Search

Accreting millisecond pulsars (AMSPs) are excellent laboratories to study reflection spectra and their features from an accretion disk truncated by a rapidly rotating magnetosphere near the neutron star surface. These systems also exhibit thermonuclear (type-I) bursts that can provide insights on the accretion physics and fuel composition. We explore spectral properties of the AMSP SRGA J144459.2-0604207 observed during the outburst that recently led to its discovery in February 2024. We aim to characterize the spectral shape of the persistent emission, both its continuum and discrete features, and to analyze type-I bursts properties. We employ XMM and NuSTAR overlapping observations taken during the most recent outburst from SRGA J1444. We perform spectral analysis of the persistent (i.e., non-bursting) emission employing a semi-phenomenological continuum model composed of a dominant thermal Comptonization plus two thermal contributions, and a physical reflection model. We also perform time-resolved spectral analysis of a type-I burst employing a blackbody model. We observe a broadened iron emission line, thus suggesting relativistic effects, supported by the physical model accounting for relativistically blurred reflection. The resulting accretion disk extends down to 6 gravitational radii, inclined at ~$53^{\circ}$, and only moderately ionized (log$\xi\simeq2.3$). We observe an absorption edge at ~9.7 keV that can be interpreted as an Fe XXVI edge blueshifted by an ultrafast ($\simeq0.04$c) outflow. Our broadband observations of type-I bursts do not find evidence of photospheric radius expansion. The burst recurrence time shows a dependence on the count rate with the steepest slope ever observed in these systems. We also observe a discrepancy of ~3 between the observed and expected burst recurrence time, which we discuss in the framework of fuel composition and high NS mass scenarios., Comment: 8 pages. Submitted to A&A

Context: NICER (Neutron star Interior Composition ExploreR) is the instrument of choice for the spectral analysis of type I X-ray bursts, as it provides high throughput at X-ray CCD resolution, down to 0.3 keV. Aims: This study investigates whether the energies of absorption lines detected in photospheric radius expansion (PRE) bursts correlate with the inferred blackbody radius. Previous reports suggested such a correlation, attributed to a combination of weaker gravitational redshift and higher blueshifts in bursts with larger radii. Methods: The analysis reexamines four previously studied PRE bursts and examines eight additional bursts from 4U 1820-303, evidencing PRE. Spectral evolution is tracked on the shortest possible timescales (tenth of a second) adopting two parallel continuum descriptions to characterise the photospheric expansion and line evolution. Applying the accretion-enhanced model, maximum blackbody radii of up to $\sim$ 900 km are inferred, with peak bolometric luminosities exceeding the Eddington limit of an Helium accretor. Absorption lines are assessed for significance using Monte Carlo simulations, and spectral lines are characterised using the state-of-art plasma codes available within {\sc{spex}} with a phenomenological continuum. A thorough parameter search explores Doppler shifts to avoid local minima. Results: Several significant (> 99.9%) absorption lines, including the previously reported 2.97 keV line, are detected. While no consistent correlation between line energies and blackbody radii is confirmed, bursts with larger radii exhibit up to four lines and the line strength is higher. The modelling suggests that the observed lines mostly originate from slightly redshifted (almost rest-frame) photo-/collisionally ionised gas in emission. For the burst with the largest PRE, a combination of photo-ionised plasma in both emission and absorption is preferred., Comment: 18 pages, 11 figures, accepted for publication in Astronomy & Astrophysics

We aim to investigate the energy-resolved pulse profile changes of the accreting X-ray pulsar V 0332+53, focusing in the cyclotron line energy range, using the full set of available NuSTAR observations. We applied a tailored pipeline to study the energy dependence of the pulse profiles and to build the pulsed fraction spectra (PFS) for the different observations. We studied the profile changes also using cross-correlation and lag spectra. We re-analysed the energy spectra to search for links between the local features observed in the PFS and spectral emission components associated with the shape of the fundamental cyclotron line. In the PFS data, with sufficiently high statistics, we observe a consistent behaviour around the cyclotron line energy. Specifically, two Gaussian-shaped features appear symmetrically on either side of the putative cyclotron line. These features exhibit minimal variation with source luminosity, and their peak positions consistently remain on the left and right of the cyclotron line energy. Associated with the cyclotron line-forming region, we interpret them as evidence for the resonant cyclotron absorption line wings, as predicted by theoretical models of how the cyclotron line profile should appear along the observer's line of sight. A phase-resolved analysis of the pulse in the energy bands surrounding these features enables us to determine both the spectral shape and the intensity of the photons responsible for these peaks in the PFS. Assuming these features correspond to a spectral component, we used their shapes as priors for the corresponding emission components, finding a statistically satisfactory description of the spectra. To explain these results, we propose that our line of sight is close to the direction of the spin axis, while the magnetic axis is likely orthogonal to it., Comment: 33 pages, 58 figures. Accepted for publication in A&A

