Research Bank
Not a member yet
5195 research outputs found
Sort by
Predicting wildfires from satellite images using deep learning
No full textDetecting the possibility of wildfires early can help individuals and organizations respond appropriately and minimise the potential damage caused. This report investigates the use of MobileNetV3 for predicting the occurrence of wildfires from satellite images that do not contain visible wildfire spots. The paper delves into current research and highlights the value of satellite imagery as a homogeneous data sources for wildfire prediction. A thorough evaluation and comparison of MobileNetV3’s performance with larger, more complicated models like ResNet50 and VGG19 is conducted. The findings demonstrate MobileNetV3’s efficacy in balancing computational efficiency with predictive power, offering a lightweight yet effective alternative to traditional models. By examining the potential of lightweight neural networks in managing complex and difficult environmental data, this paper advances wildfire prediction methodologies especially in resource constrained contexts.
This research focuses on a key challenge: developing a model designed to predict small wildfires from satellite images that do not contain any visible wildfire spots. This is achieved by training the model on a dataset comprising satellite images of areas where wildfires with a small wildfire spot size just greater an 0.01 acres have occurred. Our approach addresses a critical gap in wildfire management by focusing on predicting these small-scale fires without needing heterogeneous data collection. The ability to predict small wildfires from non-fire satellite images enhances the accuracy and utility of early warning systems. Using satellite images as a single source of data ingestion removes the need for heterogeneous data collection involving soil and atmospheric data as well as vegetative and geological data. It allows for the implementation of targeted preventive measures, such as controlled burns, the creation of firebreaks, and stricter fire bans during high-risk periods. Motivated by the capacity to generalize and the possibility to overcome challenges posed by cloud cover, haze, and diverse landscapes, this research uses MobileNet V3, a deep learning model based on transfer learning to predict wildfire occurrence from satellite images. MobileNetv3 model has given promising results, with a recall of 92 percent and an accuracy of over 82 percent. Despite being a lighter model, MobileNetv3 has shown robust results when evaluated alongside heavier models like Resnet50 and VGG19
Exploring the role of AI in enhancing sustainability within New Zealand's hospitality industry: A study on knowledge, applicability, and perception in reducing food waste
No full textThe hospitality industry in New Zealand is a noteworthy contributor towards the country’s Gross Domestic Product, however, it is also responsible for a significant amount of food waste (FW). In recent years, there has been a growing interest globally in artificial intelligence (AI) to curb FW in the hospitality industry. This study explores the role of AI in reducing FW and enhancing sustainability within New Zealand's hospitality industry, specifically, it focuses on the factors critical for its implementation - knowledge, attitudes, and perceptions of hospitality stakeholders. The study adopted a mixed-methods approach involving quantitative and qualitative data collection from 131 industry professionals.
The data suggests that the sample is skewed towards the lower end of FW, requiring caution due to inconsistent self-reported data with audits (Chisnall, 2017). A quarter were unsure or found it difficult to estimate avoidable food waste (AFW), potentially indicating a lack of awareness or difficulty measuring FW accurately. Inaccurate demand forecasting is the primary cause of FW in catering services and cafes/coffee shops, while portion control and plate waste are cited as the main causes of FW in fine-dining restaurants and pubs. A lack of FW awareness is the leading cause of FW in hotels/resorts and fast-food outlets. The diverse range of challenges faced by the hospitality industry provides insights into the areas where improvements can be made to increase efficiency, reduce waste, and enhance customer satisfaction.
The study found a general lack of awareness and knowledge about AI among hospitality professionals, yet an openness to adopting AI-driven technologies for FW reduction. Challenges such as cost-effectiveness and proven effectiveness are key causes hindering AI adoption. Increased awareness and promotion of AI's return on investment could decrease scepticism and facilitate effective integration into FW reduction strategies. Finally, this research underlines the need for collaborative efforts among industry professionals, policymakers, and technology developers to overcome existing hurdles and leverage AI for sustainable practices in the hospitality sector in New Zealand
Barriers male secondary students encounter when considering counselling at school
No full textWhat are the barriers and facilitative factors male students encounter when considering accessing a school counsellor for emotional or mental well-being support
End users engagement with GenAI
No full textLarge Language Models (LLMs) are powerful artificial intelligence (AI) systems that promise to create new and innovative user experiences and applications in different contexts, including education. The integration of AI into everyday life is rapidly increasing. As LLM technologies, such as ChatGPT, become increasingly prevalent in educational settings, they can disrupt conventional teaching and learning methods rooted in legacy learning by offering personalized, dynamic, and interactive learning experiences.
