Deep learning has emerged as a transformative approach in medicinal plant identification, addressing the critical need for accurate and scalable solutions to support biodiversity conservation, traditional medicine, and sustainable healthcare practices. This systematic literature review examines 30 papers on deep learning for medicinal plant identification, revealing diverse approaches across global contexts. Convolutional neural networks emerge as the primary technique, achieving high accuracy, particularly with leaf-based identification. Data collection methods vary, with manual fieldwork predominating. The review highlights challenges in scaling to larger species sets and using crowdsourced data, though strategies like data augmentation show promise. Plant state and maturity impact model performance, warranting further investigation. The geographical distribution of studies emphasizes the global relevance of this research, with India and China contributing the most. Mobile applications offer potential for deployment and data collection but lack robust user feedback mechanisms for model refinement. The review identifies gaps in continuous model updating and suggests exploring incremental and zero-shot learning. Overall, the field shows promise but requires more balanced datasets and context-aware approaches to maximize real-world impact in medicinal plant identification.

Smartphone-based plant identification increasingly serves as the edge tier of agricultural Internet of Things (IoT) systems, where models must adapt to crowdsourced data under bandwidth, memory, and energy constraints. No prior work has systematically investigated continual learning at the scale of thousands of fine-grained medicinal plant species, nor how retraining frequency affects the cost–performance trade-off in an IoT model-lifecycle setting. We evaluate three continual learning strategies—naïve fine-tuning, experience replay, and Learning without Forgetting—under periodic retraining schedules (updating every K increments), tested on 2,719 species (≥25 images each) from the Viet Medi Species 2026 dataset (310,647 images; 4,799 species total). All three strategies exhibit negative forgetting (performance improvement rather than degradation) in the instance-incremental setting, with naïve fine-tuning and LwF showing the strongest gains. Periodic retraining with K=2 reduces retraining operations by approximately 50% while maintaining performance. A baseline MobileNetV2 model achieves 54.07% top-10 accuracy across 2,719 species and has been deployed via TensorFlow Lite (FP16, ∼11.5 MB) in the Med Herb Lens Android application. Naïve fine-tuning is recommended as the practical default for instance-incremental agricultural IoT deployments.

Artificial intelligence (AI) is increasingly promoted as a means of providing objective player performance assessment in badminton. Compared to other sports, however, the supporting evidence remains fragmented. A systematic review based on 51 studies that satisfied the established quality and eligibility criteria from three major databases (covering the period 2018 to the end of 2025) reveals four dominating methodological schools: computer vision stroke tracking, movement-pattern recognition, spatio-temporal analysis of rally sequences, and multi-modal frameworks that integrate several data streams. Although many studies report high classification or prediction accuracy, only a small proportion of them employ shared validation datasets or evaluate repeatability across testing sessions, which limits the generalisability of their findings. Common shortcomings include small or imbalanced data samples, weak alignment with established sport-science theory, substantial computational requirements, and participant pools drawn largely from elite athletes in a single geographic region. Recent work has begun integrating explainable AI with retrieval-augmented generation (RAG) and large language model (LLM) frameworks to provide grounded, query-responsive feedback that links visual detections and performance metrics to structured match evidence. Future research should focus on larger and more diverse datasets, alignment with skill development models, transparent output formats, and validation across competitive levels and contexts with these state-of-the-art explainable AI-based RAG or LLM frameworks.

This article presents Med Herb Lens, a mobile application prototype that uses artificial intelligence to identify medicinal plants. Med Herb Lens does not attempt to identify all plant species, rather, it focuses on plants used for medicinal purposes. The app combines a deep-learning-trained image classifier system with a dynamic, user-augmented knowledge base. EfficientNetB0 was used for model development, with model performance improved via transfer learning. The model was converted into TensorFlow Lite to facilitate real-time and offline inference on Android devices. A Firebase backend architecture is employed to facilitate user contributions, synchronise the dynamic knowledge base, and enable users to access the application in remote areas. The implementation of the proof-of-concept model with a small dataset (seven medicinal plant species) yielded high classification accuracy and inference times of under 400 milliseconds on average. The app could also facilitate crowdsourcing of imagery and metadata, while allowing for iterative model improvement and development, and preserving traditional botanical knowledge. This work suggests that AI-powered tools can bridge the gap between conventional medicine and technology and be implemented in low-resource settings, focusing on under-resourced communities.

