Abstract
Optical Coherence Tomography (OCT) is the major diagnostic tool for the leading cause of blindness in developed countries: Diabetic Macular Edema (DME). Depending on the type of fluid accumulations, different treatments are needed. In particular, Cystoid Macular Edemas (CMEs) represent the most severe scenario, while Diffuse Retinal Thickening (DRT) is an early indicator of the disease but a challenging scenario to detect. While methodologies exist, their explanatory power is limited to the input sample itself. However, due to the complexity of these accumulations, this may not be enough for a clinician to assess the validity of the classification. Thus, in this work, we propose a novel approach based on multi-prototype networks with vision transformers to obtain an example-based explainable classification. Our proposal achieved robust results in two representative OCT devices, with a mean accuracy of 0.9099 ± 0.0083 and 0.8582 ± 0.0126 for CME and DRT-type fluid accumulations, respectively. © 2023 IEEE.

Abstract
Assessing the societal contributions of research is not simple, especially for research projects that produce outputs with low technology readiness level. This paper analyses the potential societal impacts of research resulting in technologies with low maturity, but with the potential to be further developed in the long-term. It uses the case of the ESGRIDS (Enhancing Smart Grids for Sustainability) collaborative research project and its outputs aimed at enhancing smart grids for sustainability. Data was collected from the four participant research teams through two sequential questionnaires about technologies' state of development and expected long-term societal effects. Among the main results, we underscore the influence of individual perceptions and organisational contexts over the process of eliciting future developments. The analysis of technologies' status, barriers for market uptake, and potential future developments was translated into a technology roadmap, which outlined the time-dimension for technology maturity evolution and implementation impacts. The technologies developed within the ESGRIDS project can contribute to support consumers' energy decision-making and to encourage them to have a more active role in the electricity market. Those technologies can also create job opportunities associated with the development of new products and services, and contribute to mitigating climate change by promoting the use of renewable energies thus reducing carbon dioxide emissions, in addition to contributing to energy cost reduction by optimizing the use of supply and demand resources. Future research avenues point towards a methodology that can be used for assessing the potential impacts of research projects with low technology readiness outputs.

Abstract
The scoping review reported by this article aimed to analyze and synthesize state-of-the-art studies focused on the integration of cyber resilience in the implementation of smart cities. An electronic search was conducted, and 11 studies were included in this review after the selection process. According to the findings, cyber resilience represents a gap of the current research related to smart cities and, therefore, additional efforts are required to guarantee that smart cities are resilient to challenging events such as cyber-attacks or natural disasters. © 2023 ITMA.

Abstract
Intelligent wheelchairs aim to improve mobility limitations by providing ingenious mechanisms to control and move the chair. This paper aims to enhance the autonomy level of intelligent wheelchair navigation by applying reinforcement learning algorithms to move the chair to the desired location. Also, as a second objective, add one more chair and move both chairs in pairs to promote group social activities. The experimental setup is based on a simulated environment using gazebo and ROS where a leader chair moves towards a goal, and the follower chair should navigate near the leader chair. The collected metrics (time to complete the task and the trajectories of the chairs) demonstrated that Deep Q-Network (DQN) achieved better results than the Q-Learning algorithm by being the unique algorithm to accomplish the pair navigation behaviour between two chairs.

Abstract
In recent years, it has been increasingly necessary for citizens to understand real life statistical data—an ability that is rarely taught in schools, where the majority of tasks in statistics classes contain fictional data without context and make no demands on students to explore or explain. Since most real-world phenomena are multivariate (See Chap. 2), there is a need to develop students’ abilities dealing with complex data and stories they encounter in the media, in order to help prepare them for informed citizenship. The ProCivicStat project has developed materials to support teaching and learning, in the form of detailed lesson plans; a large repository of resources (http://iase-web.org/islp/pcs/) (in several languages) is freely available. This chapter describes our approach to the development of teaching resources. It introduces our storytelling approach in lesson plans, where we use real data in context to encourage students to explore and understand complex data, produce narrative accounts, and often make recommendations about appropriate social actions. The structure of this chapter is as follows: we start with a brief introduction on problems in most tasks commonly encountered in statistics education, and the need for real data in statistics teaching (Sect. 7.1), followed by the presentation of the milestones that are important for creation of lesson plans (Sect. 7.2), and after that we address the use of real data and our storytelling approach (Sect. 7.3). In Sect. 7.4 we talk briefly about empowering teachers (Sect. 7.4) and describe the teachers’ version of the lesson plan (Sect. 7.5). In Sect. 7.6 we present the guidelines for designing student activities, then proceed with an excerpt of a lesson plan to exemplify products of the proposed guidelines (Sect. 7.7). We then highlight the visualization tools that help promote the data exploration step (Sect. 7.8), and finish with a conclusion (Sect. 7.9). © Springer Nature Switzerl and AG 2022.

Abstract
Deep Learning models are tremendously valuable in several prediction tasks, and their use in the medical field is spreading abruptly, especially in computer vision tasks, evaluating the content in X-rays, CTs or MRIs. These methods can save a significant amount of time for doctors in patient diagnostics and help in treatment planning. However, these models are significantly sensitive to confounders in the training data and generally suffer a performance hit when dealing with out-of-distribution data, affecting their reliability and scalability in different medical institutions. Deep Learning research on Medical datasets may overlook essential details regarding the image acquisition procedure and the preprocessing steps. This work proposes a data-centric approach, exploring the potential of attention maps as a regularisation technique to improve robustness and generalisation. We use image metadata and explore self-attention maps and contrastive learning to promote feature space invariance to image disturbance. Experiments were conducted using Chest X-ray datasets that are publicly available. Some datasets contained information about the windowing settings applied by the radiologist, acting as a source of variability. The proposed model was tested and outperformed the baseline in out-of-distribution data, serving as a proof of concept. © 2023 IEEE.