Birmingham City University Open Access Repository
Not a member yet
8823 research outputs found
Sort by
TRACE: Tracking and Real-time Analytics for Cargo and Enterprises
No full textOne of the ongoing challenges in supply chain management is ensuring traceability and transparency, especially when handling sensitive or volatile products. Conventional systems cannot frequently monitor in real time, which results in inefficiencies and makes it challenging to confirm transit conditions. To improve supply chain transparency, this study proposes an integrated solution that combines blockchain technology with Internet of Things (IoT) devices. Real-time data on product handling and environmental conditions is gathered by the IoT devices and securely stored on a blockchain ledger. This method guarantees data integrity and gives all parties involved access to an immutable record. The suggested system seeks to increase supply chain participants’ trust, guarantee adherence to legal requirements, and boost operational efficiency
How Circular Economy Innovation Can Backfire on the Environment: Quantifying the Rebound Effect of the Textiles and Clothing Sector
No full textCircular economy ( CE ) is championed as a sustainability solution, promoting reuse, recycling and resource efficiency to reduce environmental harm. However, CE innovations can trigger a rebound effect (RE), where lower costs stimulate higher consumption and production, paradoxically negating sustainability gains. This study applies a multi‐region, multi‐sector dynamic computable general equilibrium (DCGE) model to quantify the rebound effect triggered by CE innovation in the textile and clothing (TC) sector, the second most polluting industry. Our findings reveal a 155% rebound backfire, showing that CE innovations in the TC sector may exacerbate rather than mitigate environmental pressures. This challenges the assumption that CE alone can drive sustainability and underscores the need for complementary policies. As an extension, we look at complementary policies to ensure that CE strategies deliver genuine sustainability benefits. One explored policy is a uniform Pigouvian tax on TC production whereby we quantify that a minimum rate of 1.25% is required to curb the RE. However, effective implementation requires targeted regulatory interventions that also account for socio‐economic trade‐offs, particularly in low‐income countries. Achieving genuine sustainability will require degrowth‐informed policies that explicitly target reductions in production and consumption to suppress the systemic drivers of rebound effects in the TC sector
Enhancing AI transparency in IoT intrusion detection using explainable AI techniques
No full textInternet of Things (IoT) networks continue to grow and have been integrated into critical applications such as healthcare, industrial control, and national infrastructure. The interconnected nature and resource-constrained devices can create numerous entry points for malicious actors who can bring about data breaches, unauthorised access, service disruptions, and even compromise critical infrastructure. Ensuring the security of these networks is essential to maintain the integrity and availability of services that could have serious social, economic, or operational consequences. Automated Intrusion Detection Systems (IDSs) have been widely used to identify threats with high accuracy and reduced detection time. However, the complexity of machine learning and deep learning models poses a serious challenge to the transparency and interpretability of the produced detection results. The lack of explainability in AI-driven IDS undermines user confidence and limits their practical deployment, especially among non-expert stakeholders. To address these challenges, this paper investigates the use of Explainable AI (XAI) techniques to enhance the interpretability of AI-based IDSs within IoT ecosystems. Specifically, it applies SHapley Additive exPlanations (SHAP) and Local Interpretable Model-Agnostic Explanations (LIME) to different Machine learning models. The models’ performance is evaluated using standard metrics such as accuracy, precision, and recall. The results show that incorporating XAI techniques significantly improves the transparency of IDS results, allowing users to understand and trust the reasoning behind AI decisions. This enhanced interpretability not only supports more informed cybersecurity practices but also makes AI systems more accessible to non-specialist users
Comparative environmental disclosure practices: evidence from the UK and the USA
No full textThis study examines environmental reporting practices among leading firms in the United Kingdom (UK) and the United States of America (USA) by analyzing the top 50 firms from the Financial Times Stock Exchange 100 (FTSE100) and the Standard & Poor 100 (S&P100) indices during 2018 and 2019. Using volumetric metrics from annual and stand-alone reports, we assess disclosure patterns prior to the adoption of the revised Global Reporting Initiative (GRI) Standards and in anticipation of the International Financial Reporting Standards (IFRS) S1 and S2. Statistical analyses including correlation tests and both independent and paired sample t-tests reveal three key findings: (1) UK firms disclosed significantly less environmental content than US firms, suggesting strategic prioritization of other environmental, social, and governance (ESG) topics and alignment with national reporting