OARS - Open Access Repository Suffolk
Not a member yet
2036 research outputs found
Sort by
Detecting new obfuscated malware variants: a lightweight and interpretable machine learning approach
Full text linkMachine learning has been successfully applied in developing malware detection systems, with a primary focus on accuracy, and increasing attention to reducing computational overhead and improving model interpretability. However, an important question remains under explored: How well can machine learning-based models detect entirely new forms of malware not present in the training data? In this study, we present a machine learning-based system for detecting obfuscated malware that is not only highly accurate, lightweight and interpretable, but also capable of successfully adapting to new types of malware attacks. Our system is capable of detecting 15 malware subtypes despite being exclusively trained on one malware subtype, namely the Transponder from the Spyware family. This system was built after training 15 distinct random forest-based models, each on a different malware subtype from the CICMalMem-2022 dataset. These models were evaluated against the entire range of malware subtypes, including all unseen malware subtypes. To maintain the system’s streamlined nature, training was confined to the top five most important features, which also enhanced interpretability. The Transponder-focused model exhibited high accuracy, exceeding 99.8%, with an average processing speed of 5.7 µs per file. We also illustrate how the Shapley additive explanations technique can facilitate the interpretation of the model predictions. Our research contributes to advancing malware detection methodologies, pioneering the feasibility of
detecting obfuscated malware by exclusively training a model on a single or a few carefully selected malware subtype and applying it to detect unseen subtypes
Decoding neural activity for part-of-speech tagging (POS)
No full textDecoding Part of Speech(POS) tagging directly from electroencephalography (EEG) signals whilst user overtly spoke (voiced speech) sentences could improve direct speech brain-computer interfaces (BCIs) using imagined or inner speech. To the best of our knowledge, earlier work uses machine learning approach using 74,953 sentences/tokens recorded in 75 EEG sessions. The tokens can be found in 4,479 phrases consisting of terms from the English Online treebank which contains the record of weblogs, newsgroups, reviews, and Yahoo Answers. The results demonstrated the feasibility of POS decoding from EEG based on word class, word frequency, and word length with accuracy of 71%, 86%, 89%, respectively. We believe that there is significant room for improvement with more advanced artificial intelligence. In this paper, we further extend the existing work with end-to-end transformers. Our results presents transformer model outperforms benchmark traditional ML results with +20% in length, +13% for the open vs closed class and +12% in frequency. In our empirical analysis, we find the decoding performance was better when using multi-electrode recordings as compared to single-electrode recordings
The commercial determinants of violence: identifying opportunities for violence prevention through a public health-based framework analysis
Full text linkViolence has immediate and long-term repercussions for the health of individuals and communities. Recent increases in the understanding of public health approaches to violence prevention have focused on the policies and practices of government, health, and other public sector agencies. However, the roles of commercial bodies in fostering and preventing violence remain largely unaddressed. The wealth and influence of some companies now exceeds that of many countries. Consequently, it is timely to explore the roles of commercial processes in violence. Using a conceptual framework for the commercial determinants of health, we examine seven practices: political; scientific; marketing; supply chain and waste; labor and employment; financial; and reputational management. We include areas directly linked with violence (e.g., firearms) and those that indirectly impact violence through the following: design and promotion of products; employment practices; and impacts on environment, poverty, and local resources. A range of avoidable commercial behaviors are found to increase levels of violence including the following: lobbying practices; distortion of scientific processes; polluting manufacture and supply lines; poor employee protections; financial investment in organizations and regimes associated with violence; and misleading communications and marketing. We conclude commercial actors can take action to ensure their workers, clients, suppliers, and distributors help prevent, not promote, violence. New technologies such as artificial intelligence are transforming corporate processes and products and offer opportunities to implement violence prevention through commercial developments (e.g., monitoring online content). International regulation of commercial behaviors is needed to prevent interpersonal and interstate conflict and harms to health and trade
Using resistance and ruptures to strengthen the supervisory relationship
No full textDeveloping expertise as a coaching supervisor entails more than imparting coaching-specific knowledge and experience but requires expertise in the delivery of supervision itself. Supervision is increasingly understood as a distinct area of specialism, one that supports coaches’ reflection, practice and development through a skilled synthesis of educational, monitoring and supportive functions. Yet, combining and synthesising these functions can give rise to many questions and dilemmas for supervisors and those they supervise, both in terms of designing supervisory practices that are optimal for coaches receiving supervision and in responding to issues arising from the moment-to-moment interactions in a supervision session itself. As the contexts in which supervision is delivered expand and diversify, the dilemmas relating to how to balance these different functions increase.
