FPT University

Applying Artificial Intelligence (AI) to the monitoring of live fish in natural environments represents a promising approach to the sustainable management of aquatic resources. Detecting and counting fish in water through video analysis is crucial for fish population statistics. This study employs AI algorithms, specifically YOLOv10 (You Only Look Once version 10) for identifying the presence fish in video frames, combined with the DeepSORT (Deep Simple Online and Realtime Tracking) algorithm to count the number of fish individual moving across the frames. A total of 9,002 frames were extracted from 13 videos recorded in five different environments: areas with submerged tree roots, shallow marine regions, coral reefs, bleached coral reefs and seagrass meadows. To train the recognition model, the dataset was divided into training, validation and testing sets in 8:1:1 ratio. The results demonstrated that the model achieved an accuracy of 89.5%, with processing times of 6.2ms for preprocessing, 387.0ms for inference and 0.9ms for postprocessing per image. The combination of YOLO and DeepSORT enhances the accuracy of tracking objects in aquatic environments, showing great potential for the monitoring of fishery resources.

The organic cosmetics market in Vietnam is rapidly growing, especially among Generation Z consumers who prioritize sustainability and eco-friendly products. Despite this expansion, the key factors driving purchasing decisions for organic cosmetics have not been adequately researched. This study addresses this gap by examining the influence of Social Media Marketing Activities (SMMAs) on Generation Z’s purchase intentions and eWOM, with perceived quality and perceived value as mediating factors. Using a quantitative approach, data were collected from 315 Generation Z participants in Vietnam through a structured questionnaire. The study explores various dimensions of SMMAs—interaction, customization, trendiness, and entertainment—and their effects on perceived quality, perceived value, eWOM, and purchase intention. Results show that SMMAs have a stronger impact on perceived quality (β = 0.726) than on perceived value (β = 0.503), suggesting that social media marketing strategies are particularly effective in shaping how Generation Z evaluates product quality. Additionally, perceived quality significantly influences perceived value (β = 0.312), eWOM (β = 0.346), and purchase intention (β = 0.279). Similarly, perceived value positively impacts eWOM (β = 0.395) and purchase intention (β = 0.402), while eWOM itself plays a direct role in driving purchase intention (β = 0.167). These findings highlight the interconnected role of social media marketing in influencing consumer behavior through both direct and mediated effects. This research provides strategic direction for brands in Vietnam’s organic cosmetics sector targeting Generation Z. By leveraging these findings, brands develop targeted social media campaigns that emphasize quality perceptions through interactive content and customized experiences, thereby effectively driving both purchase decisions and positive eWOM among Generation Z consumers.

Previous research has demonstrated that formulaic sequences (FSs), including such types as collocations, idioms, or phrasal verbs, are ubiquitous and play an important role in second/foreign language (L2) proficiency. However, FSs can pose significant challenges for L2 learners, especially those in contexts where there is limited exposure to L2 input. For this reason, a significant amount of research has investigated various ways to improve L2 learners' knowledge of FSs. Given the limited classroom time for teaching FSs deliberately, recent studies have examined whether FSs can be incidentally picked up from different modes of input, such as reading, listening, or watching TV. In particular, being rich in FSs comparable to everyday speech, authentic TV has been gaining traction in L2 research as a potential source of input for learning FSs. This chapter presents a narrative review on learning FSs from authentic TV, aiming to address the following two primary issues: the extent to which FSs can be learned from authentic TV and the factors that might affect the learning of FSs from this type of resource. In addition, recommendations for practice and suggestions for future research on the topic will also be provided.

Image-based similar trademark retrieval is a time-consuming and labor-intensive task in the trademark examination process. This paper aims to support trademark examiners by training Deep Convolutional Neural Network (DCNN) models for effective Trademark Image Retrieval (TIR). To achieve this goal, we first develop a novel labeling method that automatically generates hundreds of thousands of labeled similar and dissimilar trademark image pairs using accompanying data fields such as citation lists, Vienna classification (VC) codes, and trademark ownership information. This approach eliminates the need for manual labeling and provides a large-scale dataset suitable for training deep learning models. We then train DCNN models based on Siamese and Triplet architectures, evaluating various feature extractors to determine the most effective configuration. Furthermore, we present an Adapted Contrastive Loss Function (ACLF) for the trademark retrieval task, specifically engineered to mitigate the influence of noisy labels found in automatically created datasets. Experimental results indicate that our proposed model (Efficient-Net_v21_Siamese) performs best at both True Negative Rate (TNR) threshold levels, TNR = 0.9 and TNR = 0.95, with respective True Positive Rates (TPRs) of 77.7% and 70.8% and accuracies of 83.9% and 80.4%. Additionally, when testing on the public trademark dataset METU_v2, our model achieves a normalized average rank (NAR) of 0.0169, outperforming the current state-of-the-art (SOTA) model. Based on these findings, we estimate that considering only approximately 10% of the returned trademarks would be sufficient, significantly reducing the review time. Therefore, the paper highlights the potential of utilizing national trademark data to enhance the accuracy and efficiency of trademark retrieval systems, ultimately supporting trademark examiners in their evaluation tasks.

