Abstract
As distributed and multi-organization Machine Learning emerges, new challenges must be solved, such as diverse and low-quality data or real-time delivery. In this paper, we use a distributed learning environment to analyze the relationship between block size, parallelism, and predictor quality. Specifically, the goal is to find the optimum block size and the best heuristic to create distributed Ensembles. We evaluated three different heuristics and five block sizes on four publicly available datasets. Results show that using fewer but better base models matches or outperforms a standard Random Forest, and that 32 MB is the best block size.

Abstract
The high rates of mortality provoked by cardiovascular disorders (CVDs) have been rated by the OMS in the top among non-communicable diseases, killing about 18 million people annually. It is crucial to detect arrhythmias or cardiovascular events in an early way. For that purpose, novel portable acquisition devices have allowed long-term electrocardiographic (ECG) recording, being the most common way to discover arrhythmias of a random nature such as atrial fibrillation (AF). Nonetheless, the acquisition environment can distort or even destroy the ECG recordings, hindering the proper diagnosis of CVDs. Thus, it is necessary to assess the ECG signal quality in an automatic way. The proposed approach exploits the feature and meta-feature extraction of 5-s ECG segments with the ability of machine learning classifiers to discern between high- and low-quality ECG segments. Three different approaches were tested, reaching values of accuracy close to 83% using the original feature set and improving up to 90% when all the available meta-features were utilized. Moreover, within the high-quality group, the segments belonging to the AF class outperformed around 7% until a rate over 85% when the meta-features set was used. The extraction of meta-features improves the accuracy even when a subset of meta-features is selected from the whole set.

Abstract
As LLMs gain an increasingly relevant role and agency, their alignment with human values, principles and goals is crucial for their responsible deployment and acceptance. The main goal of this study is to assess the alignment of different LLMs regarding the relative importance of AI Ethics principles across different domains. To this end, human experts in different domains were asked, through a questionnaire, to rate the relative importance of six AI Ethics principles in their respective domains, totaling 6 domains. Then, five publicly available LLMs were asked to rate the same Ethics principles in different domains. Multiple prompts were used multiple times, to also evaluate consistency, totaling 90 runs per LLM. Model alignment was measured through the correlation with human experts, and consistency was evaluated through the standard deviation. Results show varying degrees of alignment and consistency, with a couple of models showing satisfactory results. This makes it possible to envisage the use of such models to automatically configure and adapt data pipeline ecosystems and architectures across different domains, selecting processes, dashboard elements or monitored KPIs according to the target domain or the goals of the system.

Abstract
Many of today’s domains of application of Machine Learning (ML) are dynamic in the sense that data and their patterns change over time. This has a significant impact in the ML lifecycle and operations, requiring frequent model (re-)training, or other strategies to deal with outdated models and data. This need for dynamic and responsive solutions also has an impact on the use of computational resources and, consequently, on sustainability indicators. This paper proposes an approach in line with the concept of Frugal AI, whose main aim is to minimize the resources and time spent on training models by re-using models from a pool of past models, when appropriate. Specifically, we present and validate a methodology for similarity-based model selection in data streaming environments with concept drift. Rather than training a new model for each new block of data, this methodology considers a pool with only a subset of the models and, for each new block of data, will select the best model from the pool. The best model is determined based on the distance between its training data and the current block of data. Distance is calculated based on a set of meta-features that characterizes the data, and on the Bray-Curtis distance. We show that it is possible to reuse previous models using this methodology, leading to potentially significant saving of resources and time, while maintaining predictive quality. © The Author(s), under exclusive license to Springer Nature Switzerland AG 2025.

Abstract
The footwear industry is energy-intensive and, consequently, a source of large amounts of greenhouse gas emissions every year. Issues related to climate change and growing conflicts on a global scale that impact the prices of raw materials and energy prices have led companies in the footwear industry to take actions to mitigate these impacts. Among these actions is the growing focus on producing its energy from energy systems based on renewable sources and battery energy storage units. This paper addresses the energy-efficient manufacturing scheduling in footwear industries with onsite energy production from a photovoltaic system with batteries. The problem is formulated as a mixed integer linear programming problem. Different objectives are presented, depending on the priorities of the entity that owns the footwear factory, namely, minimizing operation costs, minimizing CO2 emissions, or both. The case study is footwear factory located in Portugal that uses a manufacturing process based on injection molding. The results show the effectiveness of the proposed approach, with active demand side management playing a fundamental role in shifting periods of higher energy consumption to periods of lower prices or lower CO2 emissions. Also, Pareto fronts are depicted to make the trade-off between CO2 emissions and operation costs. As expected, the reduction of CO2 emissions promotes an increase on operation costs. Furthermore, a sensitivity analysis is carried out on the increase in photovoltaic capacity and battery capacity. The results show that increasing photovoltaic and battery capacity promotes reductions in costs up to 30% and in the emissions up to 37%.

Abstract
Lean and green (L&G) manufacturing in Industry 4.0 (I4.0) has brought many advantages in manufacturing industries by minimizing waste and maximizing efficiency with integration of renewable energy sources and sustainable materials. Multi-layer Stream Mapping (MSM) is a new framework for the performance assessment of complex manufacturing processes. MSM is used for multi-domain analysis of manufacturing processes to assess resources, and processes, that are used to identify Non-ValueAdded (NVA) procedures or steps that consume unnecessary time and resources, and/or release emissions and waste that can no longer be reused or recycled to be eliminated or replaced to create a Value Added (VA) process flow that avoids waste in a clean, green and environmental friendly manner. This paper presents the implementation of the L&G strategy through MSM in metal working production systems. In metalworking production systems, the variables of operational performance and resources consumption considered are process time, number of operators, consumables, raw material, and energy. These can be suitably used for reduction in water emissions, gas emissions, solid waste and scrap generated in metalworking production systems.