Knowledge Transfer Research Articles

Advancements in Data Augmentation and Transfer Learning: A Comprehensive Survey to Address Data Scarcity Challenges

Abstract: Deep Learning (DL) models have demonstrated remarkable proficiency in image classification and recognition tasks, surpassing human capabilities. The observed enhancement in performance can be attributed to the utilization of extensive datasets. Nevertheless, DL models have huge data requirements. Widening the learning capability of such models from limited samples even today remains a challenge, given the intrinsic constraints of small datasets. The trifecta of challenges, encompassing limited labeled datasets, privacy, poor generalization performance, and the costliness of annotations, further compounds the difficulty in achieving robust model performance. Overcoming the challenge of expanding the learning capabilities of Deep Learning models with limited sample sizes remains a pressing concern even today. To address this critical issue, our study conducts a meticulous examination of established methodologies, such as Data Augmentation and Transfer Learning, which offer promising solutions to data scarcity dilemmas. Data Augmentation, a powerful technique, amplifies the size of small datasets through a diverse array of strategies. These encompass geometric transformations, kernel filter manipulations, neural style transfer amalgamation, random erasing, Generative Adversarial Networks, augmentations in feature space, and adversarial and meta- learning training paradigms. : Furthermore, Transfer Learning emerges as a crucial tool, leveraging pre-trained models to facilitate knowledge transfer between models or enabling the retraining of models on analogous datasets. Through our comprehensive investigation, we provide profound insights into how the synergistic application of these two techniques can significantly enhance the performance of classification tasks, effectively magnifying scarce datasets. This augmentation in data availability not only addresses the immediate challenges posed by limited datasets but also unlocks the full potential of working with Big Data in a new era of possibilities in DL applications.

CadastreVision: A benchmark dataset for cadastral boundary delineation from multi-resolution earth observation images

Approximately 70%–75% of people worldwide have no formally registered land rights. Fit-For-Purpose Land Administration was introduced to address this problem and focuses on delineating visible cadastral boundaries from earth observation imagery. Recent studies have shown the potential of deep learning models to extract these visible cadastral boundaries automatically. However, studies are limited by the small size and geographical coverage of available datasets and by the lack of information about which cadastral boundaries are visible, i.e., associated with a physical object boundary. To overcome these problems, we present CadastreVision, a benchmark dataset containing cadastral reference data and corresponding multi-resolution earth observation imagery from The Netherlands, with a spatial resolution ranging from 0.1 m to 10 m. The ratio between visible and non-visible cadastral boundaries is essential to evaluate the potential automation level in cadastral boundary extraction from earth observation images and interpret results obtained by deep learning models. We investigate this ratio using a novel analysis pipeline that overlays cadastral reference data with visible topographic object boundaries. Our results show that approximately 72% of the total length of cadastral boundaries in The Netherlands are visible. CadastreVision will enable new developments in cadastral boundary delineation and future endeavours to investigate knowledge transfer to data-scarce areas. Our data and code is available at https://github.com/jeroengrift/cadastrevision.

Two-Stage Approach for Targeted Knowledge Transfer in Self-Knowledge Distillation

Two-Stage Approach for Targeted Knowledge Transfer in Self-Knowledge Distillation

Corrections to "Toward Adaptive Knowledge Transfer in Multifactorial Evolutionary Computation".

Presents corrections to the paper, (Corrections to "Toward Adaptive Knowledge Transfer in Multifactorial Evolutionary Computation").

Bringing research into policy: understanding context-specific requirements for productive knowledge brokering in legislatures

