Abstract
The learning from continuous data streams is a relevant area within machine learning, focusing on the creation and updating of predictive models in real time as new data becomes available for training and prediction. Among the most widely used methods for this type of task, Hoeffding Trees are highly valued for their simplicity and robustness across a variety of applications and are considered the primary choice for generating decision trees in data stream contexts. However, Hoeffding Trees tend to continuously expand as new data is incorporated, resulting in increased processing time and memory consumption, often without providing significant gains in accuracy. In this study, we propose an instance selection scheme that combines different strategies to regularize Hoeffding Trees and their variants, mitigating excessive growth without compromising model accuracy. The method selects misclassified instances and a fraction of correctly classified instances during the training phase. After extensive experimental evaluation, the instance selection scheme demonstrates superior predictive performance compared to the original models (without selection), for both real and synthetic datasets for data streams, using a reduced subset of examples. Additionally, the method achieves relevant improvements in processing time, model complexity, and memory consumption, highlighting the effectiveness of the proposed instance selection scheme.

Abstract
PurposeThis paper aims to identify the critical success factors (CSFs) for managing remote project teams (RPT) within project environments. In other words, it focuses on identifying the crucial elements for the success of projects executed by RPT.Design/methodology/approachAn exploratory mixed-method was used combining a case study approach with the application of surveys. Document analysis and direct observation were also applied. The analyzed company is a well-known project-based company acting in the coffee industry and is justified due to its multilocation and multicultural perspectives.FindingsThrough an initial literature review, 93 CSFs were identified and then organized into 7 categories. The subsequent phase involved the relevance evaluation of the identified CSFs through surveys conducted in an international company. The first results analysis identified 20 CSFs. A deeper analysis identified the most relevant factors for each category (Project Managers, 33 factors; Team Leaders, 15; and Team Members, 29). Combining these results, 11 CSFs were identified.Originality/valueWith the trend of remote work that is being kept after the pandemic, this study contributes to identify the most relevant issues that must be taken into account in managing remote teams. By identifying those issues, or CSFs, managers and team members might focus on the most relevant factors.

Abstract
Inherited retinal diseases (IRDs) are rare and genetically diverse disorders that cause progressive vision loss and affect 1 in 3000 individuals worldwide. Their rarity and genetic variability pose a challenge for deep learning models due to the limited amount of data. Generative models offer a promising solution by creating synthetic data to improve training datasets. This study carried out a systematic literature review to investigate the use of generative models to augment data in IRDs and assess their impact on the performance of classifiers for these diseases. Following PRISMA 2020 guidelines, searches in four databases identified 32 relevant studies, 2 focused on IRD and the rest on other retinal diseases. The results indicate that generative models effectively augment small datasets. Among the techniques identified, Deep Convolutional Adversarial Generative Networks (DCGAN) and the Style-Based Generator Architecture of Generative Adversarial Networks 2 (StyleGAN2) were the most widely used. These architectures generated highly realistic and diverse synthetic data, often indistinguishable from real data, even for experts. The results highlight the need for more research into data generation in IRD to develop robust diagnostic tools and improve genetic studies by creating more comprehensive genetic repositories.

Abstract
Polarization optical fiber sensors are based on modifications of fiber birefringence by an external measurand (e.g., strain, pressure, acoustic waves). Yet, this means that different input states of polarization will result in very distinct behaviors, which may or may not be optimal in terms of sensitivity and signal-to-noise ratio. To tackle this challenge, this manuscript presents an optimization technique for the input polarization state using the Fisher information formalism, which allows for achieving maximal precision for a statistically unbiased metric. By first measuring the variation of the Mueller matrix of the optical fiber in response to controlled acoustic perturbations induced by piezo speakers, we compute the corresponding Fisher information operator. Using maximal information states of the Fisher information, it was possible to observe a significant improvement in the performance of the sensor, increasing the signal-to-noise ratio from 4.3 to 37.6 dB, attaining an almost flat response from 1.5 kHz up to 15 kHz. As a proof-of-concept for dynamic audio signal detection, a broadband acoustic signal was also reconstructed with significant gain, demonstrating the usefulness of the introduced formalism for high-precision sensing with polarimetric fiber sensors. (c) 2025 Optica Publishing Group. All rights, including for text and data mining (TDM), Artificial Intelligence (AI) training, and similar technologies, are reserved.

Abstract

Abstract
Industry 4.0 is transforming manufacturing environments, with robotics being a key technology that enhances various capabilities. The flexibility of Autonomous Mobile Robots has led to the rise of multi-robot systems in industrial settings. Considering the high cost of these robots, it is essential to determine the best fit of number and type before making any major investments. Simulation and modeling are valuable decision-support tools, allowing the simulation of different setups to address robot fleet sizing issues. This paper introduces a decision-support framework that combines a fleet manager software stack with the FlexSim simulator, helping decision-makers determine the most suitable mobile robots fleet size tailored to their needs. Unlike previous approaches, the developed solution integrates the same real robot coordination software in both simulation and actual deployment, ensuring that tested scenarios accurately reflect real-world conditions. A case study was conducted to evaluate the framework, involving multiple tasks of loading and unloading materials within a warehouse. Five different scenarios with varying fleet sizes were simulated, and their performances assessed. The analysis concluded that, for the case study under consideration, a fleet of three robots was the most suitable, considering relevant key performance indicators. The results confirmed that the developed solution is an effective alternative for addressing the problem and represents a novel technology with no prior state-of-the-art equivalents.