performed on the entire dataset or datasets within a category that the question writer manually specifies.In contrast,we aim to automatically find the category to which the question belongs by employing the text classification model and to find the answer corresponding to the question within the category.Due to the text classification model,we can effectively reduce the search space for finding the answers to a given question.Therefore,the proposed model improves the accuracy of the QA matching model and significantly reduces the model inference time.Furthermore,to improve the performance of finding similar sentences in each category,we present an ensemble embedding model for sentences,improving the performance compared to the individual embedding models.Using real-world QA data sets,we evaluate the performance of the proposed QA matching model.As a result,the accuracy of our final ensemble embedding model based on the text classification model is 81.18%,which outperforms the existing models by 9.81%∼14.16%point.Moreover,in terms of the model inference speed,our model is faster than the existing models by 2.61∼5.07 times due to the effective reduction of search spaces by the text classification model.

Word Embeddings and Semantic Spaces in Natural Language Processing

One of the critical hurdles, and breakthroughs, in the field of Natural Language Processing (NLP) in the last two decades has been the development of techniques for text representation that solves the so-called curse of dimensionality, a problem which plagues NLP in general given that the feature set for learning starts as a function of the size of the language in question, upwards of hundreds of thousands of terms typically. As such, much of the research and development in NLP in the last two decades has been in finding and optimizing solutions to this problem, to feature selection in NLP effectively. This paper looks at the development of these various techniques, leveraging a variety of statistical methods which rest on linguistic theories that were advanced in the middle of the last century, namely the distributional hypothesis which suggests that words that are found in similar contexts generally have similar meanings. In this survey paper we look at the development of some of the most popular of these techniques from a mathematical as well as data structure perspective, from Latent Semantic Analysis to Vector Space Models to their more modern variants which are typically referred to as word embeddings. In this review of algoriths such as Word2Vec, GloVe, ELMo and BERT, we explore the idea of semantic spaces more generally beyond applicability to NLP.

Embedding Extraction for Arabic Text Using the AraBERT Model

Nowadays,we can use the multi-task learning approach to train a machine-learning algorithm to learn multiple related tasks instead of training it to solve a single task.In this work,we propose an algorithm for estimating textual similarity scores and then use these scores in multiple tasks such as text ranking,essay grading,and question answering systems.We used several vectorization schemes to represent the Arabic texts in the SemEval2017-task3-subtask-D dataset.The used schemes include lexical-based similarity features,frequency-based features,and pre-trained model-based features.Also,we used contextual-based embedding models such as Arabic Bidirectional Encoder Representations from Transformers(AraBERT).We used the AraBERT model in two different variants.First,as a feature extractor in addition to the text vectorization schemes’features.We fed those features to various regression models to make a prediction value that represents the relevancy score between Arabic text units.Second,AraBERT is adopted as a pre-trained model,and its parameters are fine-tuned to estimate the relevancy scores between Arabic textual sentences.To evaluate the research results,we conducted several experiments to compare the use of the AraBERT model in its two variants.In terms of Mean Absolute Percentage Error(MAPE),the results showminor variance between AraBERT v0.2 as a feature extractor(21.7723)and the fine-tuned AraBERT v2(21.8211).On the other hand,AraBERT v0.2-Large as a feature extractor outperforms the finetuned AraBERT v2 model on the used data set in terms of the coefficient of determination(R2)values(0.014050,−0.032861),respectively.

Tufting Carpet Machine Information Model Based on Object Linking and Embedding for Process Control Unified Architecture

In view of the lack of research on the information model of tufting carpet machine in China,an information modeling method based on Object Linking and Embedding for Process Control Unified Architecture(OPC UA)framework was proposed to solve the problem of“information island”caused by the differentiated data interface between heterogeneous equipment and system in tufting carpet machine workshop.This paper established an information model of tufting carpet machine based on analyzing the system architecture,workshop equipment composition and information flow of the workshop,combined with the OPC UA information modeling specification.Subsequently,the OPC UA protocol is used to instantiate and map the information model,and the OPC UA server is developed.Finally,the practicability of tufting carpet machine information model under the OPC UA framework and the feasibility of realizing the information interconnection of heterogeneous devices in the tufting carpet machine digital workshop are verified.On this basis,the cloud and remote access to the underlying device data are realized.The application of this information model and information integration scheme in actual production explores and practices the application of OPC UA technology in the digital workshop of tufting carpet machine.

