Abstract
Hyper-parameter tuning of machine learning models has become a crucial task in achieving optimal results in terms of performance. Several researchers have explored the optimisation task during the last decades to reach a state-of-the-art method. However, most of them focus on batch or offline learning, where data distributions do not change arbitrarily over time. On the other hand, dealing with data streams and online learning is a challenging problem. In fact, the higher the technology goes, the greater the importance of sophisticated techniques to process these data streams. Thus, improving hyper-parameter self-tuning during online learning of these machine learning models is crucial. To this end, in this paper, we present MESSPT, an evolutionary algorithm for self-hyper-parameter tuning for data streams. We apply Differential Evolution to dynamically-sized samples, requiring a single pass-over of data to train and evaluate models and choose the best configurations. We take care of the number of configurations to be evaluated, which necessarily has to be reduced, thus making this evolutionary approach a micro-evolutionary one. Furthermore, we control how our evolutionary algorithm deals with concept drift. Experiments on different learning tasks and over well-known datasets show that our proposed MESSPT outperforms the state-of-the-art on hyper-parameter tuning for data streams.

Abstract
Autism Spectrum Disorder (ASD) affects individuals in diverse ways, making personalized therapeutic approaches crucial. In this context, we propose a personalized mobile application designed for music-based therapy tailored to people with ASD. This adaptive piano app can be customized to suit the individual abilities of each user. The paper is structured as follows: The introduction provides context on autism and the importance of personalized therapy. The background section reviews related studies on music-based therapy. The methodology section introduces Professor Piano, our adaptive and adaptable music therapy application. The results and discussion section explores the challenges encountered during development and presents the findings from a heuristic evaluation conducted by experts. Finally, the conclusion summarizes the main insights and implications of the study.

Abstract
Immersive technologies, such as virtual reality, augmented reality, and mixed reality have gained increasing interest and usage in the field of education. Attention is being paid to their effects on teaching and learning processes, one of which is self-regulation of learning, with an important role in supporting learning success. However, designing and creating immersive environments that support the development of SRL strategies is challenging. Employing a systematic approach, this literature review provides an overview of the uses of virtual, augmented, and mixed reality with the goal of supporting SRL. We map these to known educational uses of immersive environments, highlighting current gaps in these efforts and suggesting pathways for future studies on instructional design of the use of immersive technologies to support self-regulation of learning.

Abstract
Water is at the core of sustainable development, and its use for human activities, including vehicle washing, should be done in a sustainable way. There are several technical solutions for washing buses offering different performances, making it difficult to choose the one that best meets the requirements of each specific case. The literature on the topic hardly analyzes the choice of the best technical solution for washing buses and does not apply and compare the results of different multicriteria decision-making (MCDM) methods for the problem. The unique information available is from the different suppliers in the market. Whereby, this work intends to give a technical-scientific contribution to fulfill this gaps. Therefore, the main objectives of this work are (1) to select the best sustainable technical solutions for washing buses depending on the specific conditions for a case study and (2) to analyze how different multicriteria decision-making methods behave in the selection process. To achieve these objectives, the problem was approached as a case study in a public transport company in Portugal and the methodology followed the next steps: started with the identification of the different types of commercial technical solutions for washing buses; the company's experts selected four main criteria: water consumption, operating costs, quality of washing, and time spent; the criteria weights were determined using the fuzzy-AHP method; then four representative MCDM methods were selected, namely, AHP, ELECTRE, TOPSIS, and SMART; the ranks obtained for the four methods were compared; and a sensitivity analysis was performed. Considering the input data for the criteria and their weights, the results for all the methods showed that the best and the worst solution was the same, mobile portico with a brush and porticoes with three brushes, respectively. Furthermore, the results of the sensitivity analysis performed with disturbances for the weights of each criterion presented that the results are slightly affected and the similarity in rankings for the four MCDM methods was validated by Spearman's rank correlation coefficient (rs) and Kendall's coefficient of concordance (W). Considering these results, the SMART method, the less complex one, showed no difference from the others. For that reason, simple methods, such as SMART, in line with other works in the literature perform well in most cases. As a final remark of this work, it can be said that the methodology employed in this project can also be deemed applicable to other similar companies seeking technical solutions for bus or truck washing. Furthermore, the application of the SMART method, the less complex one and the most understandable for people, showed no difference from the others, being able to be applied in similar situations.

Abstract
Traditional recurrent neural networks (RNNs) are essential for processing time-series data. However, they function as read-only models, lacking the ability to directly modify the data they learn from. In this study, we introduce the corrector long short-term memory (cLSTM), a Read & Write LSTM architecture that not only learns from the data but also dynamically adjusts it when necessary. The cLSTM model leverages two key components: (a) predicting LSTM’s cell states using Seasonal Autoregressive Integrated Moving Average (SARIMA) and (b) refining the training data based on discrepancies between actual and forecasted cell states. Our empirical validation demonstrates that cLSTM surpasses read-only LSTM models in forecasting accuracy across the Numenta Anomaly Benchmark (NAB) and M4 Competition datasets. Additionally, cLSTM exhibits superior performance in anomaly detection compared to hierarchical temporal memory (HTM) models.

Abstract
All types of sports are potential application scenarios for automatic and real-time visual object and event detection. In rink hockey, the popular roller skate variant of team hockey, it is of great interest to automatically track player movements, positions, and sticks, and also to make other judgments, such as being able to locate the ball. In this work, we present a real-time pipeline consisting of an object detection model specifically designed for rink hockey games, followed by a knowledge-based event detection module. Even in the presence of occlusions and fast movements, our deep learning object detection model effectively identifies and tracks important visual elements in real time, such as: ball, players, sticks, referees, crowd, goalkeeper, and goal. Using a curated dataset consisting of a collection of rink hockey videos containing 2525 annotated frames, we trained and evaluated the algorithm's performance and compared it to state-of-the-art object detection techniques. Our object detection model, based on YOLOv7, presents a global accuracy of 80% and, according to our results, good performance in terms of accuracy and speed, making it a good choice for rink hockey applications. In our initial tests, the event detection module successfully detected an important event type in rink hockey games, namely, the occurrence of penalties.