Background: Knowledge brokering is suggested as an instrument to improve productive use of research in policy organisations. Previous research asserted that research utilisation is dependent on dynamics of knowledge exchange in institutional settings, but these claims have not received substantial empirical attention (Saarela et al, 2015; Akerlof et al, 2019; MacKillop et al, 2020). Viewing knowledge brokering as the involved role, three specific challenges are identified: high legitimacy requirements for the brokered knowledge and the broker; the need to cater for a wide range of topics, audiences and uses; and the need to compete with other evidence suppliers. Aims and objectives: The research question of the article is: how do legislative knowledge brokers navigate context-specific knowledge transfer challenges presented by their institutional context? Methods: An in-depth interpretive case study of the UK Parliamentary Office of Science and Technology. The analysis includes interviews with parliamentary actors, shadowing and participant observation. Findings: The results substantiate the challenges of legislative knowledge brokering in the UK context and inductively identify a further challenge of demonstrating impact. Legislative knowledge brokers employ multiple strategies to navigate the challenges: co-shape and adhere to the norms of impartiality, mobilise external expertise, collaborate with in-house and external research support actors, employ anticipation techniques, build broker chains, seek understanding of own role and impact. Discussion and conclusion: The article contributes to the understanding of knowledge brokering as a context-dependent role. The conclusions discuss influence of knowledge brokers’ work remit and positionality in deploying strategies to overcome the legislative challenges.

이해중심 교육과정에 기반한 호주 자국어 교육과정 문학 영역의 교육 내용 분석

Objectives This study aims to identify the advantages and limitations that may arise when the Understanding by Design (UbD) framework is fully implemented in the literature domain of the Korean language curriculum by analyzing the educational content of the literature domain in the Australian national language curriculum based on the UbD framework. Methods For this purpose, I selected educational content combined with performance verbs related to understanding (explanation, interpretation, application, perspective, empathy, self-knowledge) provided by the UbD framework from the literature domain of the Australian national language curriculum (version 9.0). The selected content was then analyzed according to sub-domains of literature and types of understanding. Results In terms of sub-domains, cognitive activities were distributed as educational content, focusing on the exploration of literature represented in socio-cultural and historical contexts (<Literature and Context>), engagement and response to literature (<Engaging with and Responding to Literature>), multifaceted exploration of literary forms (<Examining Literature>), and the transfer of literary knowledge (<Creating Literature>). Regarding types of understanding, the content was designed to ‘explain’ the reasons and methods of appreciating literature, ‘interpret’ by attributing value to understanding literature, ‘apply’ literary knowledge appropriately in new contexts, form ‘perspectives’ by relationally expanding and refining understanding of literature, and develop ‘self-knowledge’ by reflecting on personal understanding of literature. Conclusions The advantage of this approach is that it allows for a meta-understanding of literature by using literature phenomena as the organizing principle of literature education content and promotes the formation of a knowledge repertoire for transfer based on conditional knowledge. However, a limitation is that literature education activities tend to be biased towards the generalization of knowledge.

Innovative Design and Fast Iterative Manufacturing of Micro-Turbojet Engines at Low Cost

The turbojet is the foundation of the gas turbine engine, which combines knowledge of fluid dynamics, thermodynamics, heat transfer, mechanical design, manufacturing, and jet propulsion. This paper describes in detail the inspiration source, design process, fluid dynamics analysis, manufacturing steps, ignition test method and test process of small jet engines. 3D modeling software SolidWorks was used for the preliminary modeling and fluid simulation was performed to optimize the design. Part of the model was produced by 3D resin printing technology, and the turbine engine model was driven by compressed air for no-load testing to verify the feasibility and structural stability of the resin model. Subsequently, some parts were printed with metal materials, combined with traditional machining techniques to complete production and assembly, and finally successfully carried out static ignition tests. The micro turbojet engine designed in this paper has successfully achieved static ignition test and demonstrated the feasibility of efficient development with limited resources and low manufacturing cost. This exploration proves the feasibility of rapid iteration technology route. The market prospect and manufacturing cost of micro turbojet are also analyzed.