Heterogeneous data-driven aerodynamic modeling based on physical feature embedding

Aerodynamic surrogate modeling mostly relies only on integrated loads data obtained from simulation or experiment,while neglecting and wasting the valuable distributed physical information on the surface.To make full use of both integrated and distributed loads,a modeling paradigm,called the heterogeneous data-driven aerodynamic modeling,is presented.The essential concept is to incorporate the physical information of distributed loads as additional constraints within the end-to-end aerodynamic modeling.Towards heterogenous data,a novel and easily applicable physical feature embedding modeling framework is designed.This framework extracts lowdimensional physical features from pressure distribution and then effectively enhances the modeling of the integrated loads via feature embedding.The proposed framework can be coupled with multiple feature extraction methods,and the well-performed generalization capabilities over different airfoils are verified through a transonic case.Compared with traditional direct modeling,the proposed framework can reduce testing errors by almost 50%.Given the same prediction accuracy,it can save more than half of the training samples.Furthermore,the visualization analysis has revealed a significant correlation between the discovered low-dimensional physical features and the heterogeneous aerodynamic loads,which shows the interpretability and credibility of the superior performance offered by the proposed deep learning framework.

On modeling approach for embedded real-time software simulation testing

Modeling technology has been introduced into software testing field. However, how to carry through the testing modeling effectively is still a difficulty. Based on combination of simulation modeling technology and embedded real-time software testing method, the process of simulation testing modeling is studied first. And then, the supporting environment of simulation testing modeling is put forward. Furthermore, an approach of embedded real-time software simulation testing modeling including modeling of cross-linked equipments of system under testing （SUT）, test case, testing scheduling, and testing system service is brought forward. Finally, the formalized description and execution system of testing models are given, with which we can realize real-time, closed loop, mad automated system testing for embedded real-time software.

A p-version embedded model for simulation of concrete temperature fields with cooling pipes

Pipe cooling is an effective method of mass concrete temperature control, but its accurate and convenient numerical simulation is still a cumbersome problem. An improved embedded model, considering the water temperature variation along the pipe, was proposed for simulating the temperature field of early-age concrete structures containing cooling pipes. The improved model was verified with an engineering example. Then, the p-version self-adaption algorithm for the improved embedded model was deduced, and the initial values and boundary conditions were examined. Comparison of some numerical samples shows that the proposed model can provide satisfying precision and a higher efficiency. The analysis efficiency can be doubled at the same precision, even for a large-scale element. The p-version algorithm can fit grids of different sizes for the temperature field simulation. The convenience of the proposed algorithm lies in the possibility of locating more pipe segments in one element without the need of so regular a shape as in the explicit model.

Four-Beam Model for Vibration Analysis of a Cantilever Beam with an Embedded Horizontal Crack

As one of the main failure modes, embedded cracks occur in beam structures due to periodic loads. Hence it is useful to investigate the dynamic characteristics of a beam structure with an embedded crack for early crack detection and diagnosis. A new four-beam model with local flexibilities at crack tips is developed to investigate the transverse vibration of a cantilever beam with an embedded horizontal crack; two separate beam segments are used to model the crack region to allow opening of crack surfaces. Each beam segment is considered as an Euler-Bernoulli beam. The governing equations and the matching and boundary conditions of the four-beam model are derived using Hamilton＇s principle. The natural frequencies and mode shapes of the four-beam model are calculated using the transfer matrix method. The effects of the crack length, depth, and location on the first three natural frequencies and mode shapes of the cracked cantilever beam are investigated. A continuous wavelet transform method is used to analyze the mode shapes of the cracked cantilever beam. It is shown that sudden changes in spatial variations of the wavelet coefficients of the mode shapes can be used to identify the length and location of an embedded horizontal crack. The first three natural frequencies and mode shapes of a cantilever beam with an embedded crack from the finite element method and an experimental investigation are used to validate the proposed model. Local deformations in the vicinity of the crack tips can be described by the proposed four-beam model, which cannot be captured by previous methods.

