Abstract
Chest radiography is one of the most common imaging exams, but its interpretation is often challenging and timeconsuming, which has motivated the development of automated tools for pathology/abnormality detection. Deep learning models trained on large-scale chest X-ray datasets have shown promising results but are highly dependent on the quality of the data. However, these datasets often contain incorrect metadata and non-compliant or corrupted images. These inconsistencies are ultimately incorporated in the training process, impairing the validity of the results. In this study, a novel approach to detect non-compliant images based on deep features extracted from a patient position classification model and a pre-trained VGG16 model are proposed. This method is applied to CheXpert, a widely used public dataset. From a pool of 100 images, it is shown that the deep feature-based methods based on a patient position classification model are able to retrieve a larger number of non-compliant images (up to 81% of non-compliant images), when compared to the same methods but based on a pretrained VGG16 (up to 73%) and the state of the art uncertainty-based method (50%).

Abstract
This paper proposes an interactive approach for narrative extraction from semantic graphs. The proposed approach extracts events from RDF triples, maps them to their corresponding attributes, and assembles them into a chronological sequence to form narrative graphs. The approach is evaluated on the Wikidata graph and achieves promising results in terms of narrative quality and coherence. The paper also discusses several avenues for future work, including the integration of machine learning, graph embedding methods and the exploration of advanced techniques for attention-based narrative labeling and semantic role labeling. Overall, the proposed method offers a promising approach to narrative extraction from semantic graphs and has the potential to be useful in various applications, including chatbots, conversational agents, and content creation tools. © Daniil Lystopadskyi, André Santos, and José Paulo Leal;

Abstract
The Data2MV dataset contains gaze fixation data obtained through experimental procedures from a total of 45 participants using an Intel RealSense F200 camera module and seven different video playlists. Each of the playlists had an approximate duration of 20 minutes and was viewed at least 17 times, with raw tracking data being recorded with a 0.05 second interval. The Data2MV dataset encompasses a total of 1.0 0 0.845 gaze fixations, gathered across a total of 128 experiments. It is also composed of 68.393 image frames, extracted from each of the 6 videos selected for these experiments, and an equal quantity of saliency maps, generated from aggregate fixation data. Software tools to obtain saliency maps and generate complementary plots are also provided as an open source software package. The Data2MV dataset was publicly released to the research community on Mendeley Data and constitutes an important contribution to reduce the current scarcity of such data, particularly in immersive, multi-view streaming scenarios. (c) 2023 Published by Elsevier Inc. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/)

Abstract
Switched reluctance machines are gaining importance due to their low cost, simple construction, and non-use of rare earth magnets. However, for the development of advanced torque controllers, accurate torque estimation is crucial, especially under varying load conditions. There are different torque estimation methods, which fall into different well-established classes, however, the characterization of their performance and operating conditions are not well known. This paper provides a comparative study of the most significant estimation algorithms: average torque, analytical and area approximation estimators. To assess the performance of these algorithms, a set of numerical simulations is presented and their results are compared based on signal similarity criteria. Results show a better performance when using the area approximation algorithm in comparison with the other two.

Abstract
Context. Hot atomic hydrogen emission lines in pre-main sequence stars serve as tracers for physical processes in the innermost regions of circumstellar accretion disks, where the interaction between a star and disk is the dominant influence on the formation of infalls and outflows. In the highly magnetically active T Tauri stars, this interaction region is particularly shaped by the stellar magnetic field and the associated magnetosphere, covering the inner five stellar radii around the central star. Even for the closest T Tauri stars, a region as compact as this is only observed on the sky plane at sub-mas scales. To resolve it spatially, the capabilities of optical long baseline interferometry are required.Aims. We aim to spatially and spectrally resolve the Br gamma hydrogen emission line with the methods of interferometry in order to examine the kinematics of the hydrogen gas emission region in the inner accretion disk of a sample of solar-like young stellar objects. The goal is to identify trends and categories among the sources of our sample and to discuss whether or not they can be tied to different origin mechanisms associated with Br gamma emission in T Tauri stars, chiefly and most prominently magnetospheric accretion.Methods. We observed a sample of seven T Tauri stars for the first time with VLTI GRAVITY, recording spectra and spectrally dispersed interferometric quantities across the Br gamma line at 2.16 mu m in the near-infrared K-band. We used the visibilities and differential phases to extract the size of the Br gamma emission region and the photocentre shifts on a channel-by-channel basis, probing the variation of spatial extent at different radial velocities. To assist in the interpretation, we also made use of radiative transfer models of magnetospheric accretion to establish a baseline of expected interferometric signatures if accretion is the primary driver of Br gamma emission.Results. From among our sample, we find that five of the seven T Tauri stars show an emission region with a half-flux radius in the four to seven stellar radii range that is broadly expected for magnetospheric truncation. Two of the five objects also show Br gamma emission primarily originating from within the co-rotation radius, which is an important criterion for magnetospheric accretion. Two objects exhibit extended emission on a scale beyond 10 R-(sic), one of them is even beyond the K-band continuum half-flux radius of 11.3 R-(sic). The observed photocentre shifts across the line can be either similar to what is expected for disks in rotation or show patterns of higher complexity.Conclusions. Based on the observational findings and the comparison with the radiative transfer models, we find strong evidence to suggest that for the two weakest accretors in the sample, magnetospheric accretion is the primary driver of Br gamma radiation. The results for the remaining sources imply either partial or strong contributions coming from additional, spatially extended emission components in the form of outflows, such as stellar or disk winds. We expect that in actively accreting T Tauri stars, these phenomena typically occur simultaneously on different spatial scales. Through more advanced modelling, interferometry will be a key factor in disentangling their distinct contributions to the total Br gamma flux arising from the innermost disk regions.

Abstract
A recent relevant innovation in last-mile delivery is to consider the possibility of goods being delivered by couriers appointed through crowdsourcing. In this paper we focus on the setting of in-store customers delivering goods, ordered by online customers, on their way home. We assume that not all the proposed delivery tasks will necessarily be accepted, and use logistic regression to model the crowd agents' willingness to undertake a delivery. This model is then used to build a novel compensation scheme that determines reward values, based on the current plan for the professional fleet's routes and on the couriers' probabilities of acceptance, by employing a direct search algorithm that seeks to minimise the expected cost.