Abstract
In a time in which streaming data becomes the new normal in Machine Learning problems, to the detriment of batch data, new challenges arise. In the past, a data source would be static in the sense that all data were known at the moment of the training of the model. A model would be trained and it would be in use for relatively long periods of time. Nowadays, data arrive in real-time and their statistical properties may also change over time, rendering trained models outdated. In this paper we propose an approach to deal with the concept drift problem with minimal computational effort. Specifically, we continuously update an ensemble with new weak learners and adjust their weights according to their performance. This approach is suitable to be used in real-time in the form of an ever-evolving model that adapts to change in the data.

Abstract
Machine Learning has emerged in the last years as the main solution to many of nowadays' data-based decision problems. However, while new and more powerful algorithms and the increasing availability of computational resources contributed to a widespread use of Machine Learning, significant challenges still remain. Two of the most significant nowadays are the need to explain a model's predictions, and the significant costs of training and re-training models, especially with large datasets or in streaming scenarios. In this paper we address both issues by proposing an approach we deem predictive and user-centric. It is predictive in the sense that it estimates the benefit of re-training a model with new data, and it is user centric in the sense that it implements an explainable interface that produces interpretable explanations that accompany predictions. The former allows to reduce necessary resources (e.g. time, costs) spent on re-training models when no improvements are expected, while the latter allows for human users to have additional information to support decision-making. We validate the proposed approach with a group of public datasets and present a real application scenario.

Abstract

Abstract
The way students learn changed significantly over the past two years, due to the current pandemic. However, this change was neither desired not planed beforehand. As a result, in many cases, it may have been undertaken without the appropriate care. In this paper we propose a framework for online education tailored for Computer Science degrees. Its goals are twofold: to avoid disruptive changes by providing a familiar and supportive structure for teaching/learning activities, and to motivate Students to learn autonomously, despite their reduced contact with their peers or the Teacher.

Abstract
Artificial Intelligence (AI), in all its different sub-fields, has grown significantly over the past years. When compared with other scientific or technological fields, this can almost be seen as a revolution. Nonetheless, as in other revolutions, not all that revolves around AI evolved at the same pace. As a consequence, many serious legal and ethical issues on the use of Artificial Intelligence are presently being raised. This paper addresses the main root causes for these problems from a technical standpoint, and then analyzes the legal and ethical framework. Finally, the paper describes a range of techniques and methods that can be used to address the identified problems, namely by ensuring transparency, fairness, equality, explanability and avoiding bias or discrimination. The field is presently at a tipping point, which can either lead to an avoidance of Artificial Intelligence due to fear or lack of regulation, or to a wide adoption supported by increased transparency and more human-centered approaches. Given the recent developments addressed in this paper, the paper argues in favor of a tendency towards the latter.

Abstract
Water loss is one of the factors that most affect a concessionaire's financial sustainability. Early detection of any anomaly in water consumption is very valuable. This article aims to carry out a preliminary study to detect change points in consumption associated with water meter malfunction. The dataset is composed of water consumption measurements of two different companies (a hotel and a hospital) located in the north of Portugal, obtained during a complete year. Different methods were implemented in order to study its effectiveness in the detection of change points in the time series related to a sharp decrease in water consumption. Results suggest that the Seasonal Decomposition of Time Series by Loess method (STL) and the combination of several breakpoint detection methods is a suitable approach to be implemented in a software system, in order to help the company in anomaly detection and in the decision-making process of substituting the water meters.