Modeling and Analysis of Scheduling for Distributed Real-time Embedded Systems

Aimed at the deficiencies of resources based time Petri nets （RBTPN） in doing scheduling analysis for distributed real-time embedded systems, the assemblage condition of complex scheduling sequences is presented to easily compute scheduling length and simplify scheduling analysis. Based on this, a new hierarchical RBTPN model is proposed. The model introduces the definition of transition border set, and represents it as an abstract transition. The abstract transition possesses all resources of the set, and has the highest priority of each resource; the cxecution time of abstract transition is the longest time of all possible scheduling sequences. According to the characteristics and assemblage condition of RBTPN, the refinement conditions of transition border set are given, and the conditions ensure the correction of scheduling analysis. As a result, it is easy for us to understand the scheduling model and perform scheduling analysis.

A discrete model for response estimation of soil-structure systems with embedded foundations

The need for simplified physical models representing frequency dependent soil impedances has been the motivation behind many researches throughout history. Generally, such models are generated to capture impedance functions in a wide range of excitation frequencies, which leads to relatively complex models. That is while there is just a limited range of frequencies that really influence the response of the structure. Here, a new methodology based on the response-matching concept is proposed, which can lead to the development of simpler discrete models. The idea is then used to upgrade an existing simple model of surface foundations to the case of embedded foundations. The applicability of the model in both frequency domain and time domain analyses of soil-structure systems with embedded foundations is discussed. Moreover, the accuracy of the results is compared with another existing discrete model for embedded foundations.

A new method of determining the optimal embedding dimension based on nonlinear prediction

A new method is proposed to determine the optimal embedding dimension from a scalar time series in this paper. This method determines the optimal embedding dimension by optimizing the nonlinear autoregressive prediction model parameterized by the embedding dimension and the nonlinear degree. Simulation results show the effectiveness of this method. And this method is applicable to a short time series, stable to noise, computationally efficient, and without any purposely introduced parameters.

Study on a Dual Embedded Discrete Fracture Model for Fluid Flow in Fractured Porous Media

Simulation of fluid flow in the fractured porous media is very important and challenging.Researchers have developed some models for fractured porous media.With the development of related research in recent years,the prospect of embedded discrete fracture model(EDFM)is more and more bright.However,since the size of the fractures in the actual reservoir varies greatly,a very fine grid should be used which leads to a huge burden to the computing resources.To address this challenge,in the present paper,an upscaling based model is proposed.In this model,the flow in large-scale fractures is directly described by the EDFM while that in the small-scale fractures is upscaled through local simulation by EDFM.The EDFM is used to simulate the large-and small-scale fractures independently two times,so the new model is called dual embedded discrete fracture model(D-EDFM).In this paper,the detailed implementation process of D-EDFM is introduced and,through test cases,it is found the proposed model is a feasible method to simulate the flow in fractured porous media.

EFFECTIVE ELASTIC MODULI OF INHOMOGENEOUS SOLIDS BY EMBEDDED CELL MODEL

An embedded cell model is presented to obtain the effective elastic moduli for three-dimensional two-phase composites which is an exact analytic formula without any simplified approximation and can be expressed in an explicit form. For the different cells such as spherical inclusions and cracks surrounded by sphere and oblate ellipsoidal matrix, the effective elastic moduli are evaluated and the results are compared with those from various micromechanics models. These results show that the present model is direct, simple and efficient to deal with three-dimensional two-phase composites.

Computation model of proppant embedment depth based on dimensional analysis

Hydraulic fracturing is designed to form a high-conductivity fracture. The proppant will embed into the formation rock, especially in soft rock, owing to the interaction between proppant and fracture surface after fracture closure. Proppant embedment would reduce the fracture width and then lower the fracture conductivity. According to dimensional analysis, the rock is assumed to be an elastic material. Using the theory of elasticity to describe the stage of elastic deformation and analysis of the corresponding simplified embedding process, the study establish the static computation model of the two-dimensional infinite half plane and three-dimensional infinite half space model of the proppant embedment. According to laboratory results, the calculation model was modified, got an effective correction factor and analyzed the causes of errors, then discussed the factors which have impact on proppant embedment. The result calculated by the model in this paper can be reference of prop- pant optimization in on-site fracturing for a certainty degree.

