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
Class imbalance is a common issue while developing classification models. In order to tackle this problem, synthetic data have recently been developed to enhance the minority class. These artificially generated samples aim to bolster the representation of the minority class. However, evaluating the suitability of such generated data is crucial to ensure their alignment with the original data distribution. Utility measures come into play here to quantify how similar the distribution of the generated data is to the original one. For tabular data, there are various evaluation methods that assess different characteristics of the generated data. In this study, we collected utility measures and categorized them based on the type of analysis they performed. We then applied these measures to synthetic data generated from two well-known datasets, Adults Income, and Liar+. We also used five well-known generative models, Borderline SMOTE, DataSynthesizer, CTGAN, CopulaGAN, and REaLTabFormer, to generate the synthetic data and evaluated its quality using the utility measures. The measurements have proven to be informative, indicating that if one synthetic dataset is superior to another in terms of utility measures, it will be more effective as an augmentation for the minority class when performing classification tasks.

Abstract
This paper examines the idea of Industry 4.0 from the perspective of the molds industry, a vital industry in today's industrial panorama. Several technologies, particularly in the area of machining equipment, have been introduced as a result of the industry's constant modernization. This technological diversity makes automatic interconnection with production management software extremely difficult, as each brand and model requires different, mostly proprietary, interfaces and communication protocols. In the methodology presented in this paper, a development of monitoring solutions for machining devices is defined supporting the leading equipment and operations used by molds industry companies. OPC UA is employed for high-level communication between the various systems for a standardized approach. The approach combines various machine interfaces on a single system to cover a significant subset of machining equipment currently used by the molds industry, as a key result of this paper and given the variety of monitoring systems and communication protocols. This type of all-in-one approach will provide production managers with the information they need to monitor and improve the complete manufacturing process.

Abstract
In many real-world prediction tasks, the class labels contain information about the relative order between the labels that are not captured by commonly used loss functions such as multicategory cross-entropy. In ordinal regression, many works have incorporated ordinality into models and loss functions by promoting unimodality of the probability output. However, current approaches are based on heuristics, particularly non-parametric ones, which are still insufficiently explored in the literature. We analyze the set of unimodal distributions in the probability simplex, establishing fundamental properties and giving new perspectives to understand the ordinal regression problem. Two contributions are then proposed to incorporate the preference for unimodal distributions into the predictive model: 1) UnimodalNet, a new architecture that by construction ensures the output is a unimodal distribution, and 2) Wasserstein Regularization, a new loss term that relies on the notion of projection in a set to promote unimodality. Experiments show that the new architecture achieves top performance, while the proposed new loss term is very competitive while maintaining high unimodality.

Abstract
Deep learning approaches require a large amount of data to be transferred to centralized entities. However, this is often not a feasible option in healthcare, as it raises privacy concerns over sharing sensitive information. Federated Learning (FL) aims to address this issue by allowing machine learning without transferring the data to a centralized entity. FL has shown great potential to ensure privacy in digital healthcare while maintaining performance. Despite this, there is a lack of research on the impact of different types of data heterogeneity on the results. In this study, we research the robustness of various FL strategies on different data distributions and data quality for glaucoma diagnosis using retinal fundus images. We use RetinaQualEvaluator to generate quality labels for the datasets and then a data distributor to achieve our desired distributions. Finally, we evaluate the performance of the different strategies on local data and an independent test dataset. We observe that federated learning shows the potential to enable high-performance models without compromising sensitive data. Furthermore, we infer that FedProx is more suitable to scenarios where the distributions and quality of the data of the participating clients is diverse with less communication cost.