Search results

Fine-tuning pre-trained language models (PLMs), e.g. SciBERT, generally require large numbers of annotated data to achieve state-of-the-art performance on a range of NLP tasks in the scientific domain. However, obtaining fine-tuning data for scientific NLP tasks is still challenging and expensive. In this paper, the authors propose the mix prompt tuning (MPT), which is a semi-supervised method aiming to alleviate the dependence on annotated data and improve the performance of multi-granularity academic function recognition tasks.

Specifically, the proposed method provides multi-perspective representations by combining manually designed prompt templates with automatically learned continuous prompt templates to help the given academic function recognition task take full advantage of knowledge in PLMs. Based on these prompt templates and the fine-tuned PLM, a large number of pseudo labels are assigned to the unlabelled examples. Finally, the authors further fine-tune the PLM using the pseudo training set. The authors evaluate the method on three academic function recognition tasks of different granularity including the citation function, the abstract sentence function and the keyword function, with data sets from the computer science domain and the biomedical domain.

Extensive experiments demonstrate the effectiveness of the method and statistically significant improvements against strong baselines. In particular, it achieves an average increase of 5% in Macro-F1 score compared with fine-tuning, and 6% in Macro-F1 score compared with other semi-supervised methods under low-resource settings.

In addition, MPT is a general method that can be easily applied to other low-resource scientific classification tasks.

Sustainable development is inseparable from rational and responsible use of resources and promotion of green entrepreneurship. The contemporary green development agenda encompasses climate, economic, technical, social, cultural, and political dimensions. International efforts to greening the global development are conducted by the major economies, including China as the world’s largest consumer of energy and the biggest emitter of greenhouse gases. China is aware of its environmental problems, as well as of its part of the overall responsibility for the accomplishment of the sustainable development goals. By means of the decarbonization efforts, the latter are integrated both into the national development agenda (the concept of ecological civilization) and China’s international initiatives (the greening narrative within the Belt and Road Initiative). Over the past decade, China has made a breakthrough on the way to promoting green entrepreneurship and greening of its development (better quality of air and water, renewable energy, electric vehicles, and organic farming). On the other hand, emissions remain high, agricultural land loses productivity, and freshwater resources degrade due to climate change. In conventional industries (oil, coal mining, and electric and thermal energy), decarbonization faces an array of impediments. In this chapter, the authors summarize fundamental provisions of China’s approach to building an ecological civilization and measures to reduce emissions and achieve the carbon neutrality status within the nearest decades. The analysis of obstacles to the decarbonization of the economy and possible prospects for the development of green entrepreneurship summarizes China’s practices for possible use in other countries.

The mega railway infrastructure projects are faced with complex environments and multi-level management challenges. Thus, the mega railway infrastructure project management system not only needs to focus on its composition, but also needs to consider changes and impacts of internal and external environment.

This study attempts to introduce the concept of dissipative structure from the perspective of complexity theory and constructs a positive entropy and negentropy flow index system for mega railway infrastructure project management system in order to analyze the factors of management system more deeply. The Brusselator model is used to construct the structure of the mega railway infrastructure project management system, and the entropy method is used to calculate the positive entropy and negentropy values to verify whether the management system is a dissipative structure.

A plateau railway project in China was used as an example for an empirical study, not only its own characteristics are analyzed, but also the role of constraints and facilitation of the internal and external environment. Based on the research results, several effective suggestions are put forward to improve the stability and work efficiency of mega railway infrastructure project management system.

This study demonstrates that mega railway infrastructure project management system has the characteristics of dissipative structure. It can provide theoretical support for the development of mega railway infrastructure project management system from disorderly state to orderly state.

A time-varying grey Fourier model (TVGFM(1,1,N)) is proposed for the simulation of variable amplitude seasonal fluctuation time series, as the performance of traditional grey models can't catch the time-varying trend well.

The proposed model couples Fourier series and linear time-varying terms as the grey action, to describe the characteristics of variable amplitude and seasonality. The truncated Fourier order N is preselected from the alternative order set by Nyquist-Shannon sampling theorem and the principle of simplicity, then the optimal Fourier order is determined by hold-out method to improve the robustness of the proposed model. Initial value correction and the multiple transformation are also studied to improve the precision.

