Search results

This paper aims to study the agent learning from expert demonstration data while incorporating reinforcement learning (RL), which enables the agent to break through the limitations of expert demonstration data and reduces the dimensionality of the agent’s exploration space to speed up the training convergence rate.

Firstly, the decay weight function is set in the objective function of the agent’s training to combine both types of methods, and both RL and imitation learning (IL) are considered to guide the agent's behavior when updating the policy. Second, this study designs a coupling utilization method between the demonstration trajectory and the training experience, so that samples from both aspects can be combined during the agent’s learning process, and the utilization rate of the data and the agent’s learning speed can be improved.

The method is superior to other algorithms in terms of convergence speed and decision stability, avoiding training from scratch for reward values, and breaking through the restrictions brought by demonstration data.

The agent can adapt to dynamic scenes through exploration and trial-and-error mechanisms based on the experience of demonstrating trajectories. The demonstration data set used in IL and the experience samples obtained in the process of RL are coupled and used to improve the data utilization efficiency and the generalization ability of the agent.

Due to the miniaturization of electronic devices, the increased current density through solder joints leads to the occurrence of electromigration failure, thereby reducing the reliability of electronic devices. The purpose of this study is to propose a finite element-artificial neural network method for the prediction of temperature and current density of solder joints, and thus provide reference information for the reliability evaluation of solder joints.

The temperature distribution and current density distribution of the interconnect structure of electronic devices were investigated through finite element simulations. During the experimental process, the actual temperature of the solder joints was measured and was used to optimize the finite element model. A large amount of simulation data was obtained to analyze the neural network by varying the height of solder joints, the diameter of solder pads and the magnitude of current loads. The constructed neural network was trained, tested and optimized using this data.

Based on the finite element simulation results, the current is more concentrated in the corners of the solder joints, generating a significant amount of Joule heating, which leads to localized temperature rise. The constructed neural network is trained, tested and optimized using the simulation results. The ANN 1, used for predicting solder joint temperature, achieves a prediction accuracy of 96.9%, while the ANN 2, used for predicting solder joint current density, achieves a prediction accuracy of 93.4%.

The proposed method can effectively improve the estimation efficiency of temperature and current density in the packaging structure. This method prevails in the field of packaging, and other factors that affect the thermal, mechanical and electrical properties of the packaging structure can be introduced into the model.

In recent years, there has been growing interest in the use of stainless steel (SS) in reinforced concrete (RC) structures due to its distinctive corrosion resistance and excellent mechanical properties. To ensure effective synergy between SS and concrete, it is necessary to develop a time-saving approach to accurately determine the ultimate bond strength τ_u between the two materials in RC structures.

Three robust machine learning (ML) models, including support vector regression (SVR), random forest (RF) and extreme gradient boosting (XGBoost), are employed to predict τ_u between ribbed SS and concrete. Model hyperparameters are fine-tuned using Bayesian optimization (BO) with 10-fold cross-validation. The interpretable techniques including partial dependence plots (PDPs) and Shapley additive explanation (SHAP) are also utilized to figure out the relationship between input features and output for the best model.

Among the three ML models, BO-XGBoost exhibits the strongest generalization and highest accuracy in estimating τ_u. According to SHAP value-based feature importance, compressive strength of concrete f_c emerges as the most prominent feature, followed by concrete cover thickness c, while the embedment length to diameter ratio l/d, and the diameter d for SS are deemed less important features. Properly increasing c and f_c can enhance τ_u between ribbed SS and concrete.

An online graphical user interface (GUI) has been developed based on BO-XGBoost to estimate τ_u. This tool can be utilized in structural design of RC structures with ribbed SS as reinforcement.

The paper aims to tackle the classification of Remote Sensing Images (RSIs), which presents a significant challenge for computer algorithms due to the inherent characteristics of clustered ground objects and noisy backgrounds. Recent research typically leverages larger volume models to achieve advanced performance. However, the operating environments of remote sensing commonly cannot provide unconstrained computational and storage resources. It requires lightweight algorithms with exceptional generalization capabilities.

This study introduces an efficient knowledge distillation (KD) method to build a lightweight yet precise convolutional neural network (CNN) classifier. This method also aims to substantially decrease the training time expenses commonly linked with traditional KD techniques. This approach entails extensive alterations to both the model training framework and the distillation process, each tailored to the unique characteristics of RSIs. In particular, this study establishes a robust ensemble teacher by independently training two CNN models using a customized, efficient training algorithm. Following this, this study modifies a KD loss function to mitigate the suppression of non-target category predictions, which are essential for capturing the inter- and intra-similarity of RSIs.