This research investigates the link between adopting green practices and export performance among exporting SMEs in Vietnam, a booming export-oriented economy. Utilising the natural resource-based view (NRBV), it investigates if green practices directly influence SMEs’ export performance and the mediating role of incremental and radical green innovation in this relationship. The sample consists of 319 Vietnamese exporting SMEs. PLS-SEM was employed for model testing, and Harman’s single-factor test and t-tests were utilised to evaluate potential biases. Green practices positively impact export performance directly and are partially mediated by both types of green innovation. This study applies the NRBV to exporting SMEs in Vietnam to illustrate that adopting green practices, directly and indirectly, enhances export performance through incremental and radical green innovation. Secondly, the study tackles methodological limitations by conceptualising EP as a reflective-formative higher-order construct and concentrating on SMEs. It offers a more refined comprehension of how environmental commitments enhance performance in smaller exporting enterprises. This study integrates the ambidexterity framework into the green innovation literature, demonstrating that the balance of exploitative and explorative actions through green practices’ adoption enhances export performance in both financial and non-financial aspects.

We prove a Liouville type theorem for nonnegative solutions of the problem $-\Delta _p u + |\nabla u|^\gamma = u^q$ in $\mathbb {R}^N_+$ with zero Dirichlet boundary condition, where $p>2$ and $q>\gamma > p-1$. Our proof combines a recent monotonicity result with a new Liouville type theorem for nonnegative stable solutions in dimension $N<N^\sharp$, where $N^\sharp$ is explicitly computed.

This study presents a modified YOLOv5 algorithm specifically designed to enhance small‐object detection in unmanned aerial vehicle (UAV) images. Traditional object detection in UAV images is particularly challenging due to the high altitude of the cameras, which results in small object sizes and varying viewing angles. To address these challenges, the algorithm incorporates an additional prediction head to detect objects across a wide range of scales, a channel feature fusion with involution (CFFI) block to minimize information loss, a convolutional block attention module (CBAM) to highlight the crucial spatial and channel features, and a C3 structure with a Transformer block (C3TR) to capture contextual information. The algorithm additionally employs soft non‐maximum suppression to enhance the bounding box scoring of overlapping objects in dense scenes. Extensive experiments were conducted on the VisDrone‐DET2019 dataset, which demonstrated the effectiveness of the proposed algorithm. The results showed improvements with precision scores of 55.0%, recall scores of 44.6%, mean average precision scores of mAP50 = 50.9% and mAP50:95 = 33.0% on the VisDrone‐DET2019 validation set, and precision of 50.8%, recall of 37.3%, mAP50 = 44.2%, and mAP50:95 = 27.3% on the VisDrone‐DET2019 testing set. The improved performance is due to the incorporation of attention mechanisms, which allow the proposed model to stay lightweight while still extracting the features needed to detect small objects.

Recent trends in academia and the business community have seen a shift in focus in knowledge management (KM) issues from technology adoption to the interplay between organizational and technology management, with a special interest in the integration of KM technology and organizational design. Organizational structure is fundamental in this exploration as it encapsulates how an enterprise manages knowledge, communicates, allocates decision-making power, and implements control mechanisms, thereby portraying the efficacy of an organization’s operations. This study aims to formulate and test a theoretical model based on the Organizational Knowledge/Information Processing Model, examining the relationships between environmental pressure, organizational innovativeness, KM technology, organizational size, centralization, and formalization. Using a structural equation modeling approach, the study analyzed survey data collected from 220 companies to explore how these variables interact and influence each other. The analysis revealed that in response to competitive pressure, organizations primarily employ KM systems, centralization, and formalization. Additionally, KM systems were found to act as mediators, enhancing both centralization and formalization within the organizations. The study concludes that competitive pressure drives organizations to adopt KM systems, which in turn facilitate increased centralization and formalization. This highlights the critical role of KM technology in shaping organizational structure and processes, ultimately impacting the effectiveness of organizational operations.

