A design study of a low-cost heterogeneous parallel computer based on already existing hardware is presented. The UNIX operating system is used together with PVM for message passing control. The parallel machine is constructed using existing Intel PC's and Alpha workstation platforms. The new techniques and algorithms are successfully applied to the grid generation problem based on the solution of partial differential equations with control of the boundary point distribution. Two different programming models were used. One uses decomposition of the domain into as many parts as there are processors, and revealed high reductions on the total turnaround time, although load balancing was a major problem. The other programming model has the possibility of changing the position of the points of the interface between each sub-grid, and allows reductions of the total turnaround time up to 1.6 times for the case of 3 processors, while the quality characteristics of the final grid are maintained. The communication times were reduced up to 7.5% of the total turnaround times, but the recursive nature of the numerical iterative process of the grid generation method causes a substantial increase in the number of iterations, reducing the parallel machine performance.

O crescente uso de sistemas de Inteligência Artificial (IA) nas redações tem promovido novas formas de fazer jornalismo, assim como questionamentos e preocupações sobre as suas implicações na prática jornalística. Estas tecnologias são aplicadas nas várias etapas do processo jornalístico, desde a recolha, à produção e à distribuição nas plataformas digitais. Um dos subcampos da IA é o Processamento de Linguagem Natural (PLN), uma técnica utilizada na produção automática de texto. Esta tecnologia encontra-se em expansão no jornalismo, dando lugar à prática do jornalismo semiautomático, ou seja, aquele que utiliza sistemas computacionais ou algoritmos para produzir notícias automaticamente, com uma intervenção humana residual. A aplicação desta tecnologia tornou-se possível devido a um conjunto de fatores, entre eles, o desenvolvimento da capacidade de armazenamento de informação e a abundância de dados estruturados que se encontram acessíveis na Internet. Os sistemas de automação de texto têm sido aplicados maioritariamente em assuntos que envolvam dados estatísticos estruturados, entre eles: resultados desportivos, oscilações nas bolsas de valores, alertas meteorológicos e resultados eleitorais. Os textos produzidos de forma semiautomática obedecem a modelos textuais que têm como base notícias escritas por jornalistas, com uma estrutura frásica simples e repetitiva. Em Portugal, a produção semiautomática de texto nos meios de comunicação encontra-se ainda numa fase incipiente. O primeiro caso de uso regular desta tecnologia é o jornal digital desportivo ZeroZero, que desenvolveu internamente o sistema Prosebot, lançado em 2021. Este algoritmo destina-se a escrever sínteses de várias competições de futebol, em Portugal e noutros países, escrevendo em várias línguas para além do português. O software elabora sínteses instantâneas sobre cada jogo com recurso a uma base de dados do próprio jornal, que é continuamente alimentada com as informações de cada jogo. O Prosebot dá resposta a um fluxo contínuo de jogos que acontecem todos os fins de semana, em campeonatos nacionais e distritais. Esta estratégia inovadora contribui para uma disrupção no sistema informativo e um conjunto de desafios económicos, profissionais e éticos. Como tal, torna-se pertinente questionar: De que forma a produção semiautomática de texto influencia os profissionais de redação, as práticas jornalísticas e a organização noticiosa? Com base neste problema de pesquisa, o principal objetivo é compreender como é que a produção semiautomática de notícias transforma o trabalhos nas redações. Para concretizar este objetivo, a investigação estuda o caso do Prosebot/ZeroZero a partir de uma abordagem qualitativa, com entrevistas semiestruturadas e um grupo focal com os profissionais da redação.

An increase in the accuracy of identification of Activities of Daily Living (ADL) is very important for different goals of Enhanced Living Environments and for Ambient Assisted Living (AAL) tasks. This increase may be achieved through identification of the surrounding environment. Although this is usually used to identify the location, ADL recognition can be improved with the identification of the sound in that particular environment. This paper reviews audio fingerprinting techniques that can be used with the acoustic data acquired from mobile devices. A comprehensive literature search was conducted in order to identify relevant English language works aimed at the identification of the environment of ADLs using data acquired with mobile devices, published between 2002 and 2017. In total, 40 studies were analyzed and selected from 115 citations. The results highlight several audio fingerprinting techniques, including Modified discrete cosine transform (MDCT), Mel-frequency cepstrum coefficients (MFCC), Principal Component Analysis (PCA), Fast Fourier Transform (FFT), Gaussian mixture models (GMM), likelihood estimation, logarithmic moduled complex lapped transform (LMCLT), support vector machine (SVM), constant Q transform (CQT), symmetric pairwise boosting (SPB), Philips robust hash (PRH), linear discriminant analysis (LDA) and discrete cosine transform (DCT).