Adopting AI chatbots has gained significant momentum across various industries due to advancements in artificial intelligence. Despite their potential, AI chatbot adoption remains a complex process affected by numerous factors that are not fully understood. This systematic review seeks to identify and categorize the factors influencing AI chatbot adoption, including drivers and impediments. Following the PRISMA guidelines, a comprehensive review process was conducted. From 459 publications collected via Web of Science and Scopus, 84 research articles meeting eligibility criteria were analyzed to provide insights into the determinants of adoption. This review systematically examines the theoretical models employed, geographic distribution, primary domains of study, methodologies, and key factors shaping adoption. Furthermore, the study outlines future research directions to guide advancements in this area. The findings contribute to theoretical understanding by synthesizing determinants like anthropomorphism, trust, and hedonic motivation and advocating for integrating underexplored frameworks and hybrid methodologies. Practical implications are provided for developers, marketers, and policymakers, emphasizing user-centric design, privacy protection, and sector-specific strategies. This review advances knowledge in AI chatbot adoption and offers actionable insights for successful implementation across diverse industries.

The increasing number of international students (IS) enrolled in Australian higher education institutions, combined with the widespread adoption of online and hybrid learning, has significant implications for understanding the factors that influence engagement among this diverse student group. Early identification of students with low engagement facilitates academic success, prevents poor outcomes, optimises resource allocation, improves teaching strategies, increases motivation, and supports long term success. . This study's main aim is to examine the use of AI to predict student engagement. Development of a theoretically informed survey that aimed to elicit post graduate students' engagement was developed and validated by expert judgement. In total, 200 copies of the survey were distributed, 121 responses were received, and 96 were considered for this study representing a response rate of 48%. This study promotes a multidimensional approach, utilising AI and ML methodologies, to determine the influence of social and cultural contexts on student engagement This approach enables educators and institutions to create effective strategies for enhancing the learning experience of postgraduate students. Multiple AI and ML techniques have been utilised including synthetic data generation methods such GaussianCopula, TVAE, GAN, CopulaGAN, and CTGAN. These techniques are specifically employed to predict various dimensions of engagement, including personal, academic, intellectual, social, and professional engagement. . The performance of AI/ML algorithms, including SVM, KNN, DT, GBM, RF, NB, LR, and ET, was assessed using several metrics including F1 Score, Sensitivity, Specificity, Confusion Matrix, and Accuracy. The models used in this study achieved up to 85% accuracy, offering a solid foundation for guidelines and support to enhance decision making processes in higher education. These findings provide valuable insights for both academics and policy makers, laying the groundwork for evidence-based strategies to improve student engagement.

Healthcare project management requires a nuanced approach capable of navigating rapid changes while adhering to strict regulatory requirements. Traditional methodologies via PMI/PRINCE2 offer structured risk management and documentation, but their rigidity can be limiting in dynamic healthcare environments. Conversely, Agile’s flexibility allows for quick adaptation but may lack the oversight needed for regulatory compliance. This paper investigates the current state of the latest literature from 2012 to 2024 to conduct a systematic literature review, utilising PRISMA guidelines. From a total of 4,717 extract studies, this paper analysed 38 relevant studies ensuring the quality of the filtration process. The findings from the final set of selected studies suggest balancing robust planning with iterative flexibility, healthcare projects can meet regulatory demands while responding to evolving patient and technology needs. This study provides the latest challenges and trends that promise improved stakeholder engagement, regulatory adherence and overall project efficiency in the healthcare sector.