Inverting visual representations within deep neural networks (DNNs) presents a challenging and important problem in the field of security and privacy for deep learning. The main goal is to invert the features of an unidentified target image generated by a pre-trained DNN, aiming to reconstruct the original image. Feature inversion holds particular significance in understanding the privacy leakage inherent in contemporary split DNN execution techniques, as well as in various applications based on the extracted DNN features. In this paper, we explore the use of diffusion models, a promising technique for image synthesis, to enhance feature inversion quality. We also investigate the potential of incorporating alternative forms of prior knowledge, such as textual prompts and cross-frame temporal correlations, to further improve the quality of inverted features. Our findings reveal that diffusion models can effectively leverage hidden information from the DNN features, resulting in superior reconstruction performance compared to previous methods. This research offers valuable insights into how diffusion models can enhance privacy and security within applications that are reliant on DNN features.

The rapid advancement of large language models (LLMs) has opened new boundaries in the extraction and synthesis of medical knowledge, particularly within evidence synthesis. This paper reviews the state-of-the-art applications of LLMs in the biomedical domain, exploring their effectiveness in automating complex tasks such as evidence synthesis and data extraction from a biomedical corpus of documents. While LLMs demonstrate remarkable potential, significant challenges remain, including issues related to hallucinations, contextual understanding, and the ability to generalize across diverse medical tasks. We highlight critical gaps in the current research literature, particularly the need for unified benchmarks to standardize evaluations and ensure reliability in real-world applications. In addition, we propose directions for future research, emphasizing the integration of state-of-the-art techniques such as retrieval-augmented generation (RAG) to enhance LLM performance in evidence synthesis. By addressing these challenges and utilizing the strengths of LLMs, we aim to improve access to medical literature and facilitate meaningful discoveries in healthcare., Comment: 9 pages, 1 table, 1 figure, conference paper

The structure and kinematics of the Milky Way disk are largely inferred from the solar vicinity. To gain a comprehensive understanding, it is essential to find reliable tracers in less-explored regions like the bulge and the far side of the disk. Mira variables, which are well-studied and bright standard candles, offer an excellent opportunity to trace intermediate and old populations in these complex regions. We aim to isolate a clean sample of Miras in the Vista Variables in the V\'ia L\'actea survey using Gaussian process algorithms. This sample will be used to study intermediate and old age populations in the Galactic bulge and far disk. Near- and mid-infrared time-series photometry were processed using Gaussian Process algorithms to identify Mira variables and model their light curves. We calibrated selection criteria with a visually inspected sample to create a high-purity sample of Miras, integrating multi-band photometry and kinematic data from proper motions. We present a catalog of 3602 Mira variables. By analyzing photometry, we classify them by O-rich or C-rich surface chemistry and derive selective-to-total extinction ratios of $A_{K_{s}}/E(J - K_{s}) = 0.471 \pm 0.01$ and $A_{K_{s}}/E(H - K_{s}) = 1.320 \pm 0.020$. Using the Mira period-age relation, we find evidence supporting the inside-out formation of the Milky Way disk. The distribution of proper motions and distances aligns with the Galactic rotation curve and disk kinematics. We extend the rotation curve up to R$_{\rm GC} \sim 17 \ \rm{kpc}$ and find no strong evidence of the nuclear stellar disk in our Mira sample. This study constitutes the largest catalog of variable stars on the far side of the Galactic disk to date., Comment: 20 pages, 19 figures, Accepted in A&A

