Open Research
Permanent link for this community
About
Tuwhera Open Access Research Outputs provides free access to full texts of scholarly works from AUT's Schools, Research Institutes and Centres.
AUT's research is built on a foundation of innovation and excellence, with the aim that its discoveries and applications are shared in ways that enhance wellbeing and prosperity.
Adding your outputs
AUT staff research outputs are added to this collection via Research Elements. All items submitted to this collection are checked to ensure material does not breach publisher copyright and is suitable for archiving prior to being made open access.
Find out more about making your work open access
For help with Research Elements contact the Research and Innovation Office.
Browse
Browsing Open Research by Subject "0806 Information Systems"
Now showing 1 - 18 of 18
Results Per Page
Sort Options
- ItemA Lightweight, Effective and Efficient Model for Label Aggregation in Crowdsourcing(Association for Computing Machinery (ACM), 2023-10-26) Yang, Yi; Zhao, Zhong-qiu; Wu, Gongqing; Zhuo, Xingrui; Liu, Qing; Bai, Quan; Li, WeihuaDue to the presence of noise in crowdsourced labels, label aggregation (LA) has become a standard procedure for post-processing these labels. LA methods estimate true labels from crowdsourced labels by modeling worker quality. However, most existing LA methods are iterative in nature. They require multiple passes through all crowdsourced labels, jointly and iteratively updating true labels and worker qualities until a termination condition is met. As a result, these methods are burdened with high space and time complexities, which restrict their applicability in scenarios where scalability and online aggregation are essential. Furthermore, defining a suitable termination condition for iterative algorithms can be challenging. In this paper, we view LA as a dynamic system and represent it as a Dynamic Bayesian Network. From this dynamic model, we derive two lightweight and scalable algorithms: LAonepass and LAtwopass. These algorithms can efficiently and effectively estimate worker qualities and true labels by traversing all labels at most twice, thereby eliminating the need for explicit termination conditions and multiple traversals over the crowdsourced labels. Due to their dynamic nature, the proposed algorithms are also capable of performing label aggregation online. We provide theoretical proof of the convergence property of the proposed algorithms and bound the error of the estimated worker qualities. Furthermore, we analyze the space and time complexities of our proposed algorithms, demonstrating their equivalence to those of majority voting. Through experiments conducted on 20 real-world datasets, we demonstrate that our proposed algorithms can effectively and efficiently aggregate labels in both offline and online settings, even though they traverse all labels at most twice. The code is on https://github.com/yyang318/LA_onepass.
- ItemA Privacy-Preserving Word Embedding Text Classification Model Based on Privacy Boundary Constructed by Deep Belief Network(Springer Science and Business Media LLC, 2023-09-15) Ma, Bo; Lai, Edmund; Yan, Wei Qi; Wu, JinsongTo effectively extract and classify the information from reports or documents and protect the privacy of the extracted results, we propose a privacy classification named Word Embedding Combination Privacy-preserving Support Vector Machine (WECPPSVM) model to classify the text. In addition, this paper also proposes the Privacy-preserving Distribution and Independent Frequent Subsequence Extraction Algorithm (PPDIFSEA), which calculates the degree of independence of the training data input to the classification model by training the Deep Belief Network(DBN) in PPDIFSEA, then obtains the Privacy Boundary(PB). PB is an indispensable condition for both data sampling and privacy noise generation. And this model can protect privacy by injecting the privacy noise into the classification result, this method can interfere with the background knowledge-based privacy attack. Our quantitative analysis shows that the WECPPSVM proposed in this paper can approach mainstream text classification algorithms in terms of text classification accuracy while preserving privacy without increasing computational complexity. In addition, the fusion study and privacy threat evaluation also verify that the proposed PPDIFSEA method combined with WECPPSVM achieves an acceptable level of classification accuracy and privacy protection.