Refining English tutors' skills: Documentary films as pre-teaching reference

Training spoken English tutors requires a variety of resources, including audio-visual references. However, popular social media tutorials often focus on teaching specific English materials, neglecting broader teaching techniques. Ideal tutors need a positive and engaging personality, knowledge transfer strategies, adaptable approaches to classroom conditions, meticulous preparation, and effective wrap-up activities - elements that are missing from most tutorial videos. This study investigates the potential of short documentary films as a reference for teaching techniques for 13 prospective tutors in an intermediate English class. Through an integrated mentoring program for the production of short documentaries, the study explores, describes and addresses the challenges of teaching preparation through observation, interviews and literature review. It also analyzes the potential of documentaries to support teaching and formulates criteria for effective documentary content relevant to teacher preparation. The findings suggest that integrating documentary film production into the English language teaching curriculum can equip tutors with the necessary skills for successful micro-teaching experiences. This approach allows tutors to create personalized short documentaries that serve as a reference for teaching techniques tailored to their specific pre-teaching needs. Tutors gain a deeper understanding and ability to demonstrate not only teaching techniques, but also classroom management elements such as preparation, wrap-up activities, brainstorming and teaching approaches.

Reclaiming and Unlocking Ancient Heritage Knowledge from Ethiopia’s Ancient Cultural Heritages

Abstract The purpose of this paper is to reclaim and unlock heritage knowledge related to the secular value of time and labor from a sample Ethiopian ancient Ge’ez parchment manuscript. We claim that Ethiopia’s ancient Ge’ez parchment manuscripts could be a source of heritage knowledge if systematically studied. We conducted an extensive and intensive survey to identify and select relevant ancient Ge’ez parchment manuscripts from ancient churches and monasteries in Tigray, Ethiopia. We selected one Ge’ez parchment manuscript for content analysis related to secular value time and another Ge’ez parchment manuscript for the secular value of labor and work ethics. Content analysis and interpretation including language translation from Ge’ez to Tigrinya and English was used as a method of heritage knowledge extraction, as were unlocking approaches from the selected sample Ge’ez parchment manuscripts. Digital content was also produced in order to have access to the digital content, which facilitates further content analysis. A critical evaluation of the physical condition of the sample manuscripts including their palaeography, folios, marginalia, and provenance was made in order to establish their relative age and credibility. From our content analysis, we extract and unlock valuable information and heritage knowledge domains related to the secular value of time and labor. The striking aspect in this study is that the sources of such heritage knowledge are the ancient Ge’ez manuscripts preserved for centuries in religious institutions like ancient churches and monasteries. These sources of heritage knowledge could be easily accepted by Christian farmers to address and solve their earthly problems. We, therefore, assert that the analysis of manuscripts democratizes access to knowledge, making historical collections accessible to a broader audience. Therefore, this current research paper extends the scope of previous studies from merely identifying preservation challenges and making ancient manuscripts accessible through digitization to conducting content analysis, unlocking relevant heritage knowledge domains, and disseminating the knowledge as well as preventing the loss of valuable heritage knowledge. Finally, we suggest making a paradigm shift from preserving heritage for its own sake to focusing on heritage knowledge transfer and education. This is mainly because we currently stand at a crossroads, facing a dilemma between valuing and utilizing heritage on one hand and ensuring its sustainable preservation on the other.

Shortcuts to award winning research: analogies from ‘shortcuts to innovation: the use of analogies in knowledge production’

ABSTRACT In this interview, we explore the innovative research that earned Soomi Kim the Best Paper Award at DRUID 2024. As a partner of DRUID, Industry and Innovation presents an exclusive view into the research journey behind Kim’s celebrated work. Her study ‘Shortcuts to Innovation: The Use of Analogies in Knowledge Production’ examines how innovators venture into uncharted territories, using analogies to transfer knowledge from related domains. Using structural biology as the empirical setting, her findings highlight both the advantages and limitations of employing analogy-based technologies, such as machine learning, to speed up discovery. In this conversation, Kim reveals the inspiration, challenges, and breakthroughs that defined her research, providing a unique perspective on the paths leading to her pioneering work. Just as her paper examines how analogical reasoning and its technological automation serve as shortcuts in knowledge production, this conversation offers a unique lens on the pathways that shape groundbreaking research.

