Knowledge Transfer Model Research Articles

Prediction of multiaxial fatigue life with a data-driven knowledge transfer model

A data-driven model is presented for accurate prediction of multiaxial fatigue life based upon the principle of transfer learning (TL). The Tradaboost framework is explored to adjust the weights of training data from different sources, actuating information transfer from domain knowledge to the data-driven modeling of multiaxial fatigue life. Subsequently, extensive experimental results tested under the proportional and non-proportional circle loadings are collected for model evaluation. The results demonstrate that the proposed model is more accurate than domain knowledge-based, conventional data-driven, and comparable TL-based models, with a low data requirement, showcasing good applicability for multiaxial fatigue life assessment.

Dynamic collaborative learning with heterogeneous knowledge transfer for long-tailed visual recognition

Solving the long-tailed visual recognition with deep convolutional neural networks is still a challenging task. As a mainstream method, multi-experts models achieve SOTA accurary for tackling this problem, but the uncertainty in network learning and the complexity in fusion inference constrain the performance and practicality of the multi-experts models. To remedy this, we propose a novel dynamic collaborative learning with heterogeneous knowledge transfer model (DCHKT) in this paper, in which experts with different expertise collaborate to make predictions. DCHKT consists of two core components: dynamic adaptive weight adjustment and heterogeneous knowledge transfer learning. First, the dynamic adaptive weight adjustment is designed to shift the focus of model training between the global expert and domain experts via dynamic adaptive weight. By modulating the trade-off between the learning of features and classifier, the dynamic adaptive weight adjustment can enhance the discriminative ability of each expert and alleviate the uncertainty of model learning. Then, heterogeneous knowledge transfer learning, which measures the distribution differences between the fusion logits of multiple experts and the predicted logits of each expert with different specialties, can achieve message passing between experts and enhance the consistency of ensemble prediction in model training and inference to promote their collaborations. Finally, extensive experimental results on public long-tailed datasets: CIFAR-LT, ImageNet-LT, Place-LT and iNaturalist2018, demonstrate the effectiveness and superiority of our DCHKT.

Self-improved multi-view interactive knowledge transfer

Multi-view learning (MVL) is a promising data fusion technique based on the principles of consensus and complementarity. Despite significant advancements in this field, several challenges persist. First, scalability remains an issue, as many existing approaches are limited to two-view scenarios, making them difficult to extend to more complex multi-view settings. Second, implementing consensus principles in current techniques often requires adding extra terms to the model’s objective function or constraints, leading to increased complexity. Additionally, when applying complementarity principles, most studies focus on pairwise interactions between views, overlooking the benefits of deeper and broader multi-view interactions. To address these challenges, this paper proposes the multi-view interactive knowledge transfer (MVIKT) model, which enhances scalability by effectively managing interactions across multiple views, thereby overcoming the limitations of traditional two-view models. More importantly, MVIKT introduces a novel interactive knowledge transfer strategy that simplifies the application of the consensus principle by eliminating the need for additional terms. By treating margin distances as transferable knowledge and facilitating multiple rounds of interaction, MVIKT uncovers deeper complementary information, thereby improving the overall effectiveness of MVL. Theoretical analysis further supports the MVIKT model, demonstrating that transferring knowledge through margin distance is capable of lowering the upper bound of the generalization error. Extensive experiments across diverse datasets validate MVIKT’s superiority, showing statistically significant improvements over benchmark methods.

Designing a knowledge transfer model for accelerating urban climate adaptation in Malaysia

The urgency of adapting urban areas to the increasing impacts of climate change has prompted the scientific community to seek new approaches in partnership with public entities and civil society organizations. In Malaysia, Penang Island has developed a nature-based urban climate adaptation program (PNBCAP) seeking to increase urban resilience, reduce urban heat and flooding, strengthening social resilience, and build institutional capacity. The project includes a strong knowledge transfer component focused on encouraging other cities in the country to develop and implement adaptation policies, projects, and initiatives. This research develops a model adopting the most efficient processes to accelerate the transfer of knowledge to promote urban adaptation based on the PNBCAP. The methodology is developed based on a review of literature focused on innovation systems and change theories. The integration of success strategies in adaptation contributes to informing the creation of solutions around the alliance of local, state, and national government agencies, scientific institutions, and civil society organizations, in a new framework designated the Malaysian Adaptation Sharing Hub (MASH). MASH is structured in 3-steps and will function as an accelerator for the implementation of urban climate adaptation policies, with the target of creating 2 new adaptation-related policies to be adopted annually by each city member, based on knowledge gathered in the PNBCAP. It is concluded that, to speed up urban adaptation, it is necessary to reinforce and promote the sharing of knowledge resulting from or associated with pilot projects.