This study validated the student model, termed KD-enhanced network (KDE-Net), obtained through the KD process on three benchmark RSI data sets. The KDE-Net surpasses 42 other state-of-the-art methods in the literature published from 2020 to 2023. Compared to the top-ranked method’s performance on the challenging NWPU45 data set, KDE-Net demonstrated a noticeable 0.4% increase in overall accuracy with a significant 88% reduction in parameters. Meanwhile, this study’s reformed KD framework significantly enhances the knowledge transfer speed by at least three times.

This study illustrates that the logit-based KD technique can effectively develop lightweight CNN classifiers for RSI classification without substantial sacrifices in computation and storage costs. Compared to neural architecture search or other methods aiming to provide lightweight solutions, this study’s KDE-Net, based on the inherent characteristics of RSIs, is currently more efficient in constructing accurate yet lightweight classifiers for RSI classification.

This research aims to investigate the mediating role of organizational dehumanization in the relationships between supervisor/coworker ostracism and employee outcomes (i.e., increased physical strains, decreased work engagement, increased turnover intentions). Moreover, this research explores the moderating role of supervisor’s organizational embodiment and coworkers’ organizational embodiment in these indirect relationships.

A cross-sectional study (N = 625) surveying employees from various organizations while using online questionnaires was conducted.

Results highlighted that, when considered together, both supervisor ostracism and coworker ostracism are positively related to organizational dehumanization, which, in turn, detrimentally influences employees’ well-being (increased physical strains), attitudes (decreased work engagement) and behaviors (increased turnover intentions). Results further indicated that the indirect effects of supervisor ostracism on outcomes via organizational dehumanization were stronger when the supervisor was perceived as highly representative of the organization. However, the interactive effect between coworker ostracism and coworkers’ organizational embodiment on organizational dehumanization was not significant.

This research adds to theory by highlighting how and when supervisor and coworker ostracism relate to undesirable consequences for both employees and organizations. On top of simultaneously considering two sources of workplace ostracism (supervisor/coworkers), this research adds to extant literature by examining one underlying mechanism (i.e., organizational dehumanization) explaining their deleterious influence on outcomes. It further examines the circumstances (i.e., high organizational embodiment) in which victims of supervisor/coworker ostracism particularly rely on this experience to form organizational dehumanization perceptions.

In the era of big data, various industries are generating large amounts of text data every day. Simplifying and summarizing these data can effectively serve users and improve efficiency. Recently, zero-shot prompting in large language models (LLMs) has demonstrated remarkable performance on various language tasks. However, generating a very “concise” multi-document summary is a difficult task for it. When conciseness is specified in the zero-shot prompting, the generated multi-document summary still contains some unimportant information, even with the few-shot prompting. This paper aims to propose a LLMs prompting for multi-document summarization task.

To overcome this challenge, the authors propose chain-of-event (CoE) prompting for multi-document summarization (MDS) task. In this prompting, the authors take events as the center and propose a four-step summary reasoning process: specific event extraction; event abstraction and generalization; common event statistics; and summary generation. To further improve the performance of LLMs, the authors extend CoE prompting with the example of summary reasoning.

Summaries generated by CoE prompting are more abstractive, concise and accurate. The authors evaluate the authors’ proposed prompting on two data sets. The experimental results over ChatGLM2-6b show that the authors’ proposed CoE prompting consistently outperforms other typical promptings across all data sets.

This paper proposes CoE prompting to solve MDS tasks by the LLMs. CoE prompting can not only identify the key events but also ensure the conciseness of the summary. By this method, users can access the most relevant and important information quickly, improving their decision-making processes.

This paper aims to study almost Ricci–Yamabe soliton in the context of certain contact metric manifolds.

The paper is designed as follows: In Section 3, a complete contact metric manifold with the Reeb vector field ξ as an eigenvector of the Ricci operator admitting almost Ricci–Yamabe soliton is considered. In Section 4, a complete K-contact manifold admits gradient Ricci–Yamabe soliton is studied. Then in Section 5, gradient almost Ricci–Yamabe soliton in non-Sasakian (k, μ)-contact metric manifold is assumed. Moreover, the obtained result is verified by constructing an example.