- ItemA Think-Aloud Study of Novice Debugging(Association for Computing Machinery (ACM), 2023-03-30) Whalley, Jacqueline; Settle, Amber; Luxton-Reilly, AndrewDebugging is a core skill required by programmers, yet we know little about how to effectively teach the process of debugging. The challenges of learning debugging are compounded for novices who lack experience and are still learning the tools they need to program effectively. In this work, we report a case study in which we used a think-aloud protocol to gain insight into the behaviour of three students engaged in debugging tasks. Our qualitative analysis reveals a variety of helpful practices and barriers that limit the effectiveness of debugging. We observe that comprehension, evidence-based activities, and workflow practices all contribute to novice debugging success. Lack of sustained effort, precision, and methodical processes negatively impact debugging effectiveness. We anticipate that understanding how students engage in debugging tasks will aid future work to address ineffective behaviours and promote effective debugging activities.
- ItemApple Ripeness Identification from Digital Images Using Transformers(Springer Science and Business Media LLC, 2023-06-10) Xiao, Bingjie; Nguyen, Minh; Yan, Wei QiWe describe a non-destructive test of apple ripeness using digital images of multiple types of apples. In this paper, fruit images are treated as data samples, artificial intelligence models are employed to implement the classification of fruits and the identification of maturity levels. In order to obtain the ripeness classifications of fruits, we make use of deep learning models to conduct our experiments; we evaluate the test results of our proposed models. In order to ensure the accuracy of our experimental results, we created our own dataset, and obtained the best accuracy of fruit classification by comparing Transformer model and YOLO model in deep learning, thereby attaining the best accuracy of fruit maturity recognition. At the same time, we also combined YOLO model with attention module and gave the fast object detection by using the improved YOLO model.
- ItemCISO: Co-iteration Semi-supervised Learning for Visual Object Detection(Springer Science and Business Media LLC, 2023-09-19) Qi, Jianchun; Nguyen, Minh; Yan, Wei QiSemi-supervised learning offers a solution to the high cost and limited availability of manually labeled samples in supervised learning. In semi-supervised visual object detection, the use of unlabeled data can significantly enhance the performance of deep learning models. In this paper, we introduce an end-to-end framework, named CISO (Co-Iteration Semi-Supervised Learning for Object Detection), which integrates a knowledge distillation approach and a collaborative, iterative semi-supervised learning strategy. To maximize the utilization of pseudo-label data and address the scarcity of pseudo-label data due to high threshold settings, we propose a mean iteration approach where all unlabeled data is applied to each training iteration. Pseudo-label data with high confidence is extracted based on an ever-changing threshold (average intersection over union of all pseudo-labeled data). This strategy not only ensures the accuracy of the pseudo-label but also optimizes the use of unlabeled data. Subsequently, we apply a weak-strong data augmentation strategy to update the model. Lastly, we evaluate CISO using Swin Transformer model and conduct comprehensive experiments on MS-COCO. Our framework showcases impressive results, outperforms the state-of-the-art methods by 2.16 mAP and 1.54 mAP with 10% and 5% labeled data, respectively.
- ItemDoing Big Things in a Small Way: A Social Media Analytics Approach to Information Diffusion During Crisis Events in Digital Influencer Networks(Australian Journal of Information Systems, 2024-01-28) Kishore, Shohil; Errmann, AmyDigital influencers play an essential role in determining information diffusion during crisis events. This paper demonstrates that information diffusion (retweets) on the social media platform Twitter (now X) highly depends on digital influencers’ number of followers and influencers’ location within communication networks. We show (study 1) that there is significantly more information diffusion in regional (vs. national or international) crisis events when tweeted by micro-influencers (vs. meso- and macro-influencers). Further, study 2 demonstrates that this pattern holds when micro-influencers operate in a local location (are located local to the crisis). However, effects become attenuated when micro-influencers are situated in a global location (outside of the locality of the event). We term this effect ‘influencer network compression’ – the smaller in scope a crisis event geography (regional, national, or international) and influencer location (local or global) becomes, the more effective micro-influencers are at diffusing information. This shows that those who possess the most followers (meso- and macro-influencers) are less effective at attracting retweets than micro-influencers situated local to a crisis. As online information diffusion plays a critical role during public crisis events, this paper contributes to both practice and theory by exploring the role of digital influencers and their network geographies in different types of crisis events.