expectations; (2) strong correlations among volumetric disclosure measures, coupled with statistically significant shifts over time, indicate that firms exhibit both consistency and adaptability in disclosure practices; and (3) significant differences in disclosure volumes between the two countries confirm the influence of regulatory systems, national cultures, and reporting norms. These findings support legitimacy theory and suggest that the environmental disclosure volume highlights the role of institutional and cultural contexts. They illustrate how multinational corporations navigate regulatory changes and prepare for new standards, offering a valuable perspective for evaluating corporate environmental accountability. This study underscores the continuing relevance of legitimacy theory and calls for further research into the drivers of voluntary sustainability reporting in a shifting regulatory landscape
Wawk on the wild side: Context-dependence of pseudohomophone processing
No full textThe pseudohomophone (PH) effect refers to an established finding whereby in a visual lexical decision task, nonword letter strings that are pronounced like real words (e.g., WAWK) are harder to reject than nonword strings that are not pronounced like real words (e.g., FLIS). This article reports three lexical decision experiments that aimed at further exploring the underlying processing mechanisms. In Experiments 1 and 2, we compared PHs like WAWK with unpronounceable nonwords like NRUG and pronounceable nonwords like FLIS, making sure that all stimuli (including real-word fillers) were carefully matched in length, bigram frequency, and number of orthographic neighbors. Matching stimuli in this way resulted in the real-word fillers to be of low lexical frequency (lower than for the PHs’ base words). Experiment 1 employed a standard lexical decision task, whereas Experiment 2 used the two-alternative forced choice eye-tracking paradigm originally developed in Kunert and Scheepers (2014). Both experiments converged on showing a reversal of the classical PH effect: while unpronounceable strings like NRUG were correctly rejected relatively quickly, PHs like WAWK were indeed easier to reject than pronounceable nonwords like FLIS. Our final Experiment 3, by contrast, confirmed a “classical” PH effect when the same nonword stimuli were tested against high- rather than low-frequency words as fillers. We conclude that the direction of the PH effect strongly depends on the overall material context
Painterly Poetics and Difference in the Making
No full textThe work of Luce Irigaray has proven instrumental in opening up possibilities for feminist and women artists from the 1970s onward. It has provided ways to articulate and visualize feminist issues and to critique power structures that marginalize women, be they social, cultural, political, visual, linguistic, or otherwise
Between rock and a hard place: The impact of home country demand on exclusive international strategic alliances forged by new technology ventures
No full textThis study seeks to progress the relatively thin body of scholarly research on the exclusive characteristics of strategic international alliances forged globally, particularly by new technology ventures. Due to the liability of smallness and newness, these new ventures need to strategically adopt exclusivity in licensing to secure partners across the globe to help them overcome the lack of resources and market access capability. Adopting resource dependence theory, the present study suggests that market size is a key consideration for the determinants of exclusive licensing for new technology ventures. The study investigates if the home demand of a country will influence the propensity to form exclusive international partnerships for new technology ventures. Based on the dataset of 545 international partnerships across the globe, findings of the study provide strong support to the idea that new ventures based in developed countries with limited market size (i.e., small-developed countries) are disproportionately more inclined to offer exclusive partnerships. Significant and positive moderation to the above findings were found due to the effect of sub-sectors, but not due to the size of the partner firms in the international market. The post-hoc analysis considering international and domestic alliances combined sample indicated consistent findings. The findings have theoretical, practical, and policy related implications for international strategic partnerships
Augmenting safe system of working: a systems thinking approach with leading indicators embedded within
No full textComplex and evasive phenomenon such as safety requires a holistic, multifaceted and intricately monitored and managed approach as opposed to current fragmented and reductionistic methods predominating in safety management. Such gestaltism of combining componential elements of complexly integrated systems can be achieved through the adoption of systems thinking and via the use of weak but early signals known as leading indicators. Therefore, this current doctoral study seeks to engender a novel theoretical basis in the form of conceptual model for the promulgation of proactive and holistic safety management, which is founded on continual and iterative learning from past and current safety activities. Such a conceptual model is inductively developed through analysis of existing knowledge in the literature and is tested with real life case study data.