In the rapidly evolving healthcare sector, project management faces challenges due to the complexity and dynamic nature of its environment. Traditional methodologies including PMI PMBoK, and PRINCE2, combined with Agile, provide structured and adaptable approaches; however, the potential of the integration of artificial intelligence (AI) techniques is still underexplored. This study investigates how AI-driven tools, such as machine learning (ML) and predictive analytics, embedded with project management methodologies can improve project efficiency, decision-making, and resource management in the healthcare sector. Through a comprehensive literature review, we identify key AI technologies that augment task automation, real-time insights, and predictive capabilities within healthcare project management. This study presents statistical evidence from the literature on the percentage distribution of project management methodologies with key aspects including adaptability to AI, compliance, flexibility, stakeholder engagement and risk management. We discuss how a hybrid approach that leverages the strengths of PRINCE2/PMI, Agile, and AI can accelerate timelines, improve adaptability, and enhance stakeholder satisfaction. Despite challenges such as data privacy and compliance, this study presents a mapping of AI technologies and project management methodologies along with a conceptual design towards building a strategic framework that aligns AI advancements with organizational goals, optimizing healthcare project outcomes.

The proper identification of medicinal species is crucial for effective conservation strategies, ethnobotanical studies, and subsequent medical applications. This article presents a dataset of medicinal species, featuring multi-kingdom and curated imagery, from Vietnam. In addition to a sub-collection from the PlantCLEF 2022 benchmark, we have added adjacent images from our dataset based on the PlantGBIF Occurrence API. This is possible because we taxonomically linked the scientific names from the Vietnamese medicinal plant library to GBIF species keys, which allowed us to merge our imagery across datasets. In total, we constructed a dataset with more than 173,000 images for 2,787 species across four kingdoms (Plantae, Fungi, Chromista, and Bacteria). We established a baseline of performance by training convolutional neural networks using transfer learning to conduct two tasks: 1) fine-grained species-level classification, and 2) coarse kingdom classification. For species prediction, we achieved a Top-1 accuracy of 40.12% and Top-5 accuracy of 60.66%, due primarily to the dominant class imbalance and visual similarity between many of the taxa. For kingdom prediction, we achieved 96.14% accuracy and demonstrated that coarse-level classification is feasible. The findings from this work demonstrate the potential to combine local knowledge with a global dataset, advancing the development of artificial intelligence-supported tools for tracking biodiversity and traditional medicine. Future work will focus on hierarchical classification, the integration of ethnobotanical metadata, and the implementation of optimised models on mobile platforms.

Artificial Intelligence (AI) has achieved immense progress in recent years across a wide array of application domains, with biomedical imaging and sensing emerging as particularly impactful areas. However, the integration of AI in safety-critical fields, particularly biomedical domains, continues to face a major challenge of explainability arising from the opacity of complex prediction models. Overcoming this obstacle falls within the realm of eXplainable Artificial Intelligence (XAI), which is widely acknowledged as an essential aspect for successfully implementing and accepting AI techniques in practical applications to ensure transparency, fairness, and accountability in the decision-making processes and mitigate potential biases. This article provides a systematic cross-domain review of XAI techniques applied to quantitative prediction tasks, with a focus on their methodological relevance and potential adaptation to biomedical imaging and sensing. To achieve this, following PRISMA guidelines, we conducted an analysis of 44 Q1 journal articles that utilised XAI techniques for prediction applications across different fields where quantitative databases were used, and their contributions to explaining the predictions were studied. As a result, 13 XAI techniques were identified for prediction tasks. Shapley Additive eXPlanations (SHAP) was identified in 35 out of 44 articles, reflecting its frequent computational use for feature-importance ranking and model interpretation. Local Interpretable Model-Agnostic Explanations (LIME), Partial Dependence Plots (PDPs), and Permutation Feature Index (PFI) ranked second, third, and fourth in popularity, respectively. The study also recognises theoretical limitations of SHAP and related model-agnostic methods, such as their additive and causal assumptions, which are particularly critical in heterogeneous biomedical data. Furthermore, a synthesis of the reviewed studies reveals that while many provide computational evaluation of explanations, none include structured human–subject usability validation, underscoring an important research gap for clinical translation. Overall, this study offers an integrated understanding of quantitative XAI techniques, identifies methodological and usability gaps for biomedical adaptation, and provides guidance for future research aimed at safe and interpretable AI deployment in biomedical imaging and sensing.