Determinants of knowledge transfer activities among university researchers: a systematic review

Purpose This systematic review aims to identify the key determinants of knowledge transfer (KT) activities among researchers. Design/methodology/approach This study systematically reviewed KT literature in academic settings from 1995–2023. The authors searched Web of Science and Scopus using predefined keywords, following PRISMA guidelines for screening and eligibility assessment. From 158 selected articles, the authors extracted data and conducted a descriptive analysis to map KT activities’ evolution. A narrative synthesis approach categorized determinants of researchers’ KT activities. Findings The systematic review findings revealed a general conceptual framework that categorizes the identified determinants of KT into four categories. At the individual level, the factors are related to the sociodemographic characteristics of the researcher (e.g. gender, age, experience), their psychological aspects (e.g. attitude, intrinsic motivation, intention) and personal characteristics (e.g. self-efficacy, communication skills). At the research team level, leadership style and team dynamics. At the organizational level, the findings emphasize university characteristics (e.g. size, structure and ranking), KT culture installed and university resources. At the inter-organizational level, the key determinants were funding sources, network strength and trust. Research limitations/implications The studies included in our database were different in terms of contexts, country of the study, the disciplines of KT and the types of KT activities examined. This variety restricts the direct comparison of research findings thus the generalizability of our conclusions. Future research should focus on specific contexts, disciplines, countries or types of KT activities to provide generalizable findings. Practical implications A better understanding of all the factors influencing KT among university researchers is essential for several reasons. First, it will enable the government to develop effective policies to promote KT ecosystems. Second, universities can create strategies, policies and programs to support researchers’ engagement in KT activities. Finally, researchers can be more strategic in their KT efforts. Originality/value This systematic review contributes to the literature by providing a comprehensive conceptual framework that identifies KT determinants at different levels and fills a gap in the existing literature that only addresses specific aspects of KT determinants. This framework can be a theoretical reference for future empirical studies. Furthermore, it practically provides recommendations for different actors including, government, universities and researchers.

Impact of Benevolence Trust on Knowledge Transfer

The impact of benevolence-based trust on knowledge transfer is understudied in Bangladesh's NGO sector, which is home to more NGOs than any other nation in the world (with a comparable geographic size). The study aims to examine the effects of benevolence-based trust, a significant component of trust antecedent, on knowledge transfer between NGO staff and their beneficiaries in Bangladesh's NGO sector. Semi-structured interviews with a total of 54 NGO personnel were conducted using multi-staged and purposeful sampling procedures. Content analysis is used to process Critical Incidents (CIs) collected from semi-structured interviews. Findings imply that knowledge transfer to beneficiaries is influenced by the benevolence- trust of NGO staff. The results may be useful for beneficiaries and NGO workers in expanding knowledge transfer activities. Given that the experiment was conducted in a specific country, such as Bangladesh, it could be challenging to generalize the results. Similar studies may be conducted in the future in a different setting or country.

Research and Model Library Construction in Teacher-Student Learning Architectures for Knowledge Transfer

This paper summarizes and replicates multiple classical and cutting-edge knowledge transfer methods, including Factor Transfer (FT), Knowledge Distillation (KD), Deep Mutual Learning (DML), Contrastive Representation Distillation (CRD), and Born-Again Self-Distillation (BSS). Additionally, we studied three advanced knowledge transfer methods: Relational Knowledge Distillation (RKD), Similarity-Preserving (SP), and Attention-based Feature Distillation (AFD), successfully replicating an optimized version of KD, namely RKD. Based on these methods, a flexible model library was constructed in Pycharm, allowing the quick integration of multiple knowledge transfer strategies. The experimental results are visualized through a user-friendly interface, enabling intuitive comparisons of model training speed and performance across different methods. This research provides valuable insights into the challenge of building a reusable framework that efficiently integrates various knowledge transfer strategies into deep neural networks.