- ItemEnhancing Aotearoa, New Zealand’s Free Healthline Service through Image Upload Technology(Hindawi Limited, 2024-02-02) Wilson, Miriama K; Pienaar, Fiona; Large, Ruth; Wright, Matt; Todd, Verity FBackground. Healthline is one of the 39 free telehealth services that Whakarongorau Aotearoa/New Zealand Telehealth Services provides to New Zealanders. In early 2021, an image upload system for viewing service user-uploaded images was implemented into the Healthline service. Aims. The aim of this research was to understand the utilisation of Healthline’s image upload system by clinicians and service users in New Zealand. Methods. This is a retrospective observational study analysing Healthline image upload data over a two-year period: March 2021 through to December 2022. A total of 40,045 images were analysed, including demographics of the service users who uploaded an image: ethnicity, age group, and area of residence. The outcome or recommendation of the Healthline call was also assessed based on whether an image was included. Results. Images uploaded accounted for 6.0% of total Healthline calls (n=671,564). This research found that more service users were advised to go to an Emergency Department if they did not upload an image compared to service users who used the tool (13.5% vs. 7.7%), whereas a higher proportion of service users were given a lower acuity outcome if they included an image, including visiting an Urgent Care (24.0% vs. 16.9%) and GP (36.7% vs. 24.3%). Conclusion. Service users who did not upload an image had a higher proportion of Emergency Department outcomes than service users who did use the tool. This image upload tool has shown the potential to decrease stress on Emergency Departments around Aotearoa, New Zealand, through increased lower acuity outcomes.
- ItemFruit Ripeness Identification Using YOLOv8 Model(Springer Science and Business Media LLC, 2023-08-31) Xiao, Bingjie; Nguyen, Minh; Yan, Wei QiDeep learning-based visual object detection is a fundamental aspect of computer vision. These models not only locate and classify multiple objects within an image, but they also identify bounding boxes. The focus of this paper's research work is to classify fruits as ripe or overripe using digital images. Our proposed model extracts visual features from fruit images and analyzes fruit peel characteristics to predict the fruit's class. We utilize our own datasets to train two "anchor-free" models: YOLOv8 and CenterNet, aiming to produce accurate predictions. The CenterNet network primarily incorporates ResNet-50 and employs the deconvolution module DeConv for feature map upsampling. The final three branches of convolutional neural networks are applied to predict the heatmap. The YOLOv8 model leverages CSP and C2f modules for lightweight processing. After analyzing and comparing the two models, we found that the C2f module of the YOLOv8 model significantly enhances classification results, achieving an impressive accuracy rate of 99.5%.
- ItemImportance of IT and Role Identities in Information Systems Infusion(Springer, ) Hassandoust, Farkhondeh; Techatassanasoontorn, Angsana; Tan, FelixInformation systems (IS) should be infused into individuals’ work activities for organizations to extract value from these systems. Studies have identified various factors that impact IS infusion, but few have examined the importance of individuals’ identities and the role of contextual factors. Drawing on identity and status characteristics theories, this study conceptualizes individuals’ material identity as IT identity, and role identity as IS infusion role identity and examines their relationships and effects on IS infusion as well as the role status characteristics play in shaping these relationships. The models were evaluated using survey data collected from enterprise systems users. Findings suggest that individuals’ IT identity shape IS infusion role identity, and together, these identities influence their IS infusion. Additionally, work-related and personal characteristics strengthen the relationships between identities and IS infusion. This study highlights the role of individual’s IT and role identities and status characteristics in fostering IS infusion.
- ItemMachine Learning and Network Analysis for Diagnosis and Prediction in Disorders of Consciousness(Springer Science and Business Media LLC, ) Narayanan, Ajit; Magee, Wendy L; Siegert, Richard JBACKGROUND: Prolonged Disorders of Consciousness (PDOC) resulting from severe acquired brain injury can lead to complex disabilities that make diagnosis challenging. The role of machine learning (ML) in diagnosing PDOC states and identifying intervention strategies is relatively under-explored, having focused on predicting mortality and poor outcome. This study aims to: (a) apply ML techniques to predict PDOC diagnostic states from variables obtained from two non-invasive neurobehavior assessment tools; and (b) apply network analysis for guiding possible intervention strategies. METHODS: The Coma Recovery Scale-Revised (CRS-R) is a well-established tool for assessing patients with PDOC. More recently, music has been found to be a useful medium for assessment of coma patients, leading to the standardization of a music-based assessment of awareness: Music Therapy Assessment Tool for Awareness in Disorders of Consciousness (MATADOC). CRS-R and MATADOC data were collected from 74 PDOC patients aged 16-70 years at three specialist centers in the USA, UK and Ireland. The data were analyzed by three ML techniques (neural networks, decision trees and cluster analysis) as well as modelled through system-level network analysis. RESULTS: PDOC diagnostic state can be predicted to a relatively high level of accuracy that sets a benchmark for future ML analysis using neurobehavioral data only. The outcomes of this study may also have implications for understanding the role of music therapy in interdisciplinary rehabilitation to help patients move from one coma state to another. CONCLUSIONS: This study has shown how ML can derive rules for diagnosis of PDOC with data from two neurobehavioral tools without the need to harvest large clinical and imaging datasets. Network analysis using the measures obtained from these two non-invasive tools provides novel, system-level ways of interpreting possible transitions between PDOC states, leading to possible use in novel, next-generation decision-support systems for PDOC.