As environmental degradation escalates, the critical need to understand green purchasing intentions and behaviours among Vietnamese Generation Z becomes increasingly urgent. Although the Theory of Planned Behavior (TPB) has been extensively applied to study pro-environmental behaviours, there remain discrepancies in how green attitudes, green subjective norms, and green perceived behavioural control influence green purchasing intentions and green purchasing behaviours. This study aims to clarify these relationships within the unique socio-economic and cultural context of Vietnamese Generation Z, a demographic influenced by collectivistic cultural values, generational characteristics, and dynamic economic conditions. These factors may reshape the conventional dynamics of TPB. Utilising quantitative methodologies, this research analysed responses from 237 Vietnamese Generation Z consumers through structural equation modelling to assess the impacts of green attitude, green subjective norms, and green perceived behavioural control on green purchasing intentions and green purchasing behaviours, particularly focusing on the mediating role of green purchasing intentions. The findings demonstrate that green attitude, green subjective norms, and green perceived behavioural control significantly affect both green purchasing intentions and green purchasing behaviours, thereby confirming the mediating influence of green purchasing intentions. This research reaffirms TPB’s relevance in Vietnam’s distinct cultural and economic environment while contributing to the broader TPB literature by exploring the mediating effects among key variables. These results also underscore the need for policymakers and businesses to create community-oriented environmental programs and tailor marketing strategies to enhance pro-environmental purchasing among young consumers.

As e-commerce continues to reshape retail landscapes, logistics service quality (LSQ) has become a crucial determinant of customer trust, satisfaction, and long-term engagement. This study investigates the impact of logistics service quality (LSQ) dimensions on the behavioral intentions of Generation Z consumers within Vietnam’s rapidly expanding e-commerce sector. The research focuses on how various LSQ factors—timeliness, personal contact quality, order accuracy, order condition, order discrepancy handling, and return convenience—affect trust and satisfaction, which subsequently influence repurchase intention and electronic word-of-mouth (eWOM). A quantitative approach was employed, gathering data from 495 Generation Z consumers with prior online shopping experience. Partial Least Squares Structural Equation Modeling (PLS-SEM) was used to test the proposed model and its hypotheses. This study found that key LSQ attributes play a significant role in shaping both trust and satisfaction, which, in turn, drive repurchase intention and eWOM. However, the findings indicate that Generation Z’s expectations for seamless logistics experiences vary across different service attributes. While factors such as order accuracy, order condition, and order discrepancy handling contribute to satisfaction, they do not necessarily build trust, highlighting the generation’s high standards and perception of these aspects as fundamental rather than differentiating features. This study challenges traditional LSQ frameworks by highlighting the evolving expectations of digital-native consumers. It offers practical insights for e-commerce businesses, emphasizing the need for a strategic blend of technological efficiency, personalized interactions, and seamless post-purchase services to enhance customer loyalty and competitiveness in the digital marketplace.

The growth of Vietnam’s fast-food market is being spurred by the economy, urban migration, and young consumers. The increase in competitiveness makes it vital for companies to grasp the factors behind customer retention. Previous studies have focused on the service quality (SQ), pricing, and the overall ambiance of the restaurant, but brand equity in Vietnam’s unique market is still a gap that needs to be further investigated. This research fills that gap by looking into the relationships between brand equity, perceived SQ, pricing, and the level of customer satisfaction. We approached 324 respondents to understand the perspective of consumers better using the partial least squares structural equation modeling (PLS-SEM) model. According to our results, while SQ and an appropriate and clean environment promote satisfaction, reasonable prices are the most important consideration among other factors. Furthermore, brand image strengthens and enhances the effectiveness of all these factors. This study offers fast-food companies in Vietnam major insights on effective brand-building and customer loyalty strategies. By offering good-quality food in a decent setting and at reasonable prices, brands can establish a good rapport with their customers, thereby ensuring sustainability. As this market becomes more competitive, brands that ply these tricks will certainly gain an edge.