Entrepreneurial intention among Colombian university students: A theory of planned behavior analysis in Colombia

Studies on entrepreneurship have used several approaches, among which, research on entrepreneurial intention (EI) can be highlighted. Studies on EI help researchers better understand entrepreneurial behavior. In this sense, one of the most relevant frameworks in EI studies is the Theory of Planned Behavior (TPB), which stands out for its predictive capacity, mainly in a university context. The aim of this research was to determine the relationship between the variables of Personal Attitudes (PAs), Subjective Norms (SNs), and Perceived Behavioral Control (PBC) with respect to the EI of university students in Colombia. For this purpose, we constructed a model that integrates the elements of the TPB with EI with reference to the validated instrument of the Global University Entrepreneurial Spirit Student's Survey (GUESSS) project applied to a sample of 12,383 students. The estimation of the model uses structural equations (SEMs) by means of Partial Least Squares-Path Modelling (PLS-PM). The findings show that PAs and PBC are determinants of EI, while SNs have no direct influence; however, their effects are indirect. These results have implications for how to promote entrepreneurship in emerging economies through the university environment. The results indicate that the development of students' capabilities is essential for entrepreneurship and its incorporation into the university's knowledge transfer model.

Prototype-Guided Memory Replay for Continual Learning.

Continual learning (CL) is a machine learning paradigm that accumulates knowledge while learning sequentially. The main challenge in CL is catastrophic forgetting of previously seen tasks, which occurs due to shifts in the probability distribution. To retain knowledge, existing CL models often save some past examples and revisit them while learning new tasks. As a result, the size of saved samples dramatically increases as more samples are seen. To address this issue, we introduce an efficient CL method by storing only a few samples to achieve good performance. Specifically, we propose a dynamic prototype-guided memory replay (PMR) module, where synthetic prototypes serve as knowledge representations and guide the sample selection for memory replay. This module is integrated into an online meta-learning (OML) model for efficient knowledge transfer. We conduct extensive experiments on the CL benchmark text classification datasets and examine the effect of training set order on the performance of CL models. The experimental results demonstrate the superiority our approach in terms of accuracy and efficiency.

Novel calibration design improves knowledge transfer across products for the characterization of pharmaceutical bioprocesses.

Modern machine learning has the potential to fundamentally change the way bioprocesses are developed. In particular, horizontal knowledge transfer methods, which seek to exploit data from historical processes to facilitate process development for a new product, provide an opportunity to rethink current workflows. In this work, we first assess the potential of two knowledge transfer approaches, meta learning and one-hot encoding, in combination with Gaussian process (GP) models. We compare their performance with GPs trained only on data of the new process, that is,local models. Using simulated mammalian cell culture data, we observe that both knowledge transfer approaches exhibit test set errors that are approximately halved compared to those of the local models when two, four, or eight experiments of the new product are used for training. Subsequently, we address the question whether experiments for a new product could be designed more effectively by exploiting existing knowledge. In particular, we suggest to specifically design a few runs for the novel product to calibrate knowledge transfer models, a task that we coin calibration design. We propose a customized objective function to identify a set of calibration design runs, which exploits differences in the process evolution of historical products. In two simulated case studies, we observed that training with calibration designs yields similar test set errors compared to common design of experiments approaches. However, the former requires approximately four times fewer experiments. Overall, the results suggest that process development could be significantly streamlined when systematically carrying knowledge from one product to the next.

