Heriot-Watt University

Few-Shot Learning (FSL) algorithms are commonly trained through Meta-Learning (ML), which exposes models to batches of tasks sampled from a meta-dataset to mimic tasks seen during evaluation. However, the standard training procedures overlook the real-world dynamics where classes commonly occur at different frequencies. While it is generally understood that class imbalance harms the performance of supervised methods, limited research examines the impact of imbalance on the FSL evaluation task. Our analysis compares 10 state-of-the-art meta-learning and FSL methods on different imbalance distributions and rebalancing techniques. Our results reveal that 1) some FSL methods display a natural disposition against imbalance while most other approaches produce a performance drop by up to 17% compared to the balanced task without the appropriate mitigation; 2) many meta-learning algorithms will not automatically learn to balance from exposure to imbalanced training tasks; 3) classical rebalancing strategies, such as random oversampling, can still be very effective, leading to state-of-the-art performances and should not be overlooked.

The business model is the story of how a business creates value and seeks to promote its long-term commercial sustainability. Increasingly—either directly or indirectly—all business value creating systems have been impacted by the trend towards globalization to which many have adapted as they seek to maintain fitness with their environment. This globality increases the degree of risk business models face as they have to adopt to an increasingly complex and unpredictable commercial environment.

This chapter examines the form and nature of risk within the global system. This approach indicates that global risk is multi-scalar across the identified sub-systems. In so doing, the global system is identified as a system that is subject to complexity and concentration risks. Dealing with systemic risk is dependent upon short-term resilience and longer-term sustainability within and across the identified sub-systems.

Economic globalization is arguably the most advanced process of globalization across the respective sub-systems. The process has progressed but as it has done so it has increased exposure at all levels to global events and processes. The chapter explores the main drivers of economic globalization and then moves on to explore the main types of risk within global economic system that is those that readily cascade between national systems to create a multi-state impact.

The state is at the core of global business operating as distinct territorial units between which businesses operate. The conditions of business vary markedly between states and as a consequence there is a big difference between states with regard to development and levels of risks associated with operating in a given territorial context. This chapter explores the notion of state strategy and encapsulates state-based risk within a holistic framework.

The embedded risk within global systems increasingly appears as an omnipresent feature of the contemporary business environment. This risk context has evolved so far that there is a prevailing debate with regard to the notion that the global system is caught within a Polycrisis. This reflects the simultaneous crises within the system that directly intensifies the risks businesses face. This concern has refocused attention upon the need to attend to system resilience and sustainability.

Geopolitical risk is arguably the greatest emerging threat to continuity within the global system. There has been a gradual erosion of the liberal internationalist order towards a higher degree of nationalism (in its various forms). As a result, there are increased forces for fragmentation within the global system sourced from a more fractious global political system.

As technology has globalized and spread throughout the global system there has been an increased focus upon the risks associated with this process. Many of these risks are associated with the rise of the so-called ‘Fourth Industrial Revolution’ and the technologies associated with its evolution. Inevitably most of these risks occur within developed economies (which have a wider deployment of these technologies) with many of the risks created by the wider utilization of cyber systems.

This chapter explores the globalization of social risk. This is seen as an especially important issue as these risks have grown alongside the globalization of other sub-systems. As these systems have interconnected so social systems have followed with the process of creating risks with regard to the erosion of social capital and cohesion. This chapter examines these challenges across three core themes: demographic change, inequality, and health and wellbeing.

Climate change alters human habitat, and this emerges largely as a result of the impact of human activity. The risks associated with this process are twofold. The first set of risks emerges as a direct consequence of the process itself and its disruptive impact upon human activity through—for example—more frequent extreme weather events. The second set of risks emerges as a legacy of those actions to deal with the impact of climate change. On top of those human-generated risks, there are also long-standing risks with regard to geological processes such as earthquakes and volcanic activity.

Infrastructure is key to the operation of the global system. These physical structures facilitate the movement of tangibles and intangibles across borders. Systemic dependence upon such structures means that any disruption to them can undermine systemic operation and is therefore a key risk to the global system. These risks exist across all main forms of hard infrastructure namely transport, energy, and information.

In this short paper, we elucidate how the solution of certain illustrative boundary value problems for the Airy equation \(u_t+u_{xxx}=0\) on [0, 1] can be expressed as a perturbation of the solution of the purely periodic problem. The motivation is to understand the role boundary conditions play in the properties of the solution. This is particularly important in related work on the solution of linear dispersive problems with discontinuous initial data and the phenomena of revivals and fractalization.

The Antipodal Vivaldi Antenna (APVA) has been identified as a promising technology for use in 5G systems due to its high frequency, wide bandwidth, and high gain. In this study, we evaluated the performance of the APVA using Ferro A6s substrate at 60 GHz and compared it with other substrates. Our findings demonstrate that the Ferro A6s substrate exhibits superior performance in terms of return loss, gain, and bandwidth when compared with other substrates. Our research contributes to the field of high-frequency wireless communication and can pave the way for the development of better 5G systems. The results of this study have the potential to revolutionize 5G backbone communication.

In this study, we used 3D printing technology to develop and create a quad-ridged horn antenna for use in GSM mid-band and 5G networks. This antenna can function on two different frequencies, 1.8 and 3.5 GHz. Polylactic acid (PLA), a thermoplastic polymer recognised for its outstanding strength and adaptability for 3D printing, was used to create the prototype. The effect of various dimension adjustments on the antenna's performance was studied through a series of parametric experiments. Ridge heights of 3, 4, and 6.25 mm were thoroughly tested. The investigation found that the antenna had a reflection of −16.89 dB at 1.8 GHz and −28.60 dB at 3.5 GHz, proving that it could function at both frequencies with great efficiency. This study proves that 3D printing may be used to create functional prototypes of antennas for use in GSM mid-band and 5G networks. As an added bonus, it stresses the significance of parametric studies in antenna design for peak efficiency.

