Abstract
Communication between Deaf and hearing individuals remains a persistent challenge requiring attention to foster inclusivity. Despite notable efforts in the development of digital solutions for sign language recognition (SLR), several issues persist, such as cross-platform interoperability and strategies for tokenizing signs to enable continuous conversations and coherent sentence construction. To address such issues, this paper proposes a non-invasive Portuguese Sign Language (Lingua Gestual Portuguesa or LGP) interpretation system-as-a-service, leveraging skeletal posture sequence inference powered by long-short term memory (LSTM) architectures. To address the scarcity of examples during machine learning (ML) model training, dataset augmentation strategies are explored. Additionally, a buffer-based interaction technique is introduced to facilitate LGP terms tokenization. This technique provides real-time feedback to users, allowing them to gauge the time remaining to complete a sign, which aids in the construction of grammatically coherent sentences based on inferred terms/words. To support human-like conditioning rules for interpretation, a large language model (LLM) service is integrated. Experiments reveal that LSTM-based neural networks, trained with 50 LGP terms and subjected to data augmentation, achieved accuracy levels ranging from 80% to 95.6%. Users unanimously reported a high level of intuition when using the buffer-based interaction strategy for terms/words tokenization. Furthermore, tests with an LLM-specifically ChatGPT-demonstrated promising semantic correlation rates in generated sentences, comparable to expected sentences.

Abstract

Abstract
Our study provides valuable insights into the relationship between artificial intelligence (AI) and Human Resource Management (HRM). We have minimised bias and ensured reliable findings by employing a systematic literature review and the PRISMA statement. Our comprehensive syn-thesis of the studies included in this research, along with a bibliometric analysis of articles, journals, indexes, authors' affiliations, citations, keyword co-occurrences, and co-authorship analysis, has produced robust results. The discussion of our findings focuses on critical areas of interest, such as AI and Talent, AI Bias, Ethics and Law, and their impact on Human Resource (HR) management. Our research highlights the recognition by organisations of the importance of talent management in achieving a competitive advantage as higher-level skills become increas-ingly necessary. Although some HR managers have adopted AI technology for talent acquisition, our study reveals that there is still room for improvement. Our study is in line with previous research that acknowledges the potential for AI to revolutionise HR management and the future of work. Our findings emphasise the need for HR managers to be proactive in embracing technology and bridging the technological, human, societal, and governmental gaps. Our study contributes to the growing body of AI and HR management knowledge, providing essential insights and rec-ommendations for future research. The importance of our study lies in its focus on the role of HR in promoting the benefits of AI-based applications, thereby creating a larger body of knowledge from an organisational perspective.

Abstract
Water is a fundamental human resource and its scarcity is reflected in social, economic and environmental problems. Water used in human activities must be treated before reusing or returning to nature. This treatment takes place in wastewater treatment plants (WWTPs), which need to perform their functions with high quality, low cost, and reduced environmental impact. This paper aims to identify failures in real-time, using streaming data to provide the necessary preventive actions to minimize damage to WWTPs, heavy fines and, ultimately, environmental hazards. Convolutional and Long short-term memory (LSTM) autoencoders (AEs) were used to identify failures in the functioning of the dissolved oxygen sensor used in WWTPs. Five faults were considered (drift, bias, precision degradation, spike and stuck) in three different scenarios with variations in the appearance order, intensity and duration of the faults. The best performance, considering different model configurations, was achieved by Convolutional-AE.

Abstract
O presente artigo propõe-se discutir práticas pedagógicas que promovam a inventividade, a cocriação e a colaboração e tem por objetivo apresentar a ação “Tô Ligado!”, que emergiu da vivência de cidadania digital MOVEOnCibricity no contexto do I Festival Internacional de Cidadania Digital, desenvolvida pela Rede Internacional ConectaKaT. A pesquisa se desenvolve a partir do Método Cartográfico de Pesquisa-Intervenção e está fundamentada na perspectiva da Educação OnLIFE. Tem como resultados o desenvolvimento de outras ações, constituindo a característica reticular e conectiva da ConectaKaT, sendo eles: a WebSérie Entrevistas, a Campanha/Jogo Segurança na Internet e nova WebSérie TomKaT nas Escolas. Esta ação apresenta contribuições significativas na construção de práticas pedagógicas inventivas e cocriadas na perspectiva de uma Educação OnLIFE, contribuindo também para um habitar engajado no ensinar e no aprender. Além disso, a ação “Tô Ligado!” se configurou como uma potente prática na construção de conhecimentos em torno de temáticas que se relacionam ao desenvolvimento sustentável. Assim, a ConectaKaT tem se estruturado enquanto uma plataforma viva de interação ecológica, propícia à invenção de novas metodologias e práticas pedagógicas, evidenciando uma nova política cognitiva em Educação.

Abstract
Vine phenology modelling is increasingly important for winegrowers and viticulturists. Model calibration is often required before practical applications. However, when multiple models and optimization methods are applied for different varieties, it is rarely known which factor tends to mostly affect the calibration results. We mainly aim to investigate the main source of the variability in the modelling errors for the flowering timings of two important varieties of vine in the Douro Demarcated Region (DDR) of Portugal; this is based on five phenology model simulations that use optimal parameters and that are estimated by three optimization algorithms (MLE, SA and SCE-UA). Our results indicate that the main source of the variability in calibration can be affected by the initially assumed parameter boundary. Restricting the initial parameter distribution to a narrow range impedes the algorithm from exploring the full parameter space and searching for optimal parameters. This can lead to the largest variation in different models. At an identified appropriate boundary, the difference between the two varieties represents the largest source of uncertainty, while the choice of algorithm for calibration contributes least to the overall uncertainty. The smaller variability among different models or algorithms (tools for analysis) compared to between different varieties could indicate the overall reliability of the calibration. All optimization algorithms show similar results in terms of the obtained goodness-of-fit: the RMSE (MAE) is 5-6 (4-5) days with a negligible mean bias and moderately good R-2 (0.5-0.6) for the ensemble median predictor. Nevertheless, a similar predictive performance can result from differently estimated parameter values, due to the equifinality or multi-modal issue in which different parameter combinations give similar results. This mainly occurs for models with a non-linear structure compared to those with a near-linear one. Yet, the former models are found to outperform the latter ones in predicting the flowering timing of the two varieties in the DDR. Overall, our findings highlight the importance of carefully defining the initial parameter boundary and decomposing the total variance of prediction errors. This study is expected to bring new insights that will help to better inform users about the importance of choice when these factors are involved in calibration. Nonetheless, the importance of each factor can change depending on the specific situation. Details of how the optimization methods are applied and of the continuous model improvement are important.