The biophysical determinants related to swimming performance are one of the most attractive topics within swimming science. The aim of this paper was to do an update of the "state of art" about the interplay between performance, energetic and biomechanics in competitive swimming. Throughout the manuscript some recent highlights are described: (i) the relationship between swimmer's segmental kinematics (segmental velocities, stroke length, stroke frequency, stroke index and coordination index) and his center of mass kinematics (swimming velocity and speed fluctuation); (ii) the relationships between energetic (energy expenditure and energy cost) and swimmer's kinematics; and (iii) the prediction of swimming performance derived from above mentioned parameters.

Breast ultrasound images have several attractive properties that make them an interesting tool in breast cancer detection. However, their intrinsic high noise rate and low contrast turn mass detection and segmentation into a challenging task. In this article, a fully automated two-stage breast mass segmentation approach is proposed. In the initial stage, ultrasound images are segmented using support vector machine or discriminant analysis pixel classification with a multiresolution pixel descriptor. The features are extracted using non-linear diffusion, bandpass filtering and scale-variant mean curvature measures. A set of heuristic rules complement the initial segmentation stage, selecting the region of interest in a fully automated manner. In the second segmentation stage, refined segmentation of the area retrieved in the first stage is attempted, using two different techniques. The AdaBoost algorithm uses a descriptor based on scale-variant curvature measures and non-linear diffusion of the original image at lower scales, to improve the spatial accuracy of the ROI. Active contours use the segmentation results from the first stage as initial contours. Results for both proposed segmentation paths were promising, with normalized Dice similarity coefficients of 0.824 for AdaBoost and 0.813 for active contours. Recall rates were 79.6% for AdaBoost and 77.8% for active contours, whereas the precision rate was 89.3% for both methods.

Sleep apnea syndrome (SAS), which can significantly decrease the quality of life is associated with a major risk factor of health implications such as increased cardiovascular disease, sudden death, depression, irritability, hypertension, and learning difficulties. Thus, it is relevant and timely to present a systematic review describing significant applications in the framework of computational intelligence-based SAS, including its performance, beneficial and challenging effects, and modeling for the decision-making on multiple scenarios.

Este artículo analiza los mecanismos de moderación de contenidos en TikTok y las percepciones de los usuarios portugueses sobre estos procesos. La primera parte revisa la literatura sobre redes sociales en línea y moderación, distinguiendo la intervención humana de la moderación algorítmica y discutiendo las respectivas ventajas, limitaciones e impactos psicosociales. En el componente empírico, se administró un cuestionario en línea a 85 participantes, en su mayoría adultos jóvenes. Los resultados muestran un conocimiento limitado del trabajo de los moderadores y del funcionamiento de los algoritmos. Se concluye que una moderación eficaz en TikTok requiere un equilibrio entre los sistemas automáticos y la revisión humana. Se sugiere que la investigación futura incluya métodos cualitativos para aclarar el impacto de la moderación y mejorar las prácticas de gobernanza en las plataformas digitales.

Inspired by patterns of behavior generated in social networks, a prototype of a new object was designed and developed for the World Wide Web – the stigmergic hyperlink or “stigh”. In a system of stighs, like a Web page, the objects that users do use grow “healthier”, while the unused “weaken”, eventually to the extreme of their “death”, being autopoieticaly replaced by new destinations. At the single Web page scale, these systems perform like recommendation systems and embody an “ecological” treatment to unappreciated links. On the much wider scale of generalized usage, because each stigh has a method to retrieve information about its destination, Web agents in general and search engines in particular, would have the option to delegate the crawling and/or the parsing of the destination. This would be an interesting social change: after becoming not only consumers, but also content producers, Web users would, just by hosting (automatic) stighs, become information service providers too.