Abstract
Purpose: The assessment and observation of critical service factors within the retail industry have garnered increased importance in recent times, due to their perceived ability to shape superior future strategies. The aim of this study is to investigate the service elements that are deemed essential by consumers in the retail sector, specifically targeting the grocery retail industry. Design/Methodology/Approach: Our methodological framework incorporates a systematic review of previous literature and a meta-analysis of past studies that highlight the pivotal service elements within the chosen industry. Following the evaluation of existing literature, 55 studies met the inclusion criteria and were selected for further investigation. The systematic review first compiled information from multiple studies, which was then followed by a meta-analysis. This enabled us to statistically analyze the empirical data from the chosen studies, thereby drawing significant conclusions. Findings: The analyses pinpoint that elements such as personal interaction attributes, product quality and availability, and reliable service are of utmost importance to consumers. Interestingly, customer satisfaction was the only outcome that was positively influenced by all the examined service attributes. Additionally, our findings underscore that certain moderators, such as geographic region and timing of the study, sway the relationship between service attributes and customer outcomes. Originality: Despite numerous meta-analyses attempting to pinpoint the key service attributes for consumers, to the best of our understanding, this study is the first to focus on the retail industry, specifically on hypermarkets, supermarkets, or grocery stores. Therefore, this research bridges a gap in the literature and offers a significant contribution to the academic community by proposing an agenda for future research on customer service factors. It also provides invaluable insight for retail managers, outlining numerous practical implications and offering guidance.

Abstract
This paper describes the participation of the Artificial Intelligence and Information Retrieval (AIIR) Lab from the University of Southern Maine and the Laboratory of Artificial Intelligence and Decision Support (LIAAD) lab from INESC TEC in the CLEF 2023 SimpleText lab. There are three tasks defined for SimpleText: (T1) What is in (or out)?, (T2) What is unclear?, and (T3) Rewrite this!. Five runs were submitted for Task 1 using traditional Information Retrieval, and Sentence-BERT models. For Task 2, three runs were submitted, using YAKE! and KBIR keyword extraction models. Finally, for Task 3, two models were deployed, one using OpenAI Davinci embeddings and the other combining two unsupervised simplification models.

Abstract
This paper proposes a proof of concept for a reflective fiber optic sensor based on multimode interference, designed to measure glucose concentrations in aqueous solutions that mimic the range of glucose concentrations found in human saliva. The sensor is fabricated by splicing a short section of coreless silica fiber into a standard single-mode fiber. By studying the principles of multimode interference and Self-imaging it was developed a sensing head that has a total length of 29.1 mm, approximately equal to the second self-image cycle. This sensing head allowed us to detect low concentrations of glucose (ranging from 0 to 268 mg/dl).

Abstract
In recent years the great success of transformers-based models initially employed in Natural Language (NLP) tasks has led to the development of several transformers variations to be employed in a wide range of domains, such as vision. With the correct amount of training data and proper training, transformers can perform excellently compared to the Convolution Neural Networks (CNN) counterpart in the vision tasks. However, the main drawback of transformers concerns the know memory requirements that often exceed the available training platform, growing in a quadratic form regarding the input image size, and a great tendency to overfit. Several works address the memory problem by relaxing the model architecture versions, but mainly with reduced prediction capabilities. In this work, we evaluate Random Patch erasing among the image patch level of the transformer model as a regularization technique to reduce overfitting while at the same time alleviating training time. The evaluated regularization technique achieves competitive results on several image classification medical datasets. The evaluated Visual Transformers (ViT) models allow to be trained in a single GPU, reaching similar results to CNN counterparts, obtaining an accuracy 91.2%, 79.2% in two competitive image datasets, and reducing the training time on average by 22% on the transformers models.

Abstract
Abstract As a tool, the Sustainable Development Goals (SDG) guide local and regional leaders in developing policy approaches for better social development. SDGs are 17 ambitious objectives towards a greener, healthier, more peaceful and equal planet, promoted by the United Nations to achieve by 2030. Having this performance in mind, countries and regions can measure their level of SDG implementation and rethink how they could promote prosperity, cooperation among regions and progress. This study focuses on SDG-9: Industry, innovation and infrastructure in ten municipalities of the Alto Minho region, Portugal. The main idea is to assess the level of each municipality in the achievement of the indicators related to this SDG. The similarities and differences between the municipalities can underline areas for joint efforts or investments in the development policy. This paper selected a performance analysis as a tool for informing on the amount of effort required to achieve SDG-9 at a local level, i.e., the Alto Minho region in the north of Portugal. If the trend of evolution is maintained, only Viana do Castelo will reach the full range of indicators for SDG-9, and Caminha will have 50 % of the indicators achieved. The remaining municipalities will reach at least half of the indicators, thus achieving a value lower than half of the target value. This approach could be replicated in other SDGs and other regions. This assessment allows the region’s stakeholders to indicate areas of required action to achieve the SDG.

Abstract
Glaucoma is a disease that arises from increased intraocular pressure and leads to irreversible partial or total loss of vision. Due to the lack of symptoms, this disease often progresses to more advanced stages, not being detected in the early phase. The screening of glaucoma can be made through visualization of the retina, through retinal images captured by medical equipment or mobile devices with an attached lens to the camera. Deep learning can enhance and increase mass glaucoma screening. In this study, domain transfer learning technique is important to better weight initialization and for understanding features more related to the problem. For this, classic convolutional neural networks, such as ResNet50 will be compared with Vision Transformers, in high and low-resolution images. The high-resolution retinal image will be used to pre-trained the network and use that knowledge for detecting glaucoma in retinal images captured by mobile devices. The ResNet50 model reached the highest values of AUC in the high-resolution dataset, being the more consistent model in all the experiments. However, the Vision Transformer proved to be a promising technique, especially in low-resolution retinal images. © 2023, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.