Background/Objectives: Dementia is a leading cause of cognitive decline, with significant challenges for early detection and timely intervention. The lack of effective, user-centred technologies further limits clinical response, particularly in underserved areas. This study aimed to develop and describe a co-design process for creating a Diagnostic and Statistical Manual of Mental Disorders (DSM-5)-compliant, AI-powered Smart Assistant (SmartApp) to monitor neurocognitive decline, while ensuring accessibility, clinical relevance, and responsible AI integration. Methods: A co-design framework was applied using a novel combination of Agile principles and the Double Diamond Model (DDM). More than twenty iterative Scrum sprints were conducted, involving key stakeholders such as clinicians (psychiatrist, psychologist, physician), designers, students, and academic researchers. Prototype testing and design workshops were organised to gather structured feedback. Feedback was systematically incorporated into subsequent iterations to refine functionality, usability, and clinical applicability. Results: The iterative process resulted in a SmartApp that integrates a DSM-5-based screening tool with 24 items across key cognitive domains. Key features include longitudinal tracking of cognitive performance, comparative visual graphs, predictive analytics using a regression-based machine learning module, and adaptive user interfaces. Workshop participants reported high satisfaction with features such as simplified navigation, notification reminders, and clinician-focused reporting modules. Conclusions: The findings suggest that combining co-design methods with Agile/DDM frameworks provides an effective pathway for developing AI-powered clinical tools as per responsible AI standards. The SmartApp offers a clinically relevant, user-friendly platform for dementia screening and monitoring, with potential to support vulnerable populations through scalable, responsible digital health solutions.

Medicinal plants fulfil critical global health needs, but reliable identification poses barriers. This systematic review analyses 30 studies on deep learning for automated medicinal plant species classification to offer real-world insights. Convolutional neural networks (CNNs) demonstrate over 90% testing accuracy on plant organs such as leaves and flowers, enabling precise recognition models. While increasing species diversity and the use of crowdsourced data may pose performance challenges, optimisation strategies such as data augmentation and ensemble models may help mitigate accuracy declines. It appears that plant states (fresh vs. dry/sliced) may impact the model performance, although some models could distinguish maturity stages with sufficient data. Across geographical regions, CNNs show strong local identification capabilities, but generalised global models require larger, inclusive datasets. While mobile apps provide practical deployment avenues, robust mechanisms for continuous user-driven refinement are lacking. Ultimately, it appears that context-conscious deep learning approaches balancing efficiency and representation across diverse contexts are imperative for maximising real-world impact. This timely review consolidates evidence to guide the responsible development of specialised medicinal plant recognition systems.

This paper presents a study on the use of incremental and zero-shot learning for classifying Vietnamese medicinal plants using image analysis. Traditional machine learning methods often struggle with the constant emergence of new plant species and variability in appearances. Our methodology combines incremental learning, which continuously updates the model with new data while retaining prior knowledge, and zero-shot learning, which classifies unseen plant species by leveraging semantic similarities. Evaluated on a unique dataset from Vietnam, our approach shows improved adaptability, robustness, and reduced dependency on extensive labeled data, making it suitable for dynamic environments like medicinal plant identification.

Cameras have emerged as one of the most important tools to bring ubiquity to the Internet of Things (IoT) since its beginning and can be used to improve contextual accuracy substantially by using effective face recognition technology. The recent literature suggests that there is a large accuracy gap between today’s publicly available techniques and state-of-the-art private face recognition systems. This paper aims at bridging this gap by presenting the results of utilizing the OpenFace public library to perform face recognition in the wild. The key focus of this paper is to present a mechanism that guarantees higher accuracy with low training and less prediction time. In order to avoid the problem of reduced accuracy with the increase in the number of samples, this paper presents a strategy to keep up with low training time and high accuracy in comparison with a range of different recognition/classification techniques. A collection of large data samples of 245 different human faces was utilized to conduct this study, where the training was performed gradually starting from 5 classes to 245 distinct classes, incorporating a total of 7273 images to achieve reasonable accuracy. This paper also includes the details on how the classification was performed on live stream using webcams, where training and classification were done using the NVIDIA Jetson nano platform.