- ItemMarket Research and Insight: Past, Present and Future(SAGE Publications, 2022-02-22) Yallop, Anca; Baker, Jonathan J; Wardle, JudithOne hundred years have passed since the founding of the first independent market research firm in the UK in 1921. This important milestone inspired this special issue of the International Journal of Market Research that explores the role and importance of market research through a historical lens. A historical approach enables recognising and (re)framing both academic and practitioner contributions to market research through the years. Knowing the past allows better understanding and appreciation of the present, while simultaneously enabling envisioning of the future. In this introduction, we briefly review the origins and development of market research before introducing the three papers that comprise the special issue.
- ItemNUNI - Waste: Novel Semi-supervised Semantic Segmentation Waste Classification with Non-uniform Data Augmentation(Springer Science and Business Media LLC, 2024-01-25) Qi, Jianchun; Nguyen, Minh; Yan, Wei QiWaste categorization and recycling are critical approaches for converting waste into valuable and functional materials, thereby significantly aiding in land preservation, reducing pollution, and optimizing resource usages. However, real-world classification and identification of recyclable waste face substantial hurdles due to the intricate and unpredictable nature of wastes, as well as the limited availability of comprehensive waste datasets. These factors limit efficacy of the existing research work in the domain of waste management. In this paper, we utilize semantic segmentation at individual pixel level and introduce a semi-supervised metod for authentic waste classification scenarios, leveraging the Zerowaste dataset. We devise a non-standard data augmentation strategy that mimics the ever-changing conditions of real-world waste environments. Additionally, we introduce an adaptive weighted loss function and dynamically adjust the ratio of positive to negative samples through a masking method, ensuring the model learns from relevant samples. Lastly, to maintain consistency between predictions made on data-augmented images and the original counterparts, we remove input perturbations. Our method proves to be effective, as verified by an array of standard experiments and ablation studies, achieved an accuracy improvement of 3.74% over the baseline Zerowaste method.
- ItemPeering Through the Lens of High-Reliability Theory: A Competencies Driven Security Culture Model of High-Reliability Organisations(Wiley, 2023-05-17) Hassandoust, Farkhondeh; Johnston, Allen CTo improve organisational safety and enhance security efficiency, organisations seek to establish a culture of security that provides a foundation for how employees should approach security. There are several frameworks and models that provide a set of requirements for forming security cultures; however, for many organisations, the requirements of the frameworks are difficult to meet, if not impossible. In this research, we take a different perspective and focus on the core underlying competencies that high-reliability organisations (HROs) have shown to be effective in achieving levels of risk tolerance consistent with the goals of a security culture. In doing so we draw on high-reliability theory to develop a Security Culture Model that explains how a firm's supportive and practical competencies form its organisational security culture. To refine and test the model, we conducted a developmental mixed-method study using interviews and survey data with professional managers involved in the information security (InfoSec) programs within their respective HROs. Our findings emphasise the importance of an organisation's supportive and practical competencies for developing a culture of security. Our results suggest that organisations' security cultures are a product of their InfoSec practices and that organisational mindfulness, top management involvement and organisational structure are key to the development of those practices.