Nghiên cứu này được thực hiện nhằm phân tích tác động của quản trị doanh nghiệp tốt đến hiệu quả tài chính của các ngân hàng thương mại niêm yết trên Sở Giao dịch Chứng khoán Việt Nam trong giai đoạn 2021-2023. Thông qua việc thu thập và xử lý 120 báo cáo tài chính, nghiên cứu đã đo lường chất lượng quản trị doanh nghiệp dựa trên các yếu tố gồm: tính độc lập của Hội đồng quản trị, tỷ lệ thành viên độc lập trong Ban kiểm soát, quyền sở hữu của ban điều hành và sự hiện diện của Ủy ban kiểm toán, trong khi hiệu quả tài chính được đánh giá bằng tỷ suất lợi nhuận trên tài sản (ROA). Bằng phương pháp phân tích định lượng với mô hình hồi quy tuyến tính trên dữ liệu bảng và sử dụng phần mềm EViews, nghiên cứu đã chỉ ra rằng tính độc lập của Hội đồng quản trị có tác động tích cực nhưng không đáng kể về mặt thống kê đối với ROA, trong khi tỷ lệ thành viên độc lập trong Ban kiểm soát, quyền sở hữu của ban điều hành và Ủy ban kiểm toán đều có tác động tích cực và có ý nghĩa thống kê đến hiệu quả tài chính. Khi xét tổng thể, các yếu tố quản trị doanh nghiệp này đồng thời cho thấy tác động tích cực và đáng kể đối với ROA, với mức độ giải thích lên tới 84,82%, qua đó khẳng định vai trò quan trọng của việc nâng cao chất lượng quản trị doanh nghiệp trong việc cải thiện hiệu quả tài chính của các ngân hàng niêm yết tại Việt Nam.

Object pose estimation using learning-based methods often necessitates vast amounts of meticulously labeled training data. The process of capturing real-world object images under diverse conditions and annotating these images with 6 Degrees of Freedom (6DOF) object poses is both time-consuming and resource-intensive. In this study, we propose an innovative approach to monocular 6D pose estimation through self-supervised learning, eliminating the need for labor-intensive manual annotations. Our method initiates by training a multi-task neural network in a fully supervised manner, leveraging synthetic RGBD data. We leverage semantic segmentation, instance-level depth estimation, and vector-field prediction as auxiliary tasks to enhance the primary task of pose estimation. Subsequently, we harness advancements in multi-task learning to further self-supervise the model using unlabeled real-world RGB data. A pivotal element of our self-supervised object pose estimation is a geometry-guided pseudo-label filtering module that relies on estimated depth from instance-level depth estimation. Our extensive experiments conducted on benchmark datasets demonstrate the effectiveness and potential of our approach in achieving accurate monocular 6D pose estimation. Importantly, our method showcases a promising avenue for overcoming the challenges associated with the labor-intensive annotation process, offering a more efficient and scalable solution for real-world object pose estimation.

Amid a rapidly developing era, people can inevitably have problems with stress, depression, pressure, or difficulty sleeping due to frequent overthinking. To overcome the above problems, yoga will be an excellent solution to help adjust thoughts and harmonize body and soul, helping us relax, relax the mind, and retain positive thoughts. Negative and evil auras will be pushed away, and the worldview will improve. Yoga practice has incorrectly caused many unwanted injuries for practitioners. Therefore, we present an approach grounded in skeleton-based feature extraction and neural networks to find a solution to the recognition of yoga postures, creating a premise for researching a smart virtual trainer that supports home workouts for users from input image data converted into skeleton data through MoveNet. The classification models were used to train recognition and classification of yoga poses. The models were trained and evaluated on a dataset of 3939 images of 10 yoga poses. Experimental results show that the proposed algorithms are entirely suitable for the classification task when achieving good results on different metrics such as Precision, Recall, F1-score, and Accuracy.

The world we live in is rarely black and white. Real-life problems often involve ambiguity, uncertainty, and degrees of truth. Traditional logic, with its crisp boundaries and absolute values, can struggle in these messy situations. Fuzzy logic, however, offers a powerful tool for navigating the complexities of everyday life. This chapter explores the application of fuzzy logic in real-world problem solving. The book chapter begins by introducing the core concepts of fuzzy logic, including fuzzy sets and membership degrees. It then delves into various practical applications of fuzzy logic, showcasing its effectiveness in diverse domains such as AI, Pattern Recognition, Control Systems, Risk Assessment, and Medical Diagnosis. Through a chosen case study, the book chapter illustrates the problem-solving capabilities of fuzzy logic in a specific scenario. The chapter concludes by highlighting the benefits of fuzzy logic and its potential to become an even more valuable tool in tackling the messy problems we encounter in daily life.