A Solution Based on Modeling and Code Generation for Embedded Control System

With the development of computer technology, embedded control system plays an important role in modern industry. For the embedded system, traditional development methods are time-consuming and system is not easy to maintain. Domain-specific modeling provides a solution for the problems. In this paper, we proposed development architecture for embedded control systems based on MIC. GME is used to construct meta-model and application model, model in-terpreter interprets model and stores model information in xml format document. The final cross-platform codes are automatically generated by different templates and xml format document. This development method can reduce time and cost in the lifecycle of system development.

Software/Hardware Integrating Model of Embedded Reconfigurable Computing System for Digital Signal Processing

The system model of embedded reconfigurable system is discussed firstly, followed by the construction of the soft- ware and hardware platform. Based on the platform, the function trait and control models of digital signal processing tasks are described. According the functional unit model and control model, a software/hardware integrating model with dataflow control infrastructure is constructed. The software/hardware partition is carded out in the experiments of ultrasonic signal processing, and the results show that the system model is practicable.

An Embedded Software Modeling and Process by Using Aspect-Oriented Approach

In recent years, mobile devices have become widespread and refined, and they have offered increased convenience in human life. For these reasons, a variety of embedded systems have been designed. Therefore, improving methods for developing of embedded software systematically has become an important issue. Platform-based design is one example of an embedded-system design method that can reduce the design cost via improving a design’s abstraction level. However, platform-based design lacks precise definitions for platforms and design processes. This paper provides an approach that combines the aspects and platform-based design methods for developing embedded software. The approach is built on platform-based design methodology and uses the separating of concerns (SoC) concept to define the aspects and to reduce the crosscutting concerns in embedded system modeling. For aspect issues, we use the extended UML notation with aspects to describe both the static structure and the dynamic structure of the embedded system. We used an example of a digital photo frame system to demonstrate our approach.

基于计数模型的Word Embedding算法

Word Embedding是当今非常流行的用于文本处理任务的一种技术。基于计数模型的Word Embedding相比预测模型具有简单、快捷、易训练、善于捕捉词语相似性等优势。基于计数模型,选取2种上下文环境,运用2种权重计算方法和2种相似度计算方法,构建了5种Word Embedding模型。在词语相似性任务上比较和分析了5种Word Embedding模型,发现采用降维策略后的词表达效果要优于降维前的词表达效果;5种模型中,选取窗口上下文,PMI权重计算方法和余弦相似度计算方法的Word Embedding模型在词语相似性任务上表现最为出色。将5种模型和基于预测的Skip-gram模型进行了对比,结果表明在选取训练向量维度为100维时,基于计数的大部分模型在词语相似性任务上可以达到和Skip-gram一样甚至更好的性能。

An optimized cluster density matrix embedding theory

We propose an optimized cluster density matrix embedding theory(CDMET).It reduces the computational cost of CDMET with simpler bath states.And the result is as accurate as the original one.As a demonstration,we study the distant correlations of the Heisenberg J_(1)-J_(2)model on the square lattice.We find that the intermediate phase(0.43≤sssim J_(2)≤sssim 0.62)is divided into two parts.One part is a near-critical region(0.43≤J_(2)≤0.50).The other part is the plaquette valence bond solid(PVB)state(0.51≤J_(2)≤0.62).The spin correlations decay exponentially as a function of distance in the PVB.

Knowledge Graph Embedding Based on Adaptive Negative Sampling

Knowledge graph embedding aims at embedding entities and relations in a knowledge graph into a continuous, dense, low-dimensional and realvalued vector space. Among various embedding models appeared in recent years, translation-based models such as TransE, TransH and TransR achieve state-of-the-art performance. However, in these models, negative triples used for training phase are generated by replacing each positive entity in positive triples with negative entities from the entity set with the same probability;as a result, a large number of invalid negative triples will be generated and used in the training process. In this paper, a method named adaptive negative sampling (ANS) is proposed to generate valid negative triples. In this method, it first divided all the entities into a number of groups which consist of similar entities by some clustering algorithms such as K-Means. Then, corresponding to each positive triple, the head entity was replaced by a negative entity from the cluster in which the head entity was located and the tail entity was replaced in a similar approach. As a result, it generated a set of high-quality negative triples which benefit for improving the effectiveness of embedding models. The ANS method was combined with the TransE model and the resulted model was named as TransE-ANS. Experimental results show that TransE-ANS achieves significant improvement in the link prediction task.