The presentation aims to report the work in progress of an experimental study investigating user engagement with LLMs using a real-world scenario to understand better the prompts end users use when interacting with LLMs. The study uses a multi-case methodology and collects data from a sample (n=30) of participants who are asked to use ChatGPT and complete a predefined task. This is followed by a short survey to collect data about their background and ChatGPT usage experience. Then, the researchers collected the prompts the participants used for further analysis.
The collected prompts are analysed using several analytical techniques, such as lexical, semantic, and other qualitative measures, to gauge users’ GenAI proficiency levels and abilities. This analysis aims to uncover how end users’ prompts proficiency influences their user experience and interaction with GenAI tools. The results of this study will provide valuable insights into user engagement with generative AI tools, the metacognitive skills of the participants, their level of ChatGPT satisfaction, and the factors influencing their use of LLMs. The study findings will inform researchers and practitioners interested in end users' engagement with GenAI, including educational researchers and practitioners.
What is GenAI used for?
Rate of GenAI adoption in the workplacein the USA 2023, by industry
Examples of Gen AI application using LLMs
How to get the most out of GenAI?
Prompt engineering
Research objectives
Research method
The current stage of the study
Theoretical and practical implication
"She’ll be right" - Culture change through the lens of redeveloping culinary management and leadership
No full textThis Master's thesis explores the integration of hauora (wellbeing) practices into culinary education, challenging the traditional "She'll be right" mindset pervasive in the hospitality industry. Through a mixed-method approach grounded in autoethnography, the Cynefin framework, and te ao Māori principles like Manaakitaka and Te Whare Tapa Whā, this research presents the Kai Hauora model. The model emphasises holistic wellbeing by addressing physical, mental, social, and spiritual health, fostering a culture of care within culinary management and leadership. The thesis is delivered through an innovative podcast series, featuring industry professionals who critically engage with the Kai Hauora framework and its practical implications. This accessible format bridges the gap between academia and industry, offering real-world strategies to cultivate supportive kitchen environments. Findings underscore the pivotal role of conscious leadership in promoting wellbeing. By embedding hauora into teaching practices this research advocates for a sustainable transformation of kitchen culture. This work contributes to the growing discourse on hauora in hospitality, positioning culinary education as a catalyst for industry-wide change. The Kai Hauora model serves as a blueprint for fostering a nurturing and inclusive professional environment, paving the way for a healthier and more resilient workforce
Modern Moana: Cultural principles contributing to the wellbeing of Tamaitai Samoa (Samoan women) in Aotearoa NZ
No full textIn navigating life in the diaspora, young Samoan women in Aotearoa face unique challenges to their wellbeing. This research explores how a blend of tradition and modernity influences the lived experiences of these women. Through an autoethnographic approach, this study analyzes the significance of three core Samoan cultural principles—Sa (Sacred), Moa (Centre), and Va Tapuia (Sacred Relationships)—as living frameworks that promote identity, resilience, and holistic wellbeing.
Supplementary material link is for a University of Auckland Master Thesi
Hybrid YOLOv9-DETR model for strawberry disease detection: A non-end-to-end object detection approach
No full textDetection of plant diseases is critical for both agricultural productivity and food security. In this thesis, we propose a hybrid object detection system that combines YOLOv9 and DETR for the first time to identify disease symptoms in strawberry plants. YOLOv9, a state-of-the-art detection model known for its speed and efficient feature extraction, was fine-tuned on a custom dataset for disease symptom detection. DETR, a transformer-based architecture, was trained separately to refine predictions using attention mechanisms and global context understanding.
This study explored four sequential experiments to determine the most effective way to integrate YOLOv9 and DETR. Initially, the goal was to develop a fully end-to-end model, where both architectures would be optimized jointly. To achieve this, a feature bridging mechanism was introduced to align YOLOv9’s outputs with DETR’s input format. However, architectural incompatibilities and computational constraints revealed that end-to-end training was impractical, as YOLOv9’s convolutional feature maps did not naturally align with DETR’s transformer-based processing, leading to unstable training dynamics.
