Search results

To understand the meaning of a sentence, humans can focus on important words in the sentence, which reflects our eyes staying on each word in different gaze time or times. Thus, some studies utilize eye-tracking values to optimize the attention mechanism in deep learning models. But these studies lack to explain the rationality of this approach. Whether the attention mechanism possesses this feature of human reading needs to be explored.

The authors conducted experiments on a sentiment classification task. Firstly, they obtained eye-tracking values from two open-source eye-tracking corpora to describe the feature of human reading. Then, the machine attention values of each sentence were learned from a sentiment classification model. Finally, a comparison was conducted to analyze machine attention values and eye-tracking values.

Through experiments, the authors found the attention mechanism can focus on important words, such as adjectives, adverbs and sentiment words, which are valuable for judging the sentiment of sentences on the sentiment classification task. It possesses the feature of human reading, focusing on important words in sentences when reading. Due to the insufficient learning of the attention mechanism, some words are wrongly focused. The eye-tracking values can help the attention mechanism correct this error and improve the model performance.

Our research not only provides a reasonable explanation for the study of using eye-tracking values to optimize the attention mechanism but also provides new inspiration for the interpretability of attention mechanism.

The purpose of this paper is to classify Chinese word semantic relations, which are synonyms, antonyms, hyponyms and meronymys.

Basically, four simple methods are applied, ontology-based, dictionary-based, pattern-based and morpho-syntactic method. The authors make good use of search engine to build lexical and semantic resources for dictionary-based and pattern-based methods. To improve classification performance with more external resources, they also classify the given word pairs in Chinese and in English at the same time by using machine translation.

Experimental results show that the approach achieved an average F1 score of 50.87 per cent, an average accuracy of 70.36 per cent and an average recall of 40.05 per cent over all classification tasks. Synonym and antonym classification achieved high accuracy, i.e. above 90 per cent. Moreover, dictionary-based and pattern-based approaches work effectively on final data set.

For many natural language processing (NLP) tasks, the step of distinguishing word semantic relation can help to improve system performance, such as information extraction and knowledge graph generation. Currently, common methods for this task rely on large corpora for training or dictionaries and thesauri for inference, where limitation lies in freely data access and keeping built lexical resources up-date. This paper builds a primary system for classifying Chinese word semantic relations by seeking new ways to obtain the external resources efficiently.

The purpose of this study is to discuss the principles and factors that influence the site selection of emergency medical facilities for public health emergencies. This paper discusses the selection of the best facilities from the available facilities, proposes the capacity of new facilities, presents a logistic regression model and establishes a site selection model for emergency medical facilities for public health emergencies in megacities.

Using Guangzhou City as the research object, seven alternative facility points and the points' capacities were preset. Nine demand points were determined, and two facility locations were selected using genetic algorithms (GAs) in MATLAB for programing simulation and operational analysis.

Comparing the results of the improved GA, the results show that the improved model has fewer evolutionary generations and a faster operation speed, and that the model outperforms the traditional P-center model. The GA provides a theoretical foundation for determining the construction location of emergency medical facilities in megacities in the event of a public health emergency.

First, in this case study, there is no scientific assessment of the establishment of the capacity of the facility point, but that is a subjective method based on the assumption of the capacity of the surrounding existing hospitals. Second, because this is a theoretical analysis, the model developed in this study does not consider the actual driving speed and driving distance, but the speed of the unified average driving distance and the driving distance to take the average of multiple distances.

The results show that the method increases the selection space of decision-makers, provides them with stable technical support, helps them quickly determine the location of emergency medical facilities to respond to disaster relief work and provides better action plans for decision makers.

The results show that the algorithm performs well, which verifies the applicability of this model. When the solution results of the improved GA are compared, the results show that the improved model has fewer evolutionary generations, faster operation speed and better model than the intermediate model GA. This model can more successfully find the optimal location decision scheme, making that more suitable for the location problem of megacities in the case of public health emergencies.

The research findings provide a theoretical and decision-making basis for the location of government emergency medical facilities, as well as guidance for enterprises constructing emergency medical facilities.

Drawing on self-determination theory (Deci and Ryan, 2012) and Hew’s (2016) five-factor model, our study aimed to investigate the impact of two online training design factors (instructor accessibility and active learning) on learner’s self-efficacy and learning outcome amid the COVID-19 pandemic.

A randomized pretest-posttest control group field experiment was designed to assess participants’ self-efficacy and learning outcome of an online training program - territory business management plan, under three conditions: instructor accessibility, active learning, and controlled. Participants (N=87) were medical sales representatives from a Fortune-Global 500 pharmaceutical company’s subsidiary office in China. Data was analyzed with 2 (time) x 3 (group) MANOVA with time (pretest and posttest) as a repeated measure to investigate differences in changes in self-efficacy and learning outcome between three groups.

Overall, participants’ self-efficacy and learning outcome were significantly improved via the online training program for all three groups. Specifically, the impact of the training on learning outcome was the strongest for the active learning group, less strong for the instructor accessibility group, and the least strong for the control group.

Our research contributes towards understandings of the effectiveness of online talent training programs by examining two critical instructional design factors during a time of crisis. Our findings suggest that active learning (interactions with the training materials by purposeful self-reflection) might be a stronger predictor for increasing learning outcome than instructor accessibility (receiving feedback and tutoring sessions from the instructor) for online training programs.

Different external supply scenarios faced by the retailers will affect their choice of strategy when supply is disrupted and becomes far less than demand, urgently. This study focuses on analyzing both demand and supply side response strategies to meet customer demand and reduce the impact of the shortage during supply disruptions.