Granular transfer learning

Transfer learning is aimed at supporting the design of machine learning models in the target domain Dt, given that the knowledge (model) has already been constructed in the source domain Ds. The domains Dt and Ds (as well as the corresponding tasks Ts and Tt) are similar, yet not identical. As a result, the model transferred from Ds to Dt in this new environment exhibits its relevance (credibility) only to some limited extent. In this study, we develop an original approach, where we advocate that the knowledge transfer (model transfer) gives rise to a granular model where the level of information granularity associated with the produced results quantifies the relevance (quality or credibility) of the transferred model. In other words, we stress that the quality of knowledge transferred to Dt becomes captured through a granular generalization of the original numeric model. The overall systematic design process is elaborated on by focusing on the development process of granular neural networks carried out on a basis of the numeric neural networks coming from Ds. The key aspect of the design is to elevate the existing numeric neural network to its granular counterpart by admitting that the connections of the developed model come in the form of information granules, in particular intervals and fuzzy sets. The optimization process is guided by adjusting (optimizing) the level of information granularity being regarded as an essential design asset. The optimized performance index builds upon the descriptors of information granules commonly encountered in Granular Computing. In particular, coverage and specificity measures are treated as sound performance indicators of the quality of knowledge transfer (viz. the performance of the granular neural network expressed in the target domain). Several illustrative examples are provided to visualize the performance of the established design environment.

The impact of knowledge characteristics on process performance: experimenting with the conversion perspective on knowledge transfer velocity

PurposeWith shorter product cycles and a growing number of knowledge-intensive business processes, time consumption is a highly relevant target factor in measuring the performance of contemporary business processes. This research aims to extend prior research on the effects of knowledge transfer velocity at the individual level by considering the effect of complexity, stickiness, competencies, and further demographic factors on knowledge-intensive business processes at the conversion-specific levels.Design/methodology/approachWe empirically assess the impact of situation-dependent knowledge transfer velocities on time consumption in teams and individuals. Further, we issue the demographic effect on this relationship. We study a sample of 178 experiments of project teams and individuals applying ordinary least squares (OLS) for regression analysis-based modeling.FindingsThe authors find that time consumed at knowledge transfers is negatively associated with the complexity of tasks. Moreover, competence among team members has a complementary effect on this relationship and stickiness retards knowledge transfers. Thus, while demographic factors urgently need to be considered for effective and speedy knowledge transfers, these influencing factors should be addressed on a conversion-specific basis so that some tasks are realized in teams best while others are not. Guidelines and interventions are derived to identify best task realization variants, so that process performance is improved by a new kind of process improvement method.Research limitations/implicationsThis study establishes empirically the importance of conversion-specific influence factors and demographic factors as drivers of high knowledge transfer velocities in teams and among individuals. The contribution connects the field of knowledge management to important streams in the wider business literature: process improvement, management of knowledge resources, design of information systems, etc. Whereas the model is highly bound to the experiment tasks, it has high explanatory power and high generalizability to other contexts.Practical implicationsTeam managers should take care to allow the optimal knowledge transfer situation within the team. This is particularly important when knowledge sharing is central, e.g. in product development and consulting processes. If this is not possible, interventions should be applied to the individual knowledge transfer situation to improve knowledge transfers among team members.Social implicationsFaster and more effective knowledge transfers improve the performance of both commercial and non-commercial organizations. As nowadays, the individual is faced with time pressure to finalize tasks, the deliberated increase of knowledge transfer velocity is a core capability to realize this goal. Quantitative knowledge transfer models result in more reliable predictions about the duration of knowledge transfers. These allow the target-oriented modification of knowledge transfer situations so that processes speed up, private firms are more competitive and public services are faster to citizens.Originality/valueTime consumption is an increasingly relevant factor in contemporary business but so far not been explored in experiments at all. This study extends current knowledge by considering quantitative effects on knowledge velocity and improved knowledge transfers.

