Abstract
The application of pattern recognition techniques to data collected from accelerometers available in off-the-shelf devices, such as smartphones, allows for the automatic recognition of activities of daily living (ADLs). This data can be used later to create systems that monitor the behaviors of their users. The main contribution of this paper is to use artificial neural networks (ANN) for the recognition of ADLs with the data acquired from the sensors available in mobile devices. Firstly, before ANN training, the mobile device is used for data collection. After training, mobile devices are used to apply an ANN previously trained for the ADLs' identification on a less restrictive computational platform. The motivation is to verify whether the overfitting problem can be solved using only the accelerometer data, which also requires less computational resources and reduces the energy expenditure of the mobile device when compared with the use of multiple sensors. This paper presents a method based on ANN for the recognition of a defined set of ADLs. It provides a comparative study of different implementations of ANN to choose the most appropriate method for ADLs identification. The results show the accuracy of 85.89% using deep neural networks (DNN).

Abstract
We study the electromagnetically induced transparency (EIT) phenomenon in a hollow-core fibre filled with rubidium gas. We analyse the impact of the guiding effect and of the temperature on the properties of the EIT phenomenon. The refractive index felt by the probe laser is found to vary due to the radial dependence of the fibre mode field at the pump frequency. Several results are presented for the transmission, dispersion, and group velocity of the probe field, considering both the free propagation regime and the guided propagation along the hollow-core fibre. We note that the EIT occurring in a waveguide has a great potential for practical applications since it can be controlled by adjusting the gas and the fibre properties.

Abstract
The human gastrointestinal tract harbors approximately 100 trillion microorganisms with different microbial compositions across geographic locations. In this work, we used RNASeq data from stomach samples of non-disease (164 individuals from European ancestry) and gastric cancer patients (137 from Europe and Asia) from public databases. Although these data were intended to characterize the human expression profiles, they allowed for a reliable inference of the microbiome composition, as confirmed from measures such as the genus coverage, richness and evenness. The microbiome diversity (weighted UniFrac distances) in gastric cancer mimics host diversity across the world, with European gastric microbiome profiles clustering together, distinct from Asian ones. Despite the confirmed loss of microbiome diversity from a healthy status to a cancer status, the structured profile was still recognized in the disease condition. In concordance with the parallel host-bacteria population structure, we found 16 human loci (non-synonymous variants) in the European-descendent cohorts that were significantly associated with specific genera abundance. These microbiome quantitative trait loci display heterogeneity between population groups, being mainly linked to the immune system or cellular features that may play a role in enabling microbe colonization and inflammation.

Abstract
Despite recent efforts, accuracy in group emotion recognition is still generally low. One of the reasons for these underwhelming performance levels is the scarcity of available labeled data which, like the literature approaches, is mainly focused on still images. In this work, we address this problem by adapting an inflated ResNet-50 pretrained for a similar task, activity recognition, where large labeled video datasets are available. Audio information is processed using a Bidirectional Long Short-Term Memory (Bi-LSTM) network receiving extracted features. A multimodal approach fuses audio and video information at the score level using a support vector machine classifier. Evaluation with data from the EmotiW 2020 AV Group-Level Emotion sub-challenge shows a final test accuracy of 65.74% for the multimodal approach, approximately 18% higher than the official baseline. The results show that using activity recognition pretraining offers performance advantages for group-emotion recognition and that audio is essential to improve the accuracy and robustness of video-based recognition.

Abstract
Background: Suffering in nurses is associated with the delivery of care to patients in suffering and factors related to the working conditions. It is a multidimensional experience that occurs in situations of loss, damage, or threat to human integrity. Objective: To compare the mean scores in the dimensions of suffering (Emotional Pain, Relational Loss, and Avoidance) based on the sociodemographic and professional variables of nurses. Methodology: A descriptive and cross-sectional study with a quantitative approach was conducted with a sample of 100 nurses. A self-administered questionnaire was applied, as well as the Caregiver Grief Scale for assessing suffering. Results: Women with children, with a partner, without specialization in nursing, and with more years of service had higher mean scores of suffering. In men, the highest mean scores were found in nurses without children, without a partner, with specialization in nursing, and with more years of service. Conclusion: Nurses showed higher mean scores of suffering in the dimension of Emotional Pain, followed by Relational Loss, and Avoidance, and suffering was higher among women.

Abstract
One of the main goals of any power grid is sustainability. The given study proposes a new method, which aims to reduce users' anxiety especially at slow charging stations and improve the smart charging model to increase the benefits for the electric vehicles' owners, which in turn will increase the grid stability. The issue under consideration is modelled as an optimisation problem to minimise the cost of charging. This approach levels the load effectively throughout the day by providing power to charge EVs' batteries during the off-peak hours and drawing it from the EVs' batteries during peak-demand hours of the day. In order to minimise the costs associated with EVs' charging in the given optimisation problem, an improved version of an intelligent algorithm is developed. In order to evaluate the effectiveness of the proposed technique, it is implemented on several standard models with various loads, as well as compared with other optimisation methods. The superiority and efficiency of the proposed method are demonstrated, by analysing the obtained results and comparing them with the ones produced by the competitor techniques.