- ItemPose Estimation for Swimmers in Video Surveillance(Springer Science and Business Media LLC, 2023-09-01) Cao, Xiaowen; Yan, Wei QiTraditional models for pose estimation in video surveillance are based on graph structures, in this paper, we propose a method that breaks the limitation of template matching within a range of pose changes to obtain robust results. We implement our swimmer pose estimation method based on deep learning. We take use of High-Resolution Net (HRNet) to extract and fuse visual features of visual object and complete the object detection using the key points of human joint. The proposed model could be applied to all kinds of swimming styles throughout appropriate training. Compared with the methods that require multimodel combinations and training, the proposed method directly achieves the end-to-end prediction, which is easily to be implemented and deployed. In addition, a cross-fusion module is added between parallel networks, which assists the network to make use of the characteristics of multiple resolutions. The proposed network has achieved ideal results in the pose estimation of swimmers by comparing HRNet-W32 and HRNet-W48. In addition, we propose an annotated key point dataset of swimmers which was created from the view of underwater swimmers. Compared with side view, the torso of swimmers collected by the underwater view is much suitable for a broad spectrum of machine vision tasks.
- ItemSign Language Recognition from Digital Videos Using Feature Pyramid Network with Detection Transformer(Springer Science and Business Media LLC, ) Liu, Yu; Nand, Parma; Hossain, Md Akbar; Nguyen, Minh; Yan, Wei QiSign language recognition is one of the fundamental ways to assist deaf people to communicate with others. An accurate vision-based sign language recognition system using deep learning is a fundamental goal for many researchers. Deep convolutional neural networks have been extensively considered in the last few years, and a slew of architectures have been proposed. Recently, Vision Transformer and other Transformers have shown apparent advantages in object recognition compared to traditional computer vision models such as Faster R-CNN, YOLO, SSD, and other deep learning models. In this paper, we propose a Vision Transformer-based sign language recognition method called DETR (Detection Transformer), aiming to improve the current state-of-the-art sign language recognition accuracy. The DETR method proposed in this paper is able to recognize sign language from digital videos with a high accuracy using a new deep learning model ResNet152 + FPN (i.e., Feature Pyramid Network), which is based on Detection Transformer. Our experiments show that the method has excellent potential for improving sign language recognition accuracy. For instance, our newly proposed net ResNet152 + FPN is able to enhance the detection accuracy up to 1.70% on the test dataset of sign language compared to the standard Detection Transformer models. Besides, an overall accuracy 96.45% was attained by using the proposed method.
- ItemSociomateriality in Action: Theorizing Change in Sociomaterial Practices of Working from Home(Springer, 2023-04-04) Waizenegger, Lena; Schaedlich, Kai; Doolin, BillThe COVID-19 pandemic has led to an enforced ‘big bang’ adoption of working from home, involving the rapid implementation and diffusion of digital collaboration technologies. This radical shift to enforced working from home led to substantial changes in the practice of work. Using a qualitative research approach and drawing on the interview accounts of 29 knowledge workers required to work from home during the pandemic, the study identified five sociomaterial practices that were significantly disrupted and required reconfiguration of their constitutive social and material elements to renew them. The paper further shows evidence of the ongoing evolution of those sociomaterial practices among the participants, as temporary breakdowns in their performance led to further adjustments and fine-tuning. The study extends the body of knowledge on working from home and provides a fine-grained analysis of specific complexities of sociomaterial practice and change as actors utilize conceptual and contextual sensemaking to perceive and exploit possibilities for action in their unfolding practice of work. Against the backdrop of the increasing adoption of hybrid working in the aftermath of the pandemic, the paper offers four pillars derived from the findings that support the establishment of a conducive working from home environment.
- ItemThe Ethics of Using Generative AI for Qualitative Data Analysis(John Wiley & Sons Ltd, 2024-01-21) Davison, Robert M; Chughtai, Hameed; Nielsen, Petter; Marabelli, Marco; Iannacci, Federico; van Offenbeek, Marjolein; Tarafdar, Monideepa; Trenz, Manuel; Techatassanasoontorn, Angsana; Diaz Andrade, Antonio; Panteli, Niki
- ItemThe Importance of Theory at the Information Systems Journal(Wiley, 2023-04-12) Diaz Andrade, Antonio; Tarafdar, Monideepa; Davison, Robert M; Hardin, Andrew; Techatassanasoontorn, Angsana; Lowry, Paul Benjamin; Chatterjee, Sutirtha; Schwabe, Gerhard