Model for Knowledge Transfer in Agent Organizations: a Case Study on Moise+

Knowledge transfer enables the development of complex multi-agent systems featuring agents that share knowledge to execute tasks. Environments shaped by knowledge transfer involve agents assuming specific roles, reasoning about the environment and other agents, and forming organizations that transfer knowledge through well-defined plans and strategies. This paper presents an organizational model for knowledge transfer, introducing a centralizing role, named organizer, responsible for managing relations between the system, agents, and their roles. An interaction protocol is established to guide the step-by-step communication between the organizer and other agents in the system during knowledge transfer. To model knowledge transfer within agent organizations, we employ a dynamic implementation of the Moise+ organizational model, called MoiseLight, enabling the creation of organizations at runtime. We validate the proposal by adapting the model to MoiseLight, demonstrating the organizer's ability to facilitate knowledge transfer among agents following the specified interaction protocol.

Toward an Integrated Framework of Knowledge Transfer: When Non-Profit Organizations Become International

Existing literature lacks an in-depth exploration of Non-Profit Organizations (NPOs) studying transnational religious rituals and scripture dissemination, leaving a notable research gap. The study integrated the SECI model with an institutional distance framework to create a new path analysis, this research aims to reveal interorganizational knowledge transfer dynamics within such religious groups, focusing on Tiandi Teachings (TT) in Taiwan and its Branches of the Japan Diocese (BJD). Employing extended case methodology with participant observation and semi-structured interviews, the study analyzed data using grounded theory and NVivo 12 software. It identified insights into knowledge transfer processes within international religious organizations, elucidating mechanisms for establishing local legitimacy and transferring religious knowledge across borders. This research contributes to understanding how religious organizations globalize, enriches institutional theory, and underscores the importance of integrating institutional concepts with knowledge transfer models. Its findings offer practical benefits to religious leaders, policymakers, and academics.

Entropic Artificial Intelligence and Knowledge Transfer

An overview of entropy applications to illustrate a few possible uses of entropy in knowledge transmission and artificial intelligence. Artificial intelligence uses entropy as a fundamental concept in many diverse applications, including reinforcement learning, data compression, and decision-making. It assists artificial intelligence models in producing wellinformed forecasts and judgments by assessing uncertainty and information content. As such, the purpose of this work is to highlight the importance of entropy and draw the attention of the artificial intelligence research community to it as a potent tool for advancing artificial intelligence. This work also addresses the importance of knowledge transfer (KT), especially intergenerational KT (IGT), in knowledge management. Knowledge entropy (KE) is a concept that is used to measure the complexity of knowledge distribution within an organisation and evaluate the effectiveness of KT activities. Furthermore, the KT model—which is predicated on the ideas of information content and tacitness—is presented. It blends techniques for customisation and codification. A few challenging open problems are presented along with future study options

Learning to Prompt Knowledge Transfer for Open-World Continual Learning

This paper studies the problem of continual learning in an open-world scenario, referred to as Open-world Continual Learning (OwCL). OwCL is increasingly rising while it is highly challenging in two-fold: i) learning a sequence of tasks without forgetting knowns in the past, and ii) identifying unknowns (novel objects/classes) in the future. Existing OwCL methods suffer from the adaptability of task-aware boundaries between knowns and unknowns, and do not consider the mechanism of knowledge transfer. In this work, we propose Pro-KT, a novel prompt-enhanced knowledge transfer model for OwCL. Pro-KT includes two key components: (1) a prompt bank to encode and transfer both task-generic and task-specific knowledge, and (2) a task-aware open-set boundary to identify unknowns in the new tasks. Experimental results using two real-world datasets demonstrate that the proposed Pro-KT outperforms the state-of-the-art counterparts in both the detection of unknowns and the classification of knowns markedly. Code released at https://github.com/YujieLi42/Pro-KT.

Supporting the transfer of knowledge in high‐risk major accident environment

AbstractA key element in learning from accidents is the skill associated with the transfer of knowledge gained by the operator from historical incidents. These incidents can include accidents and near‐misses that occurred on site or in similar companies outside the plant. Knowledge transfer within the enterprise can be supported by a suitable framework or model that is easily understood by a wide range of people who are interested in the lessons learned from accidents. The application of some of the knowledge transfer models used so far can be quite time consuming and uncomfortable for the participants. For this reason, this paper aims to propose a simple model designed to support knowledge transfer. This model is proposed based on the widely used PDCA (plan‐do‐check‐act) framework. Its use is demonstrated by the case of a major accident that occurred in the Czech Republic. The model can be used not only for learning from major accidents that occur in the subject company but also for learning from near‐misses or events that occurred in the past in similar plants. Thus, the model can easily help in increasing the efficiency of the accident‐learning system in particular.