The double empathy problem proposes a lack of shared understanding between autistic and non-autistic people leads to interaction difficulties between them. Social self-efficacy (confidence in one’s social abilities) makes an important contribution to social interactions. Research has shown that autistic people show a social identity (one’s sense of self based on membership of social groups) with other autistic people, and this impacts positively on well-being. In addition, research shows that autistic people have a number of social identities. This study aims to understand whether social self-efficacy relates to mental well-being in autistic adults, and whether social identity plays a role in this relationship. A total of 512 autistic adults completed measures of social self-efficacy, social identity, mental well-being and autistic traits. In-group social self-efficacy was found to be higher than out-group social self-efficacy. Second, in-group social self-efficacy was positively associated with well-being. Finally, social identity did not mediate the relationships between social self-efficacy and mental well-being. These findings suggest social self-efficacy may relate to the nature of the group one interacts with. The findings support further research into the relationship between social self-efficacy and well-being, as well as using the double empathy problem as a framework for further investigations in other social groups. Lay abstract In the past, research has suggested that autistic people are not able to communicate well with non-autistic people because of autistic people’s communication difficulties. However, newer theories question this conclusion. It is now thought that the communication difficulties may be because autistic and non-autistic people both struggle to understand each other. This study explores how these differences in shared understanding relate to autistic people’s mental well-being, confidence in social situations and social identities (groups that we belong to and that influence how we see ourselves). We created an online survey taken by 512 autistic adults, which included questions about their confidence being social with people from different groups (such as other autistic people or people they share a hobby with), about the social groups they felt they belong to and about their mental well-being. First, participants reported higher social confidence when interacting with members of a social group they belonged to. Second, being confident during these interactions was linked to higher mental well-being. Finally, the groups participants belonged to did not influence the link between social confidence and mental well-being. These findings are important as they help us better understand autistic people’s experiences of social interactions and what contributes to good and poor mental well-being in autistic people. They also help us to think further about how to improve autistic people’s well-being.

In human-altered environments, wildlife is impacted by a range of factors; therefore, understanding on responsible factors can aid in developing conservation strategies. Irrawaddy Dolphins (IRD: Oracella brevirostris) are a group of cetaceans listed as highly endangered species under - threatened species red list of IUCN. A subpopulation of IRD is found in Chilika lagoon, a brackish-water ecosystem in the eastern seaboard of India, which is also the single largest habitat of IRD in the world. They are survived by only a few hundred individuals in small pockets in Chilika and are therefore given maximum protection under the Indian Wildlife Protection Act (1972), but much remains to be done for their conservation as they have not drawn enough investigation. This article comprehensively synthesizes existing literature on IRD, has established linkages between their behavior and physiology with threats (both climate-driven and anthropogenic), and has finally laid down a conservation and management roadmap to safeguard their future in Chilika. Damming and diversion of rivers opening in Chilika, unsustainable dolphin watching tourism, pollution, and net entanglement are some of the prominent threats faced by IRD in Chilika. Predicted climate change scenarios such as potential warming of the lagoon, increased littoral drift, sea level rise, and increased frequency of cyclones would exacerbate their survival crisis. It is also emphasized that the conservation strategy of IRD must be moved from species-centric to ecosystem-based for their long-term sustenance.

Identifying the factors that influence the citation of articles helps authors improve the impact and reach of their research. Analysis of publications in the Journal of Fish Biology between 2008‐2021 revealed that variables such as the number of keywords, abstract length, number of authors and page length were associated with higher impact papers. These trends applied to both review and regular papers. These findings suggest that papers that are more informative, have higher numbers of authors, and more keywords are more likely to be cited. Adoption of some simple ‘best‐practice’ behaviours can improve the likelihood that a paper is cited. This article is protected by copyright. All rights reserved.

AIMS To succeed in clinical trials for glioblastoma we need in vitro models capable of more faithfully replicating dis- ease biology and more accurately predicting patient drug responses. To this end, new bioprinting technologies have the potential to biofabricate clinically relevant biomimetic tissues which can accelerate drug discovery and additionally serve as a platform for personalized medicine. METHOD We evaluated the effect of individual biomaterials and combinations of biomaterials, including decellularised pig brain extracellular matrix (dECM), fibrin, gelatin-methacryloyl (GelMA), hyaluronic acid-methacrylate (HAMA), Matrigel and alginate, on the proliferation and invasion of aggressive brain cancer cells (U87) in vitro. Cell viability was assessed using propidium iodide. Invasiveness was studied employing confocal microscopy. Data generated from Z-stacks was analysed using ImageJ to determine the size and circularity of cells. RESULTS Although Matrigel supports rapid cell proliferation and invasion, it has mechanical properties unsuited to bio- printing. In contrast, HAMA displays a pronounced shear-thinning behaviour and rapid controllable photo- crosslinking. Combinations of HAMA-fibrin provided results comparable to those seen with Matrigel or HAMA- Matrigel. However, high levels of crosslinking affected these biomaterial mixtures, resulting in a decreased ability of cells to grow and spread. CONCLUSIONS Initial results indicate that the addition of fibrin to HAMA promoted the growth and spreading of U87 cells. In further work, we aim to improve our printed constructs by including porcine brain dECM, microglia and recently established cell lines from paediatric patients. The project will test whether these bioprinted models can provide drug testing data with closer results to human disease than current, simpler alternatives.