Exploring the key drivers of inter-organizational knowledge transfer in projects: evidence from international construction projects

PurposeDespite the knowledge transfer in project contexts which has been extensively studied by scholars, the study on inter-organizational knowledge transfer in international construction projects is still scattered and unsystematic. This research aims to explore the key factors influencing inter-organizational knowledge transfer of international construction projects and analyze how these factors interact to create a synthetic framework that enhances the effectiveness of knowledge transfer (EKT).Design/methodology/approachAt first, eight factors influencing inter-organizational knowledge transfer within international construction projects were identified, which were from the four dimensions of subject, relationship, channel and context, namely cultural distance, connection strength, organizational climate, intercultural competence, information technology capability, transmit willingness, receive willingness and richness of transfer channels. Then, a conceptual model was developed and 13 hypotheses were formulated, which were derived from a literature review and in-depth survey. After that, data from 353 respondents were collected and analyzed, and the hypotheses were tested by structural equation modeling analysis and bootstrapping test.FindingsThe results suggest that cultural distance hinders transfer willingness, which further affects EKT. Connection strength affects EKT by positively influencing transfer willingness and channel. Besides, organizational climate and intercultural competence positively influence transfer willingness and further affect EKT, while information technology capability affects the richness of transfer channels.Originality/valueThis research gives a thorough examination of the determinants influencing inter-organizational knowledge transfer of international construction projects, thus formulating available approaches that project managers and personnel can employ to effectively facilitate EKT.

Lightweight and efficient deep learning models for fruit detection in orchards

The accurate recognition of apples in complex orchard environments is a fundamental aspect of the operation of automated picking equipment. This paper aims to investigate the influence of dense targets, occlusion, and the natural environment in practical application scenarios. To this end, it constructs a fruit dataset containing different scenarios and proposes a real-time lightweight detection network, ELD(Efficient Lightweight object Detector). The EGSS(Efficient Ghost-shuffle Slim module) module and MCAttention(Mix channel Attention) are proposed as innovative solutions to the problems of feature extraction and classification. The attention mechanism is employed to construct a novel feature extraction network, which effectively utilizes the low-latitude feature information, significantly enhances the fine-grained feature information and gradient flow of the model, and improves the model’s anti-interference ability. Eliminate redundant channels with SlimPAN to further compress the network and optimise functionality. The network as a whole employs the Shape-IOU loss function, which considers the influence of the bounding box itself, thereby enhancing the robustness of the model. Finally, the target detection accuracy is enhanced through the transfer of knowledge from the teacher’s network through knowledge distillation, while ensuring that the overall network is sufficiently lightweight. The experimental results demonstrate that the ELD network, designed for fruit detection, achieves an accuracy of 87.4%. It has a relatively low number of parameters (4.3×105\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$4.3 \ imes 10^5$$\\end{document}), a GLOPs of only 1.7, and a high FPS of 156. This network can achieve high accuracy while consuming fewer computational resources and performing better than other networks.

Traditional African Vegetables Knowledge Translation: A Scoping Review

Traditional African vegetables (TAVs) are an important component of an ecologically sustainable approach to addressing food and nutrition security in Africa. This review aims to identify the scope of the knowledge translation literature that exists across multiple aspects of the TAV value chain and to identify gaps in knowledge translation research. A scoping review following the Preferred Reporting Items for Systematic Reviews and Meta-Analyses extension for Scoping Reviews (PRISMA-ScR) was conducted. The Population, Concept, and Context (PCC) framework was used to guide the development of inclusion and exclusion criteria for this review. The Web of Science, Scopus, Science Direct, and ABI/Inform databases were searched to identify the relevant studies which were then screened using the Covidence 2024 online software to select studies and extract information for analysis and reporting. A total of 182 studies/articles were selected and examined. The most significant knowledge documented about TAVs pertains to their vast genetic diversity contributing to their rich nutritional and medicinal properties, their ability to withstand harsh environmental conditions such as drought, pests, and diseases, as well as the effects of various post-handling techniques on the nutrient content of TAVs, with the production phase of the TAV value chain being the most studied. This review found that, although a breadth of knowledge about TAVs has been generated, there is still little documentation about transferring and utilizing this knowledge across the TAV value chain. This review identified that more research is needed to document the strategies and practices for the transfer and utilization of knowledge to ensure that it is utilized to enhance nutritional outcomes across Africa.