We prove that if the metric g admits an almost (α, β)-Ricci–Yamabe soliton with α ≠ 0 and potential vector field collinear with the Reeb vector field ξ on a complete contact metric manifold with the Reeb vector field ξ as an eigenvector of the Ricci operator, then the manifold is compact Einstein Sasakian and the potential vector field is a constant multiple of the Reeb vector field ξ. For the case of complete K-contact, we found that it is isometric to unit sphere S²ⁿ⁺¹ and in the case of (k, μ)-contact metric manifold, it is flat in three-dimension and locally isometric to Eⁿ⁺¹ × Sⁿ(4) in higher dimension.

All results are novel and generalizations of previously obtained results.

The purpose of this paper is to understand if and how industrialized house building (IHB) could support productivity developments for housebuilding on project and industry levels. The take is that fragmentation of construction is one explanation for the lack of productivity growth, and that IHB could be an integrating method of overcoming horizontal and vertical fragmentation.

Singe-factor productivity measures are calculated based on data reported by IHB companies and compared to official produced and published research data. The survey covers the years 2013–2020 for IHB companies building multi-storey houses in timber. Generalization is sought through descriptive statistics by contrasting the data samples to the used means to control vertical and horizontal fragmentation formulated as three theoretical propositions.

According to the results, IHB in timber is on average more productive than conventional housebuilding at the company level, project level, in absolute and in growth terms over the eight-year period. On the company level, the labour productivity was on average 10% higher for IHB compared to general construction and positioned between general construction and general manufacturing. On the project level, IHB displayed an average cost productivity growth of 19% for an employed prefabrication degree of about 45%.

Empirical evidence is presented quantifying so far perceived advantages of IHB. By providing analysis of actual cost and project data derived from IHB companies, the article quantifies previous research that IHB is not only about prefabrication. The observed positive productivity growth in relation to the employed prefabrication degree indicates that off-site production is not a sufficient mean for reaching high productivity and productivity growth. Instead, the capabilities to integrate the operative logic of conventional housebuilding together with logic of IHB platform development and use is a probable explanation of the observed positive productivity growth.

This study examines the impact of the Association to Advance Collegiate Schools of Business (AACSB) accreditation on education quality. We discern the prospective influences of AACSB, focusing on shifts in teaching methods and content and assessment procedures.

Using a case study approach, in-depth interviews are conducted with a Japanese-accredited business school’s faculty members to understand their perceptions of the school’s education-quality issues. The data were thematically analyzed.

Respondents acknowledged that AACSB accreditation has positively influenced teaching, encouraging active learning and the case method. However, they also indicated that accreditation had a restrictive effect on assessment activities, pushing toward compliance rather than genuine learning evaluation. This dichotomy suggests a need for balancing standard adherence with the flexibility to maintain educational depth and assessment integrity.

Convenience sampling may introduce self-selection bias. Furthermore, the qualitative case study approach does not allow for statistical generalization. However, when combined with existing literature, the findings can be analytically generalized and transferred to other contexts.

We provide insights regarding AACSB accreditation’s impact on business education, encompassing shifts in teaching methods and content and faculty perceptions of assessment. This study enhances the scholarly understanding of business school accreditation and offers guidance to accredited or accreditation-seeking academic institutions.

The study examines how the three dimensions of homophily (attitude, background, and value) influence the perceived usefulness, credibility, and enjoyability of travel content and follower behavior (i.e. willingness to search for more information and intention to visit the destination and purchase the tourism product). Likewise, the study investigates how content perception influences follower behavior.

A sample of 621 Instagram users from generations Y and Z who follow at least one travel influencer and intend to travel in the next twelve months was collected through an online survey. Partial least squares structural equation modeling was adopted to examine the data.

Attitudinal homophily influences follower behavior, value homophily impacts content perception, and background homophily has a counterproductive effect. Likewise, content perceived as useful and credible induces the intention to visit and purchase the tourism product.

The generalization of the results must be performed with care, as the context of analysis is limited to a social platform and only includes Portuguese individuals.

The findings help managers better understand which homophily cues influence content perception and maximize influencer persuasion. Based on the results, they can better decide which travel influencers should endorse their tourism products.

Research on homophily has neglected the multidimensionality of the concept and its analysis in the tourism context. By using a consolidated approach to homophily, content perception, and follower behavior, this study contributes to the tourism marketing literature and expands influencer marketing research.