This conceptual paper considers, and offers ways of responding to, some of the dilemmas and challenges that arise when the educational, monitoring and supportive functions of supervision become misaligned, giving rise to resistance or ruptures within the supervisory relationship. The supervision literature as a whole suggests that these phenomena are common yet can be experienced as difficult to discuss openly. Where these experiences remain unspoken however, there is risk of a breakdown in the supervisory relationship, threatening both the effectiveness of supervision and potentially the coaching that is being supervised. Developing ways of conceptualising and working effectively with resistance and ruptures is therefore an important consideration for advancing the practice of supervision in the coaching field generally, and for adapting supervision to different contexts (including the professional setting in which the coaching is occurring, the nature of the coaching being delivered and the stage of coach career development to name but a few).
One approach to working with resistance and ruptures is the PURE Supervision Flower, developed by Corrie and Lane, a visual heuristic for assisting supervisors in identifying and enhancing specific areas of knowledge and skill that comprise the development of an effective approach. In brief, the PURE Supervision Flower comprises 12 petals each of which focuses on a specific area that is central to effective supervision and in which supervisors need to be skilled. These petals are grouped into four forms of activity captured by the acronym ‘PURE’ (thus, P=Preparing for supervision; U=Undertaking supervision; R=Refining supervision and E=Enhancing supervision). This framework, introduced and described in the paper, is grounded in the literatures of adult learning and practice education, supervision across professional groups and fields related to coaching including the literature on expert performance and data from other helping professions.
Through an exploration of how the PURE Supervision Flower can support a process of conceptualising, addressing and using resistance and ruptures to strengthen the working relationship, this paper seeks both to equip coaching supervisors with specific approaches they can use to enhance their practice as well as contribute to a collective understanding of the philosophies, purposes and practices that underpin this complex yet highly rewarding coaching specialism
Two-level dynamic programming-enabled non-metric data aggregation technique for the Internet of Things
Full text linkThe Internet of Things (IoT) has become a transformative technological infrastructure, serving as a benchmark for automating and standardizing various activities across different domains to reduce human effort, especially in hazardous environments. In these networks, devices with embedded sensors capture valuable information about activities and report it to the nearest server. Although IoT networks are exceptionally useful in solving real-life problems, managing duplicate data values, often captured by neighboring devices, remains a challenging issue. Despite various methodologies reported in the literature to minimize the occurrence of duplicate data, it continues to be an open research problem. This paper presents a sophisticated data aggregation approach designed to minimize the ratio of duplicate data values in the refined set with the least possible information loss in IoT networks. First, at the device level, a local data aggregation process filters out outliers and duplicates data before transmission. Second, at the server level, a dynamic programming-based non-metric method identifies the longest common subsequence (LCS) among data from neighboring devices, which is then shared with the edge module. Simulation results confirm the approach’s exceptional performance in optimizing the bandwidth, energy consumption, and response time while maintaining high accuracy and precision, thus significantly reducing overall network congestion
Hamstring injuries, from the clinic to the field: a narrative review discussing exercise transfer.