To achieve the research aim, the research philosophies of interpretivism and critical realism were adopted to study the phenomenon under investigation and develop new theoretical insights. Within this overarching epistemology, the research strategy of sequential mixed methods was employed by combining a systematic literature review and case study using combination of data analysis methods such as thematic analysis, content analysis, cross-comparison analysis and framework analysis. The research process follows two phases viz., in phase 1 pertinent literature is systematically reviewed with inductive reasoning and in phase 2 the research outcome from the preceding phase is tested with real case data using abductive and deductive reasoning. Consequently, the phase 1 of the study engenders a novel conceptual model for leading indicators’ development and implementation. To test this research outcome, a proof-of-concept is designed at phase 2 by adopting the development step of the conceptual model viz., by seeking to develop leading indicators from a combination of case study data and their relevant normative documents. In addition to testing the conceptual model, this step engenders a novel analytical framework which provides the systematic development of leading indicators from the qualitative dataset. As a result, a total of 484 new leading indicators were identified by using the analytical framework. Subsequently, all these three research outcomes (i.e. proof-of-concept model, analytical framework and examples of leading indicators) are validated through focus group interview of experts.
Consequently, the study has developed multiple research outcomes, viz., main contributions such as proof-of-concept model in Figure 7.9; analytical framework in Figure 8.3; as well as other research contributions such as guidance note for training efficacy assessment in Figure 7.4; Safety-in-cohesion model in Figure 7.7; and Dynamic theory of incident evolution in Figure 8.5. These research findings generated create the groundwork for: proliferation of systems thinking in understanding safety, its management and maintenance; propagation of proactive and pre-emptive stance in development of safety countermeasures; and promulgation of a dynamic and adaptable approach in the generation of safety intelligence for continuous improvement. Therefore, these emergent theoretical and practical contributions stemming from this current doctoral work will become instrumental in mitigating asset and personal risks related to frontline workers’ interaction with operating vehicles and construction machinery on highway work sites as well as in other safety critical industries and sectors. Moreover, the work will be influential in continuously monitoring safety status of complex systems and simultaneously preventing unfavourable events from taking place and learning from both failures and successes
Robust multi-label surgical tool classification in noisy endoscopic videos
No full textOver the past few years, surgical data science has attracted substantial interest from the machine learning (ML) community. Various studies have demonstrated the efficacy of emerging ML techniques in analysing surgical data, particularly recordings of procedures, for digitising clinical and non-clinical functions like preoperative planning, context-aware decision-making, and operating skill assessment. However, this field is still in its infancy and lacks representative, well-annotated datasets for training robust models in intermediate ML tasks. Also, existing datasets suffer from inaccurate labels, hindering the development of reliable models. In this paper, we propose a systematic methodology for developing robust models for surgical tool classification using noisy endoscopic videos. Our methodology introduces two key innovations: (1) an intelligent active learning strategy for minimal dataset identification and label correction by human experts through collective intelligence; and (2) an assembling strategy for a student-teacher model-based self-training framework to achieve the robust classification of 14 surgical tools in a semi-supervised fashion. Furthermore, we employ strategies such as weighted data loaders and label smoothing to enable the models to learn difficult samples and address class imbalance issues. The proposed methodology achieves an average F1-score of 85.88% for the ensemble model-based self-training with class weights, and 80.88% without class weights for noisy tool labels. Also, our proposed method significantly outperforms existing approaches, which effectively demonstrates its effectiveness
CODE-ACCORD: A Corpus of building regulatory data for rule generation towards automatic compliance checking
No full textAutomatic Compliance Checking (ACC) within the Architecture, Engineering, and Construction (AEC) sector necessitates automating the interpretation of building regulations to achieve its full potential. Converting textual rules into machine-readable formats is challenging due to the complexities of natural language and the scarcity of resources for advanced Machine Learning (ML). Addressing these challenges, we introduce CODE-ACCORD, a dataset of 862 sentences from the building regulations of England and Finland. Only the self-contained sentences, which express complete rules without needing additional context, were considered as they are essential for ACC. Each sentence was manually annotated with entities and relations by a team of 12 annotators to facilitate machine-readable rule generation, followed by careful curation to ensure accuracy. The final dataset comprises 4,297 entities and 4,329 relations across various categories, serving as a robust ground truth. CODE-ACCORD supports a range of ML and Natural Language Processing (NLP) tasks, including text classification, entity recognition, and relation extraction. It enables applying recent trends, such as deep neural networks and large language models, to ACC