Recognizing these challenges, the research pivoted toward a non-end-to-end hybrid inference approach, where both models were trained separately and their outputs were merged at inference time. Instead of a feature bridging module, a novel bounding-box selection strategy was implemented to unify the results of both models. By calculating IoU values between potentially overlapping YOLOv9 and DETR detections, bounding boxes exceeding a 0.5 IoU threshold were filtered to retain only the most confident prediction. Additionally, when one model missed an object entirely, its bounding box was taken from the other model to ensure more comprehensive coverage. This method effectively reduced redundancy and leveraged the complementary strengths of YOLOv9’s fast detection and DETR’s refined bounding box alignment.
The final hybrid inference strategy achieved a [email protected] of 0.96, demonstrating high detection sensitivity. However, dataset imbalance and computational resource limitations impacted overall generalization and accuracy. Despite these constraints, this research lays a foundation for future hybrid architectures, emphasizing the importance of dataset balancing, feature alignment strategies, and robust model integration.
This study suggests several potential enhancements, including extended training cycles, dataset expansion, and real-time deployment for agricultural applications. The proposed YOLOv9-DETR hybrid system marks a key milestone toward automated plant disease monitoring, contributing to sustainable and intelligent agricultural practices
Challenges of the Muisca cultural revival: Lessons from Māori architectural resistance
No full textColombia’s Indigenous Muisca culture, once flourishing in the Bogotá region, has undergone significant transformation due to Spanish colonisation in the sixteenth century, which led to the displacement of Muisca traditions and their merging with Spanish traditions over centuries. Despite these changes, the Muisca community maintains a distinct cultural presence on Bogotá’s outskirts, particularly in Bosa municipality, now known as El Porvenir. The Muisca community has pursued further recognition and land rights to safeguard their heritage. However, they have faced challenges from urban development and historical injustices, highlighting the need for strategic architectural interventions to preserve their identity.
This article explores how architectural practices could empower Indigenous communities through an analysis of historical, cultural and social contexts, identifying Māori architectural strategies that act as a form of resistance to prevent assimilation by the Western culture. The goal is to gather insights from Māori architectural experience that could be applicable to the revitalisation of Muisca culture in Bogotá, Colombia. Cultural, historical and social contexts will be analysed, focusing on the role of architecture in shaping and preserving identity
The making of a professional: Insights from four decades within vocational education
No full textThis study delves into the complex construction of professional identity among professional administrative staff within New Zealand’s vocational education sector, marking a significant contribution to the discourse on professional identity in this context.
Since the beginning of the 21st century, research into the professional identity of staff within higher education has emerged, predominantly within university settings or focusing on leadership roles. However, there has been a lack of research specifically addressing the special character of vocational education in New Zealand.
Originating from my personal experiences, observations, and reflections on my work identity and the factors and environments that have influenced both my own identity and those of others, this study draws upon my time working in the New Zealand vocational education sector from 1979 to 2024. It combines my journey with the stories of eighteen participants and insights from three eminent practitioners, employing a methodological mix of autoethnography, narrative inquiry, and qualitative description. This approach sheds light on the intricate interaction between individual experiences and the broader institutional, societal, and political forces that shape professional identity.
The findings introduce new knowledge and perspectives to the conversation about professional identity construction within higher education. The study highlights the critical themes of the importance of nomenclature, the influence of professional space, self-perception, shaped through the foundational influences of family and whānau, and the crucial roles of leadership and professional networks in shaping identity. These themes reveal the complex interplay between individual and collective identity narratives, particularly within the unique context of vocational education in the New Zealand higher education sector.
At this pivotal moment for vocational education, this study provides practical implications for practice and policy. By enhancing the understanding and cultivation of professional identity among non-teaching staff, a heightened sense of professional identity not only enriches the professional lives of staff but also strengthens their commitment and engagement, thereby contributing to the delivery of applied vocational education relevant to the workplace of the 21st century
Environmental, social and governance (ESG) controversies and integrated reporting assurance: An event study of top 100 Johannesburg Stock Exchange (JSE) listed companies
No full textBackground
Motivations and objectives of the study
ESG controversies and their effects on CSR reporting and assurance
Research hypotheses
Methodology
Research models
Main findings
Additional analyses
Contribution of the stud