Full text linkPurpose: The optimal approach to hamstring training is heavily debated. Eccentric exercises reduce injury risk; however, it is argued that these exercises transfer poorly to improved hamstring function during sprinting. Some argue that other exercises, such as isometric exercises, result in better transfer to running gait and should be used if when training to improve performance and reduce injury risk. Given the performance requirements of the hamstrings’ during the terminal swing phase, where they are exposed to high strain, exercises should aim to improve the torque production during this phase. This should improve the hamstrings’ ability to resist over-lengthening consequently improving performance and limiting strain injury. Most hamstring training studies fail to assess running kinematics post intervention. Of the limited evidence available, only eccentric exercises demonstrate changes in swing phase kinematics following training. Studies of other exercise modalities investigate effects on markers of performance and injury risk, but do not investigate changes in running kinematics. Conclusions: Despite being inconsistent with principles of transfer, current evidence suggests that eccentric exercises result in transfer to swing phase kinematics. Other exercise modalities may be effective, but the effect of these exercises on running kinematics is unknown
Editorial: Strength and conditioning in football - driving physical performance through research and innovation
Full text linkSupporting Autistic People Through Pregnancy and Childbirth
No full textThis comprehensive and accessible guide is for every birthing and health professional looking to improve their care during pregnancy, birth, and aftercare for autistic women. With a distinct lack of scientifically approached work in this area, this much-needed book takes an intersectional, feminist approach and covers the background of modern birth practices and autism as a diagnosis.
With intersectionality as a core feature, the impact of cultural differences, underdiagnoses, stigma, and stereotypes amongst ethnic minorities is also included. It discusses how pain functions in the autistic brain as well as co-occurring conditions such as alexithymia, chronic pain, epilepsy, and Ehlers-Danlos Syndrome.
This multidisciplinary author team includes two well-established autism experts, and an experienced midwife and lecturer who provides invaluable birthing insight, as well as approaches for sensation management during birth, insider knowledge on midwifery protocols, and accessible tools for autistic pregnant people and families to use
Towards AI literacy: 101+ creative and critical practices, perspectives and purposes
Full text linkThis is the second open crowdsourced collection by #creativeHE which presents creative and critical practices, perspective and purposes from educators, researchers and students between September 2023 and January 2024.
We are grateful for all 119 contributions from 22 countries: Argentina, Australia, Bangladesh, Brazil, Bangladesh, Canada, China, Egypt, France, Germany, Greece, Ireland, Israel, Italy, Mexico, Netherlands, Slovenia, South Africa, Spain, United Arab Emirates, United Kingdom, Uruguay.
As the collection is made available under the Creative Commons License CC-BY-NC-SA licence, anybody can use the collection as open data to further interrogate the use of AI in Education. Please share any resulting outcomes with the editorial team and the wider community.
The collection has been generously supported by the Imagination Lab Foundation through the Playful Hybrid Higher Education project (https://playhybrid.education/) led by Sandra Abegglen and situated in the School of Architecture, Planning and Landscape at the University of Calgary.
Thanks go also to #creativeHE of which we, the editors, are all part of and that has acted as supporter of the creative AI collections from the very beginning. The #creativeHE community hosts all calls and dissemination activities for the AI collections on their website: https://creativehecommunity.wordpress.com
The diversity debate: Is midwifery higher education addressing the challenges of systemic racism?
Full text linkThis series of six articles is inspired by themes initially arising from the Royal College of Midwives (RCM 2023) State of Midwifery Education report. The series explores the current landscape and challenges in educating the future midwifery workforce, particularly those that pertain to the Higher Education workforce. This third article highlights the lack of Global Majority midwifery academics. We chart the decisions and experiences of Global Majority applicants and students in midwifery Higher Education. The barriers these students face include systemic and individual racism. These experiences likely impact on Global Majority midwives choosing academic careers, yet we need greater diversity and representation in our education workforce. We therefore explore how this might be achieved by examining the sources of systemic racism and how we decolonise midwifery curricula to tackle the inequality not just for educators and professionals but for the women in our care