According to the quantity of products that the external market can provide, the external supply scenarios were divided into sufficient-type external supply and learning-type external supply. A two-echelon perishable goods supply chain was analyzed, and three kinds of contingency strategy models for downstream retailers were investigated. First, in the sufficient external supply scenario, the optimal price and transshipment quantity to maximize retailer's profits is discussed. Second, in the scenario of learning-type external supply, this study analyzes the optimal decision in three mechanisms of the hybrid strategy and their application: price priority mechanism, quantity priority mechanism and price–quantity balance mechanism. Furthermore, the influence of penalty cost and supply on the priority orders of different mechanisms was studied.

Results show that comparing the two pure strategies (pricing strategy and transshipment strategy)it was noted that the hybrid strategy produces the best results in sufficient-type external supply scenario. In the learning-type external supply scenario, a numerical study has shown the existence of three areas in case of penalty cost and supplier's capacity, and each areas has different priority orders of the three mechanisms. Under the situation of learning external supply, the retailer's optimal strategy is affected by parameters such as penalty cost and supply volume.

The main innovation of the work lies in the following: First; the external supply situation was divided into sufficiency type and learning type, which improves the external situation faced by retailers after the outbreak of emergencies, helps retailers understand the external situation, conforms to the actual situation and has certain practical application value. Second; in the context of learning external supply, there are three coping strategies for retailers, including: Price priority mechanism, Quantity priority mechanism and Pricing and transshipment balance mechanism. This will help retailers make strategic choices, make more scientific management decisions and improve the supply chain emergency management theory. Third; the demand side response was managed through the change of external supply during supply side recovery period and supply disruption. The proposed model enables managing and analyzing supply disruption efficiently and effectively via handling uncertainty by considering all aspects of decision-making process. The proposed model can be applied in various fields such as vegetable and fruit, fresh food, etc.

This paper seeks to answer two questions: (1) where do a country's entrepreneurship policies come from? (2) How do they evolve and shape entrepreneurial activities?

Drawing upon the comparative political economy literature and the institutional perspective, this paper proposes a theoretical model of the origin and evolution of entrepreneurship policies. We use China as a case study to apply the theoretical model and demonstrate the evolution of entrepreneurship policies in three stages during the period 1978 to 2012.

The case analysis of China provides evidence and support for our theoretical model and unpacks the process by which entrepreneurship policies originate and evolve as the result of the interplay among constantly changing policymaking, production, and knowledge regimes.

Because of the research context, findings may lack generalisability. Additional studies on policymaking and production regimes of different kinds and their respective roles in shaping entrepreneurship policies are encouraged to further advance this line of research.

This paper offers important implications concerning entrepreneurship policy and activities for policymakers, practitioners and other stakeholders in emerging economies.

Our study fills a gap in the entrepreneurship literature by expanding scholarly understanding of the origin and evolution of entrepreneurship policies.

Pervasive computing enhances the environment by embedding many computers that are gracefully integrated with human users. The purpose of this paper is to describe the creation of a smart context‐aware environment in which computation follows people and serves them everywhere. Building such smart environments is still difficult and complex due to lacking a uniform infrastructure that can adapt to diverse smart domains.

To address this problem, the paper proposes an agent‐based pluggable infrastructure which integrates a mobile agent system named pvMogent, establishes an ontology‐based context model and introduces a workflow‐based application model with the open services gateway initiative (OSGi) framework. By plugging corresponding domain context in ontology model and different applications, the infrastructure can be customized to various domains.

Through the implementation of several context‐aware applications, it was found that the infrastructure can largely reduce the development complexity as well as keep the domain extensibility by plugging corresponding domain context in ontology model.

In this paper, a number of key techniques are explored which are suitable for building context‐awareness. The experiences and lessons learned from the system development could further facilitate and inspire the research in this direction.

This paper aims to help scholars to know the frontiers in the strategic management field. On studying, it was noted that business strategic management originated from America in the 1960s and has experienced more than half a century. However, strategic management development lacks systematical summary in the twenty-first century. The scientometric method was appliedto find out the frontiers and progress of the research of strategic management in the twenty-first century, based on the literature from 2001 to 2012 in the Strategic Management Journal.

In the paper, the authors mainly used the scientometric method and applied keywords, co-occurrence method combined with multistatistical methods and mutation words analysis, author co-citation, literature co-citation and keywords co-occurrence (national).

The findings show that the strategic management research focuses on the following theories and academic thoughts: knowledge-based view, network organization research and dynamic capability are the mainstream; besides, strategy risk, the stakeholders analysis of strategy management, corporate reputation and strategic concept also attract the attention of researchers; Barney, Teece and Porter have made significant contributions to strategy management research since the twenty-first century.

The findings in the paper will help scholars in the field of strategic management to know the main frontiers of the theory, as well as the main contributors.

As a crucial institutional form established since the Chinese economic reform, the system of competitive local governments has been shaping the characteristics of China's socialist market economy to a considerable degree.

This study not only adopts the view of existing studies that attribute the economic motive of local governments to rent and consider land public finance as a means through which local governments carry out strategic investment but also attempts to further develop the view within a Marxist analytical framework.

As a result, the local governments have helped to maintain an incredibly high investment rate over a considerable period of time, facilitating the continuous, rapid growth of the Chinese economy.

This study concludes that China's local governments function as the productive allocator and user of rent in the strategic investment based on land public finance and thereby embed themselves in the relative surplus-value production initially arising from competition amongst enterprises, forming the dual structure of relative surplus-value production unique to China's economy.