The new model has a broader applicability range as a result of the new grey action, attaining higher fitting and forecasting accuracy. The numerical experiment of a generated monthly time series indicates the proposed model can accurately fit the variable amplitude seasonal sequence, in which the mean absolute percentage error (MAPE) is only 0.01%, and the complex simulations based on Monte-Carlo method testify the validity of the proposed model. The results of monthly electricity consumption in China's primary industry, demonstrate the proposed model catches the time-varying trend and has good performances, where MAPE_F and MAPE_T are below 5%. Moreover, the proposed TVGFM(1,1,N) model is superior to the benchmark models, grey polynomial model (GMP(1,1,N)), grey Fourier model (GFM(1,1,N)), seasonal grey model (SGM(1,1)), seasonal ARIMA model seasonal autoregressive integrated moving average model (SARIMA) and support vector regression (SVR).

The parameter estimates and forecasting of the new proposed TVGFM are studied, and the good fitting and forecasting accuracy of time-varying amplitude seasonal fluctuation series are testified by numerical simulations and a case study.

This paper aims to provide a reference for the development of digital transformation from the perspective of manufacturing process links.

This paper applies canonical correlation analysis based on digital technology patents in the key links of manufacturing industries (product design, procurement, product manufacturing, warehousing and transportation, and wholesale and retail) and the related indicators of economic benefits of regions in China.

(1) The degree of digitalization of manufacturing process links is significantly correlated with economic benefits. (2) The improvement of the degree of digitalization in the “product design” link, the “warehousing and transportation” link, the “product manufacturing” link and the “wholesale and retail” link has significant impacts on the economic benefits of manufacturing industry. (3) The digital degree of the “procurement” link has no obvious influence on the economic benefits of manufacturing industry.

The research results can provide reference for the formulation and implementation of micro policies. The strategy of improving the level of digital transformation of key links of manufacturing industry is put forward to better promote both the digital transformation of manufacturing industry and economic development.

This paper innovatively studies the relationship between digitalization of manufacturing process links and economic benefits. The findings can provide theoretical and empirical support for the digital transformation of China's manufacturing industry and high-quality development of economy.

Coordinating low-carbonization and digitalization is a practical implementation pathway to achieve high-quality economic development. Regions are under great emission reduction pressure to achieve low-carbon development. However, why and how regional emission reduction pressure influences enterprise digital transformation is lacking in the literature. This study empirically tests the impact of emission reduction pressure on enterprise digital transformation and its mechanism.

This article takes the data of non-financial listed companies from 2011 to 2020 as a sample. The digital transformation index is measured by entropy value method. The bidirectional fixed effect model was used to test the hypothesis.

The research results show that emission reduction pressure forces enterprise digital transformation. The mechanism lies in that emission reduction pressure improves digital transformation by promoting enterprise innovation, and digital economy moderates the nexus between emission reduction pressure and digital transformation. Furthermore, the effect of emission reduction pressure on digital transformation is more significant for non-state-owned, mature and high-tech enterprises.

This paper discusses the mediating role of enterprise innovation between carbon emission reduction pressure and enterprise digital transformation, as well as the moderating role of digital economy. The research expands the body of knowledge about dual carbon targets, digitization and technological innovation. The author’s findings help update the impact of regional digital economy development on enterprise digital transformation. It also provides theoretical guidance for the realization of digital transformation by enterprise innovation.

With the development of intelligent technology, deep learning has made significant progress and has been widely used in various fields. Deep learning is data-driven, and its training process requires a large amount of data to improve model performance. However, labeled data is expensive and not readily available.

To address the above problem, researchers have integrated semi-supervised and deep learning, using a limited number of labeled data and many unlabeled data to train models. In this paper, Generative Adversarial Networks (GANs) are analyzed as an entry point. Firstly, we discuss the current research on GANs in image super-resolution applications, including supervised, unsupervised, and semi-supervised learning approaches. Secondly, based on semi-supervised learning, different optimization methods are introduced as an example of image classification. Eventually, experimental comparisons and analyses of existing semi-supervised optimization methods based on GANs will be performed.