Guangdong’s New R&D Institutes

In pursuit of technological development, China has created new organizations to promote innovation. This brief examines New Research and Development Institutes (NRDIs), which are designed to foster knowledge transfer to industry. NRDIs were pioneered in Guangdong province in the 1990s, and have gained prominence in China’s national science, technology, and innovation policies since the 13th Five-Year Plan (2016–2020). NRDIs are defined by their market orientation and extremely flexible organizational form. They work by establishing “innovation platforms” with local governments and private knowledge actors to carry out research and development (R&D), commercialize scientific and technological achievements, incubate local technology industries, and cultivate high-end talent. NRDIs have been instrumental to regional development in Guangdong, and especially Shenzhen, where they have succeeded in attracting talent from outside the region. NRDIs have important policy implications for international competition for talent. Understanding NRDIs is crucial for other countries that want to improve their own inter-regional innovation resources and respond to the challenge of China’s drive to attract global talent and knowledge resources.

Increasing the collaboration of data science stakeholders with a knowledge management system

PurposeThis paper presents the design and implementation of collaborative data science framework (CoDS), a knowledge management system for consolidating data science activities in an enterprise.Design/methodology/approachThe development of the CoDS framework is grounded on the design science research methodology for information systems research. In our case study, we first designed the initial framework for CoDS based on a systematic literature review. Then, we collected the expert opinions of eight data scientists to validate the need for generic content for such a knowledge management system. In the second iteration, a portfolio prototype is developed by the same data scientists as a part of our technical action research. Finally, a survey is conducted with 57 data analyst candidates in the last iteration.FindingsUsing the CoDS portfolio strengthened the communication among data scientists and stakeholders to improve development and scaling activities. It eased the reuse or modification of existing analytical solutions in other company processes.Practical implicationsThe CoDS presents a platform on which business details, data-related knowledge, modeling procedures and deployment steps are shared for (1) mediating and scaling ongoing projects, (2) enriching knowledge transfer among stakeholders, (3) facilitating ideation of new products and (4) supporting the onboarding of new employees and developers.Originality/valueThis study proposes a novel structure and a roadmap for creating a data science knowledge management system for the collaboration of all stakeholders in an enterprise.

Development of English Board Learning Media to Increase Learning Motivation in English Subjects in Grade 3 of Elementary Madrasah

Students are less motivated to follow the English learning process. The methods and approaches used by teachers are less interesting to students, still centered on the teacher. This study developed English board learning media, and to increase student enthusiasm and make it easier for teachers to transfer knowledge to students. English Board is expected to be fun, interesting, and motivating, and not make students bored when participating in English learning activities. The "English Board" learning media can increase student enthusiasm in learning and provide a pleasant learning experience for students. English Board media uses a research and development model. This study was only conducted in five stages, namely: 1) potential and problem stage, 2) data collection stage, 3) product design stage, 4) product validation stage, and 5) product revision stage. The results of the media validity obtained an average percentage of 89.77%, this result is included in the category of very good or very suitable for use in implementing English learning activities.

Mitigating not-invented-here syndrome in consultant knowledge transfer by developing appropriate implementation guidelines

Purpose This study aims to propose a framework for the development of implementation guidelines (IGs) that can help consultants mitigate not-invented-here (NIH) syndrome during a consultant intervention as a specific type of knowledge transfer. Design/methodology/approach The authors adopted a design science research approach for proposing an NIH-mitigating IG development framework. Inspired by findings and rich primary data from two consultant interventions, the authors, through theory building, ground five core principles in the general theory of NIH attitude functions. Finally, the authors revisit two consultant interventions to identify and describe mechanisms that led to the enactment of the principles. Findings The proposed framework provides five principles for developing NIH-mitigating IGs. The present research proposes that successful knowledge transfer and the mitigation of NIH syndrome as a prerequisite for this success are conditioned by adequately developed IGs. Originality/value The originality and value of the present research lie in the proposed NIH-mitigating IG development framework containing a set of principles for IG development as a proactive rather than reactive approach to NIH mitigation. To the best of the authors’ knowledge, this is the first research to address the problematics of mitigating NIH syndrome in consultant knowledge transfer by focusing on developing appropriate IGs. By developing and implementing IGs based on the proposed framework, a more successful transfer of knowledge from consultants to clients should take place, thus, increasing the value that clients receive from consultancy.