KDTM: Multi-Stage Knowledge Distillation Transfer Model for Long-Tailed DGA Detection

As the most commonly used attack strategy by Botnets, the Domain Generation Algorithm (DGA) has strong invisibility and variability. Using deep learning models to detect different families of DGA domain names can improve the network defense ability against hackers. However, this task faces an extremely imbalanced sample size among different DGA categories, which leads to low classification accuracy for small sample categories and even classification failure for some categories. To address this issue, we introduce the long-tailed concept and augment the data of small sample categories by transferring pre-trained knowledge. Firstly, we propose the Data Balanced Review Method (DBRM) to reduce the sample size difference between the categories, thus a relatively balanced dataset for transfer learning is generated. Secondly, we propose the Knowledge Transfer Model (KTM) to enhance the knowledge of the small sample categories. KTM uses a multi-stage transfer to transfer weights from the big sample categories to the small sample categories. Furthermore, we propose the Knowledge Distillation Transfer Model (KDTM) to relieve the catastrophic forgetting problem caused by transfer learning, which adds knowledge distillation loss based on the KTM. The experimental results show that KDTM can significantly improve the classification performance of all categories, especially the small sample categories. It can achieve a state-of-the-art macro average F1 score of 84.5%. The robustness of the KDTM model is verified using three DGA datasets that follow the Pareto distributions.

Preliminary study of knowledge transfer aspect in the creation of a framework for education and training of digital manufacturing technologies

Digitalization is changing the industry. As this change accelerates its speed, it also requires a transformation process where knowledge transfer between industry and research institutes play a significant role. There is a need to be more fluent, flexible, and efficient in order to get the latest research results into industrial implementation as quickly as possible. The challenge in knowledge transfer is that its speed in the current stage is too slow compared to the speed of development and changes required by digitalization of traditional manufacturing industries. The motivation for this study is the gap in knowledge transfer. One emerging digital transformation is the establishment of modern digital manufacturing technologies, e.g., additive manufacturing (AM). There are different approaches to supporting the industry in this transformation. Knowledge transfer can happen, for example, through education (e.g., master students) and industrial training, but also the fluent transfer of the latest research results from research institutes to companies is needed. University education needs to support the requirements of the manufacturing industry by providing future experts with skills to smooth the transformation process and bring novel technology applications, such as AM, to industrial-scale use. The article discusses how university education can support future competence-building in the industry. Similar needs are also present in industrial training, which universities often provide. Both education and training need to be improved from fundamental approaches to explain how this new knowledge should be created, i.e., how knowledge transfer happens most efficiently. The outcome of this article is the basis of the framework for education and training of digital manufacturing technologies by using modern learning methods and tools. More detailed pedagogical and knowledge transfer models can be developed and applied when this framework is created.

Reverse knowledge transfer in digital era and its effect on ambidextrous innovation: A simulation based on system dynamics

Knowledge transfer plays a prominent role in fostering innovation among multinational corporations (MNCs), however, the exploration of reverse knowledge transfer in today's digital era remains insufficient. Therefore, we explore the impact of reverse knowledge transfer on ambidextrous innovation and the process mechanism against the backdrop of digital technology. A system dynamics model of reverse knowledge transfer is constructed from three aspects: knowledge acquisition, knowledge absorption and knowledge application. Our simulation results indicate that the dual embeddedness of multinational R&D centers promotes knowledge acquisition and absorption, enhances the knowledge base of parent companies, and facilitates ambidextrous innovation. In particular, the application of digital technology further promotes the process through knowledge absorption, with the long-term impact more pronounced. This study not only enriches the literature on reverse knowledge transfer but also expands the application of digital technology in the field of MNCs' innovation.