Following the analysis of the selected studies, we summarize the problems that existed during the research process and propose future research directions.

This paper reviews and analyzes research on generative adversarial networks for image super-resolution and classification from various learning approaches. The comparative analysis of experimental results on current semi-supervised GAN optimizations is performed to provide a reference for further research.

Government organizations often store large amounts of data and need to choose effective data governance service to achieve digital government. This paper aims to propose a novel multi-attribute group decision-making (MAGDM) method with multigranular uncertain linguistic variables for the selection of data governance service provider.

This paper presents a MAGDM method based on multigranular uncertain linguistic variables and minimum adjustment consensus. First, a novel transformation function is proposed to unify the multigranular uncertain linguistic variables. Then, the weights of the criteria are determined by building a linear programming model with positive and negative ideal solutions. To obtain the consensus opinion, a minimum adjustment consensus model with multigranular uncertain linguistic variables is established. Furthermore, the consensus opinion is aggregated to obtain the best data governance service provider. Finally, the proposed method is demonstrated by the application of the selection of data governance service provider.

The proposed consensus model with minimum adjustments could facilitate the consensus building and obtain a higher group consensus, while traditional consensus methods often need multiple rounds of modifications. Due to different backgrounds and professional fields, decision-makers (DMs) often provide multigranular uncertain linguistic variables. The proposed transformation function based on the positive ideal solution could help DMs understand each other and facilitate the interactions among DMs.

The minimum adjustment consensus-based MAGDM method with multigranular uncertain linguistic variables is proposed to achieve the group consensus. The application of the proposed method in the selection of data governance service provider is also investigated.

The Internet of Things (IoT) platform empowers the digital transformation of the manufacturing industry by providing information technology services. Simultaneously, it enters the market by offering smart products to consumers. In light of different service fee scenarios, this article explores the optimal decision-making for the platform. It investigates the pricing models and entry decisions of IoT platforms.

In this study, we have formulated a game-theoretic model to scrutinize the influence of the IoT platform ventured into the smart device market on the pre-existing suppliers operating under subscription-based and usage-based pricing agreements.

Our outcome shows that introducing an IoT platform’s smart device has a differential effect on manufacturers depending on their contract type. Notably, our research indicates that introducing the platform’s own smart device within the subscription-based model does not negatively impact the profitability of incumbent manufacturers, so long as there is a noticeable discrepancy in the quality of the smart devices. However, our findings within the usage-based model demonstrate that despite the variance in smart device quality differentiation, the platform’s resolution to launch their device and impose their pricing agreements adversely affects established manufacturers. Additionally, we obtain valuable Intel regarding the platform’s entry strategies and contractual inclinations. We demonstrate that the platform is incentivized to present its smart device when reasonable entry costs remain. Furthermore, the platform prefers subscription-based contracts when the subscription fee is relatively high in non-platform entry and entry cases.

These findings hold significant practical implications for firms operating in an IoT-based supply chain.

Stock prices are subject to the influence of news and social media, and a discernible co-movement pattern exists among multiple stocks. Using a knowledge graph to represent news semantics and establish connections between stocks is deemed essential and viable.

This study presents a knowledge-driven framework for predicting stock prices. The framework integrates relevant stocks with the semantic and emotional characteristics of textual data. The authors construct a stock knowledge graph (SKG) to extract pertinent stock information and use a knowledge graph representation model to capture both the relevant stock features and the semantic features of news articles. Additionally, the authors consider the emotional characteristics of news and investor comments, drawing insights from behavioral finance theory. The authors examined the effectiveness of these features using the combined deep learning model CNN+LSTM+Attention.

Experimental results demonstrate that the knowledge-driven combined feature model exhibits significantly improved predictive accuracy compared to single-feature models.

The study highlights the value of the SKG in uncovering potential correlations among stocks. Moreover, the knowledge-driven multi-feature fusion stock forecasting model enhances the prediction of stock trends for well-known enterprises, providing valuable guidance for investor decision-making.