Recently, the scientific community has placed great emphasis on the recognition of human activity, especially in the area of health and care for the elderly. There are already practical applications of activity recognition and unusual conditions that use body sensors such as wrist-worn devices or neck pendants. These relatively simple devices may be prone to errors, might be uncomfortable to wear, might be forgotten or not worn, and are unable to detect more subtle conditions such as incorrect postures. Therefore, other proposed methods are based on the use of images and videos to carry out human activity recognition, even in open spaces and with multiple people. However, the resulting increase in the size and complexity involved when using image data requires the use of the most recent advanced machine learning and deep learning techniques. This paper presents an approach based on deep learning with attention to the recognition of activities from multiple frames. Feature extraction is performed by estimating the pose of the human skeleton, and classification is performed using a neural network based on Bidirectional Encoder Representation of Transformers (BERT). This algorithm was trained with the UP-Fall public dataset, generating more balanced artificial data with a Generative Adversarial Neural network (GAN), and evaluated with real data, outperforming the results of other activity recognition methods using the same dataset.
We present DELIGHT, or Deep Learning Identification of Galaxy Hosts of Transients, a new algorithm designed to automatically and in real time identify the host galaxies of extragalactic transients. The proposed algorithm receives as input compact, multiresolution images centered at the position of a transient candidate and outputs two-dimensional offset vectors that connect the transient with the center of its predicted host. The multiresolution input consists of a set of images with the same number of pixels, but with progressively larger pixel sizes and fields of view. A sample of 16,791 galaxies visually identified by the Automatic Learning for the Rapid Classification of Events broker team was used to train a convolutional neural network regression model. We show that this method is able to correctly identify both relatively large (10″ < r < 60″) and small (r ≤ 10″) apparent size host galaxies using much less information (32 kB) than with a large, single-resolution image (920 kB). The proposed method has fewer catastrophic errors in recovering the position and is more complete and has less contamination (<0.86%) recovering the crossmatched redshift than other state-of-the-art methods. The more efficient representation provided by multiresolution input images could allow for the identification of transient host galaxies in real time, if adopted in alert streams from new generation of large -etendue telescopes such as the Vera C. Rubin Observatory.
Context . Massive stars play a fundamental role in galactic evolution through their strong stellar winds, chemical enrichment, and feedback mechanisms. Accurate modelling of their atmospheres and winds is critical for understanding their physical properties and evolutionary pathways. Traditional spectroscopic analyses often rely on the β -law approximation for wind-velocity profiles, which may not capture the complexity of observed phenomena. Aims . This study aims to introduce and validate the grId of Stellar atmOSphere and hydrodynamiC modELs for massivE Stars (ISOSCELES), a grid-based framework for the quantitative spectroscopic analysis of massive stars. The project leverages hydrodynamic wind solutions derived from the m-CAK theory, including both fast and δ -slow solutions, to improve the accuracy of derived stellar and wind parameters. Methods . We constructed a comprehensive grid of models based on hydrodynamic wind solutions from the HYDWIND code and synthetic spectral line profiles generated by the FASTWIND code. The grid spans a broad parameter space covering OBA-type stars with solar metallicity. A semi-automatic fitting procedure was developed to analyse key spectral lines and derive the stellar and wind parameters. Results . Applying ISOSCELES to six stars demonstrates its ability to reproduce observed spectral profiles with high fidelity. The δ- slow solution proved effective for two early-type B supergiants. The grid also highlights the difference of using the β -law in modelling stellar winds compared with the m-CAK wind solutions. Conclusions . The ISOSCELES database represents a step forward in quantitatively analysing massive stars, offering an alternative to the β -law approximation. Future work will address the inclusion of UV lines and metallicity effects to further refine its applicability across diverse stellar populations.
Context . The advent of large-scale spectroscopic surveys will deliver tens of thousands of spectra of blue massive stars spanning a wide range of spectral types (SpTs) and luminosity classes (LCs). This data volume renders traditional spectral classification techniques increasingly impractical and motivates the development of automated classification tools. Aims . We aim to develop a robust machine learning framework for the automated spectral classification of massive OB-type stars by assessing the performance and evaluating the interpretability of several tree-based ensemble algorithms. Methods . Using a large catalog of high-quality optical spectra of OB-type stars (plus a few A supergiants) compiled within the IACOB project, we designed three hierarchical experiments of increasing complexity: (1) classification into broad O, B, and A SpTs; (2) fine-grained classification within the O and B domains; and (3) joint classification of spectral subtype and LC. We evaluated six algorithms, ranging from single decision trees to ensemble methods (random forest and extra trees) and gradient boosting techniques (XGBoost, LightGBM, and histogram-based gradient boosting). We further analyzed model consensus and probabilistic outputs to assess the relative impact of epistemic and aleatoric uncertainties and employed Shapley additive explanations (SHAP) value analysis to provide an interpretation of the model decisions by identifying the features that drive each classification. Results . The models achieve excellent performance for broad SpT classification when compared with SIMBAD, with LightGBM reaching an accuracy above 98%. For fine-grained spectral subtypes, random forest provides the most robust results (89%). For the most challenging task combining SpT subtypes and LCs, XGBoost reaches a ∼ 77% accuracy. The decrease in performance is primarily driven by intrinsic degeneracies in the classical spectral classification scheme. Our SHAP analysis confirms that the models rely on meaningful spectral features. We also find a high level of agreement among different algorithms, indicating that current performance limits are largely set by the intrinsic complexity of the classification problem rather than by model choice. Conclusions . Tree-based ensemble methods provide a reliable and interpretable framework for the automated spectral classification of massive OB-type stars. The combination of full-spectrum modeling, algorithm comparison, and model interpretability offers a robust approach for future large-scale spectroscopic surveys.