Since the pandemic organizations have been required to build agility to manage risks, stakeholder engagement, improve capabilities and maturity levels to deliver on strategy. Not only is there a requirement to improve performance, a focus on employee engagement and increased use of technology have surfaced as important factors to remain competitive in the new world. Consideration of the strategic horizon, strategic foresight and support structures is required to manage critical factors for the formulation, execution and transformation of strategy. Strategic foresight and Artificial Intelligence modelling are ways to predict an organizations future agility and potential through modelling of attributes, characteristics, practices, support structures, maturity levels and other aspects of future change. The application of this can support the development of required new competencies, skills and capabilities, use of tools and develop a culture of adaptation to improve engagement and performance to successfully deliver on strategy. In this paper we apply an Artificial Intelligence model to predict an organizations level of future agility that can be used to proactively make changes to support improving the level of agility. We also explore the barriers and benefits of improved organizational agility. The research data was collected from 44 respondents in public and private Australian industry sectors. These research findings together with findings from previous studies identify practices and characteristics that contribute to organizational agility for success. This paper contributes to the ongoing discourse of these principles, practices, attributes and characteristics that will help overcome some of the barriers for organizations with limited resources to build a framework and culture of agility to deliver on strategy in a changing world.

The concept of a smart factory, under Industry 4.0 relies heavily on cyber physical systems (CPS) and intra-enterprise-wide-networks (IWN). Cloud-based implementation is incumbent to accomplish the promises of enterprise integration, automation, seamless information exchange and intelligent self-organisation. Extensive research has been conducted in this domain, however, there is still much research to be done from the perspective of such frameworks in small to medium size enterprises (SMEs). In this context, the agent-oriented smart factory (AOSF) framework provides a generic end-to-end supply chain (SC) model, compliant with CPS and Industry 4.0 standards. In order to support the crucial side of warehouse management, this paper presents AOSF's recommended agent-oriented storage and retrieval (AOSR) warehouse planner with hybrid logic-based strategy, which yields a smart time-stamped plan to manage product placement and retrieval efficiently. The AOSF-associated AOSR-planner uses the hierarchical task network (HTN) AI planning to ensure different warehouse operations in a timely manner.

With recent advances in the information revolution, Digital Twin, with its complementary technologies, especially Big Data, Internet of Thing (IoT), Artificial Intelligence (AI) and Multi-Agent Systems (MAS), is becoming the core of novel strategies to maintain sustainability in Industry 4.0 networks and smart manufacturing. Despite the potential advantages of Digital Twin, such as virtual accessibility, remote monitoring, and timely customisation, not all sorts of enterprises have the resources or capabilities to incorporate such an advanced system, particularly Small to Medium size Enterprises (SMEs), due to their volatile nature of Supply Chains (SC). In this context, a Cyber-Physical Systems (CPS)-based, Agent Oriented Smart Factory (xAOSF) framework presents an over-arching SC architecture, with an associated Agent Oriented Storage and Retrieval (AOSR) based warehouse management strategy to help bridge the gap between Industry 4.0 frameworks and SME-oriented setups. This paper presents an approach towards realising the concept of Digital Twin via the xAOSF/AOSR framework, utilising state-of-the-art semantic modelling and analytical industrial tools. An amalgamation of CPS and Big Data Analytics is important to establish an effective Digital Twin to improve system scalability, security, and efficiency. This paper brings attention to this critical intersection and highlights how the xAOSF/AOSR framework can be scaled to implement Digital Twin effectively, which helps in analysing the bottleneck and threshold-states in real-time, especially in manufacturing organisations, which can lead towards full autonomy in an Industry 4.0 environment.

The Fourth Industrial Revolution (Industry 4.0), with the help of cyber-physical systems (CPS), the Internet of Things (IoT), and Artificial Intelligence (AI), is transforming the way industrial setups are designed. Recent literature has provided insight about large firms gaining benefits from Industry 4.0, but many of these benefits do not translate to SMEs. The agent-oriented smart factory (AOSF) framework provides a solution to help bridge the gap between Industry 4.0 frameworks and SME-oriented setups by providing a general and high-level supply chain (SC) framework and an associated agent-oriented storage and retrieval (AOSR)-based warehouse management strategy. This paper presents the extended heuristics of the AOSR algorithm and details how it improves the performance efficiency in an SME-oriented warehouse. A detailed discussion on the thorough validation via scenario-based experimentation and test cases explain how AOSR yielded 60–148% improved performance metrics in certain key areas of a warehouse.