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.

In large-scale environments and unstructured scenarios, the accuracy and robustness of traditional light detection and ranging (LiDAR) simultaneous localization and mapping (SLAM) algorithms are reduced, and the algorithms might even be completely ineffective. To overcome these problems, this study aims to propose a 3D LiDAR SLAM method for ground-based mobile robots, which uses a 3D LiDAR fusion inertial measurement unit (IMU) to establish an environment map and realize real-time localization.

First, we use a normal distributions transform (NDT) algorithm based on a local map with a corresponding motion prediction model for point cloud registration in the front-end. Next, point cloud features are tightly coupled with IMU angle constraints, ground constraints and gravity constraints for graph-based optimization in the back-end. Subsequently, the cumulative error is reduced by adding loop closure detection.

The algorithm is tested using a public data set containing indoor and outdoor scenarios. The results confirm that the proposed algorithm has high accuracy and robustness.

To improve the accuracy and robustness of SLAM, this method proposed in the paper introduced the NDT algorithm in the front-end and designed ground constraints and gravity constraints in the back-end. The proposed method has a satisfactory performance when applied to ground-based mobile robots in complex environments experiments.

The purpose of this paper is to investigate the consequences of two strategies of coordinating the online procurement capability and the online channel management capability on competitive performance.

A research model is presented to examine the performance impacts of these two coordination strategies, namely the balancing strategy (achieving a close match relationship) and the complementing strategy (maintaining the synergy effect), and tested using firm-level data collected from 196 manufacturing firms in China. Garen's two-stage econometric technique was used to identify the impacts of two coordination strategies on competitive performance.

Our study discusses and compares two different coordination strategies of mitigating the operational tensions across processes and deploying resource configurations for improving competitive performance. Our results show that while the balancing strategy can mitigate the risks resulted, the complementing strategy does not create synergistic effects on the focal firms' competitive performance.

The results extend our understanding of the nature of B2B digital process coordination both in IS management and supply chain operations.

The purpose of this study is to investigate the impacts of surface ozone pollution on rice profit, output and variable inputs in China.

This study estimates the rice profit function using county-level rice production data and ozone monitoring data in 2014 and 2015 to capture the impact of ozone pollution on rice profit. Then, it uses dual approach to identify the impacts of ozone on the supply of rice and the demand for variable inputs. The ozone concentration data are obtained from 1,412 monitoring stations established by the National Environmental Monitoring Centre of China.

The results show that surface ozone would significantly reduce rice profits; a 1% increase in (the daily average ozone concentration from 9 am to 4 pm) leads to a 0.1% decrease in profits. In addition, ozone has a negative impact on the levels of inputs and the supply of rice, and the elasticities of rice output, fertilizer input and labour input with respect to are −0.87, −0.86 and −0.78%, respectively. These results suggest that ozone pollution affects rice production via two channels: the direct damage on rice growth and the indirect negative impact of reducing variable inputs.

This study estimates the impacts of surface ozone pollution on rice profit and output, and quantifies its influence on variable inputs in China, which provides a better understanding of farmers' adaptation behaviour.

Escalation in the number of online food ordering platforms, along with extensive junk food marketing, lucrative offers and discounts, innovation in food flavors, and doorstep delivery of food, have triggered the consumption of high-calorie and unhealthy food products which pose serious threats to the health and future well-being of individuals by making them more obese. To date, several public policy frameworks have been developed to confront obesity; however, their efficacy seems debatable. Directionally, the objective of this study is to highlight the potential influence of “digital nudging” which aims at steering individuals in desired directions, at the same time delimiting their freedom of choice. The study also establishes the effectiveness of digital nudges promoting a healthy lifestyle by steering individuals toward healthier food choices. The author strongly believes that this conceptual perusal will offer immense inputs to healthy food marketers and researchers alike in addressing the matters of obesity. Addressing the menace of obesity calls for joint efforts of the government, the public, researchers, and more specifically food product manufacturers/marketers who should incorporate healthier food options into their portfolios. E-tailers are also urged to adopt such practices in virtual markets and promote healthier food options to effectively tackle obesity.

Although a substantial body of literature investigates the determinants of audit report lag (ARL), scant empirical evidence exists on the consequences of ARL. The purpose of this paper is to examine the association between abnormally long ARL and future stock price crash risk.

This quantitative study employed a large scale (14,445 firm-year observations) of annual financials, audit and ownership information for the Chinese listed companies during 2002–2013 which were retrieved from the China Stock Market and Accounting Research database.

This study finds evidence that abnormally long ARL increases the risk of a future stock price crash. Furthermore, the study finds that this adverse consequence is more pronounced for firms with a poor internal control environment.

Recently literature started to explore the consequences of abnormal ARL such as going concern audit opinion and restatements in the subsequent periods. This paper reveals that abnormal ARL has consequences for investor wealth losses as well. This is relevant in China, where the ongoing economic growth has attracted, and will continue to attract, a growing body of domestic and international investors. Understanding what factors could expose investors to wealth losses is of paramount importance for allocating their scarce capital.

This study extends the scant literature on the consequences of ARL, and provides useful insights for the Chinese regulatory authorities when considering the appropriateness of the current filing deadline for listed firms.

The purpose of this paper is to present and develop an ontology-based approach for automatic generation of assembly sequences.

In this approach, an assembly sequence planning ontology is constructed to represent the structure and interrelationship of product geometry information and assembly process information. In the constructed ontology, certain reasoning rules are defined to describe the knowledge and experience. Based on the ontology with reasoning rules, the algorithm for automatically generating assembly sequences is designed and implemented.

The effectiveness of this approach is verified via applying it to generate the assembly sequences of a gear reducer.

The main contribution of the paper is presenting and developing an ontology-based approach for automatically generating assembly sequences. This approach can provide a feasible solution for the issue that mathematics-based assembly sequence generation approaches have great difficulty in explicitly representing assembly experience and knowledge.

Social group system and social tagging system, which leverage the psychological mechanisms underlying group influence and social tags to drive consumer behaviors, have been prevalent in the social commerce platform. However, limited studies have examined how the affordances of social group system and social tagging system influence consumers’ social shopping behavior. The purpose of this study is to examine the formation of social shopping behavior in the social commerce platform.

Combining affordance theory with dual-congruity theory, we develop a model to examine how the affordances of social group system and social tagging system influence consumers’ social shopping behavior through the underlying self-congruity and functional-congruity processes. We empirically validate the research model using a multimethod approach, including an instrument development study and a field survey study.

Our empirical findings show that social support positively influences relational identity, while it has a nonsignificant effect on social identity. Social interactivity positively influences relational identity and social identity. Furthermore, social tagging quality and social endorser credibility positively affect perceived diagnosticity and perceived serendipity. Finally, relational identity, social identity, perceived diagnosticity and perceived serendipity collectively determine consumers’ social shopping intention.

This study contributes to the theoretical understanding of social shopping in social commerce and offers practical implications for designing an effective social group system and social tagging system to boost product sales.

The paper aims to know about energy condition’s impacts on inflation comprehensively.

This paper constructs China’s energy condition index (ECI) by bringing in three variables (China’s energy price, consumption and production) based on the financial condition index.

The result of empirical analysis shows that the index can predict China’s inflation well.

China’s ECI can predict China’s inflation well.