What has been said about the reinvention of post-COVID tourism: A bibliometric analysis

Abstract The purpose of this study is to analyse the relationship between the COVID-19 impacts on tourism and the political approaches adopted to overcome the crisis in this sector. There has been a deep reflection on a new tourism paradigm, guided by an increasing commitment to sustainability and innovation, which was triggered by the COVID-19 outbreak. This perspective reinforces the role of decentralisation and autonomy of policies at the regional level. A network governance model emerges as a strategy to overcome this crisis. A bibliometric analysis (Rafael & Pires, 2021) was carried out using the SCOPUS and Web of Science (WOS) databases, from January 2020 to April 2021. The search criteria followed the assumptions of the non-probabilistic snow-ball sampling technique (Vinuto, 2014). The results were processed in SPSS and VosViewer software. Four main dimensions were found that mediate the relationship between tourism and the effects of the COVID-19 pandemic, namely: (1) policies, enabling network governance and regional autonomy; (2) research, enhancing knowledge transfer and new teaching models; (3) economic collaborative approaches that support DMO's activities; and (4) network innovation that fosters positive change. Understanding the impacts of COVID-19 in tourism politics and strategies offers important guidelines to both private and public decision-makers.

Research on the knowledge transfer mechanism of digital platform in the digital innovation ecosystem: an improved model of SIR embedded in symbiosis theory

PurposeIn the digital economy era, digital platforms are vital infrastructure for innovation subjects to perform digital innovation activities. Achieving efficient and smooth knowledge transfer between innovation subjects through digital platforms has become a novel research subject. This study aims to examine the knowledge transfer mechanism of digital platforms in the digital innovation ecosystem through modeling and simulation to offer a theoretical basis for digital innovation subjects to acquire digital value through knowledge-sharing and thus augment their competitive advantage.Design/methodology/approachThis study explores the optimal symbiotic interaction rate between different users based on the classic susceptible-infected-removed (SIR) model. Additionally, it constructs a knowledge transfer mechanism model for digital platforms in the digital innovation ecosystem by combining the theories of communication dynamics and symbiosis. Finally, Matrix Laboratory (MATLAB) software is used for the model and numerical simulation.FindingsThe results demonstrate that (1) the evolutionary path of the symbiotic model is key to digital platforms' knowledge transfer in the digital innovation ecosystem. In the symbiotic model, the knowledge transfer path of digital platforms is “independent symbiosis—biased symbiosis (user benefit)—reciprocal symbiosis,” aligning with the overall interests of the digital innovation ecosystem. (2) Digital platforms' knowledge transfer effects within the digital innovation ecosystem show significant differences. The most effective knowledge transfer model for digital platforms is reciprocal symbiosis, whereas the least effective is parochial symbiosis (platform benefit). (3) The symbiotic rate has a significant positive impact on the evolutionary dynamics of knowledge transfer on digital platforms, especially in the reciprocal symbiosis model.Originality/valueThis study's results aid digital innovators in achieving efficient knowledge transfer through digital platforms and identify how symbiotic relationships affect the knowledge transfer process across the ecosystem. Accordingly, the authors propose targeted recommendations to promote the efficiency of knowledge transfer on digital platforms.

PKAMNet: A Transmission Line Insulator Parallel- Gap Fault Detection Network Based on Prior Knowledge Transfer and Attention Mechanism

Insulator parallel gaps in a transmission line protect the insulator from arc burn and provide protection from lightning. Detection of insulator parallel-gap faults based on aerial inspection images of transmission lines may be hindered by various factors, such as complex background, scale distortion, and occlusion. In this study, we propose a prior knowledge transfer and attention mechanism network (PKAMNet) to detect faulty insulator parallel gaps. First, the capability of PKAMNet to learn the features of different types of parallel-gap faults is improved by constructing a prior knowledge transfer model based on visual saliency. Owing to the difficulty of effectively expressing the features of the fault target due to the deviation of the shooting angle and complex background of the inspection image, we enhance the feature expression ability of the fault area by embedding the coordinate attention block. Additionally, the parameters of the detection network are simplified, and the modules are optimized to improve detection efficiency. Comparison results on 3-year aerial inspection videos of 500 kV high-voltage transmission lines show that PKAMNet has a high detection accuracy and can ameliorate the phenomenon of missed and erroneous detection caused by insufficient expression of parallel-gap faults.