This study aims to investigate how credibility cues from doctors and celebrities influence trust formation and consumer intention in cosmetic surgery livestreams. Guided by source credibility theory and signaling theory, we explore the mediating role of utilitarian and hedonic values between source credibility and two dimensions of trust—cognitive and affective. A mixed-method approach combines qualitative analysis of livestream content and comments with a quantitative survey (N = 190) analyzed using PLS-SEM. Findings reveal that doctor credibility enhances utilitarian value, fostering cognitive trust, while celebrity credibility significantly impacts both hedonic value and trust, emphasizing emotional engagement. Trust emerges as a critical driver of consumer intention to use cosmetic services, highlighting the complementary roles of doctors and celebrities in livestreaming contexts. This study contributes to digital healthcare marketing by offering actionable guidance: livestream strategies should assign informational roles to doctors to boost cognitive trust, and use celebrities to enhance emotional engagement, ultimately increasing conversion intentions.

Natural Language Processing (NLP) has become essential in text classification, sentiment analysis, machine translation, and speech recognition applications. As these tasks become complex, traditional machine learning and deep learning models encounter challenges with optimization, parameter tuning, and handling large-scale, high-dimensional data. Bio-inspired algorithms, which mimic natural processes, offer robust optimization capabilities that can enhance NLP performance by improving feature selection, optimizing model parameters, and integrating adaptive learning mechanisms. This review explores the state-of-the-art applications of bio-inspired algorithms—such as Genetic Algorithms (GA), Particle Swarm Optimization (PSO), and Ant Colony Optimization (ACO)—across core NLP tasks. We analyze their comparative advantages, discuss their integration with neural network models, and address computational and scalability limitations. Through a synthesis of existing research, this paper highlights the unique strengths and current challenges of bio-inspired approaches in NLP, offering insights into hybrid models and lightweight, resource-efficient adaptations for real-time processing. Finally, we outline future research directions that emphasize the development of scalable, effective bio-inspired methods adaptable to evolving data environments.

The politicization of environmental issues often creates divides between pro-economic conservatives and pro-ecology liberals. This study suggests that an integrative approach can bridge these ideological gaps, balancing ecological performance with economic profit for sustainable development. By combining the Diffusion of Innovations Theory (DIT) and the Theory of Planned Behavior (TPB), the study investigates how individual attitudes, subjective norms, perceived behavioral control, and systemic factors influence Green Technological Innovation (GTI) adoption. Using Smart PLS 4, the analysis underscores the pivotal role of government support in enhancing cost-effectiveness, the importance of individual attitudes in fostering resource-conserving ideologies, and the need for understanding the synergy between ecological and economic benefits. The findings provide actionable insights for policymakers, organizations, and practitioners to promote sustainable technology adoption and achieve environmental goals. Future research should explore factors like organizational culture and financial constraints to further enhance understanding and facilitate widespread GTI acceptance.

Social constructivism has been proved to be an effective approach toward learner-centered lessons and active learning with the essential role of teachers’ quality questions to trigger discussions. This research study was conducted at a private university in Vietnam where social constructivism has been implemented in Sociolinguistics for English-majored students with constructivist questions contributed by lecturers and attached to the official syllabus. This paper aims to analyze those questions using four criteria of constructivism to shed light on how the ready-made questions correlate with the social constructivist criteria in the implementation of lessons of Sociolinguistics for English majored students. Analysis results show that elicitation questions and non-constructive questions dominate at 31 percent each while reflection questions comprise a modest 3 percent. The study results imply a pattern for teachers’ quality questions together with suggested examples in the subject of Sociolinguistics so that they can authentically serve as useful tools to foster features of a social constructivist classroom.

This chapter explores the potential of Extended Reality (XR) technologies to revolutionize environmental education. By immersing learners in simulated ecological environments, XR can provide a more engaging and impactful learning experience. This chapter will delve into specific applications of XR in environmental education, such as virtual field trips to endangered habitats, interactive simulations of climate change impacts, and gamified experiences that promote sustainable behaviors. Additionally, it will discuss the pedagogical benefits of XR, including increased student motivation, deeper understanding of complex environmental issues, and the development of critical thinking skills. Finally, the chapter will address the challenges and opportunities associated with integrating XR into educational settings, highlighting the need for equitable access and ethical considerations.