We present GazeGen, a user interaction system that generates visual content (images and videos) for locations indicated by the user's eye gaze. GazeGen allows intuitive manipulation of visual content by targeting regions of interest with gaze. Using advanced techniques in object detection and generative AI, GazeGen performs gaze-controlled image adding/deleting, repositioning, and surface style changes of image objects, and converts static images into videos. Central to GazeGen is the DFT Gaze (Distilled and Fine-Tuned Gaze) agent, an ultra-lightweight model with only 281K parameters, performing accurate real-time gaze predictions tailored to individual users' eyes on small edge devices. GazeGen is the first system to combine visual content generation with real-time gaze estimation, made possible exclusively by DFT Gaze. This real-time gaze estimation enables various visual content generation tasks, all controlled by the user's gaze. The input for DFT Gaze is the user's eye images, while the inputs for visual content generation are the user's view and the predicted gaze point from DFT Gaze. To achieve efficient gaze predictions, we derive the small model from a large model (10x larger) via novel knowledge distillation and personal adaptation techniques. We integrate knowledge distillation with a masked autoencoder, developing a compact yet powerful gaze estimation model. This model is further fine-tuned with Adapters, enabling highly accurate and personalized gaze predictions with minimal user input. DFT Gaze ensures low-latency and precise gaze tracking, supporting a wide range of gaze-driven tasks. We validate the performance of DFT Gaze on AEA and OpenEDS2020 benchmarks, demonstrating low angular gaze error and low latency on the edge device (Raspberry Pi 4). Furthermore, we describe applications of GazeGen, illustrating its versatility and effectiveness in various usage scenarios., Comment: 12 pages, 10 figures

With the emergence of the Metaverse and focus on wearable devices in the recent years gesture based human-computer interaction has gained significance. To enable gesture recognition for VR/AR headsets and glasses several datasets focusing on egocentric i.e. first-person view have emerged in recent years. However, standard frame-based vision suffers from limitations in data bandwidth requirements as well as ability to capture fast motions. To overcome these limitation bio-inspired approaches such as event-based cameras present an attractive alternative. In this work, we present the first event-camera based egocentric gesture dataset for enabling neuromorphic, low-power solutions for XR-centric gesture recognition. The dataset has been made available publicly at the following URL: https://gitlab.com/NVM_IITD_Research/xrage.

Following their success in natural language processing (NLP), there has been a shift towards transformer models in computer vision. While transformers perform well and offer promising multi-tasking performance, due to their high compute requirements, many resource-constrained applications still rely on convolutional or hybrid models that combine the benefits of convolution and attention layers and achieve the best results in the sub 100M parameter range. Simultaneously, task adaptation techniques that allow for the use of one shared transformer backbone for multiple downstream tasks, resulting in great storage savings at negligible cost in performance, have not yet been adopted for hybrid transformers. In this work, we investigate how to achieve the best task-adaptation performance and introduce PETAH: Parameter Efficient Task Adaptation for Hybrid Transformers. We further combine PETAH adaptation with pruning to achieve highly performant and storage friendly models for multi-tasking. In our extensive evaluation on classification and other vision tasks, we demonstrate that our PETAH-adapted hybrid models outperform established task-adaptation techniques for ViTs while requiring fewer parameters and being more efficient on mobile hardware.

We present a new unified graph-based representation of medical data, combining genetic information and medical records of patients with medical knowledge via a unique knowledge graph. This approach allows us to infer meaningful information and explanations that would be unavailable by looking at each data set separately. The systematic use of different databases, managed throughout the built knowledge graph, gives new insights toward a better understanding of oncology medicine. Indeed, we reduce some useful medical tasks to well-known problems in theoretical computer science for which efficient algorithms exist., Comment: 19 pages, 1 figure, 14 tables, CIBB 2024 conference

Low-Latency and Low-Power Edge AI is essential for Virtual Reality and Augmented Reality applications. Recent advances show that hybrid models, combining convolution layers (CNN) and transformers (ViT), often achieve superior accuracy/performance tradeoff on various computer vision and machine learning (ML) tasks. However, hybrid ML models can pose system challenges for latency and energy-efficiency due to their diverse nature in dataflow and memory access patterns. In this work, we leverage the architecture heterogeneity from Neural Processing Units (NPU) and Compute-In-Memory (CIM) and perform diverse execution schemas to efficiently execute these hybrid models. We also introduce H4H-NAS, a Neural Architecture Search framework to design efficient hybrid CNN/ViT models for heterogeneous edge systems with both NPU and CIM. Our H4H-NAS approach is powered by a performance estimator built with NPU performance results measured on real silicon, and CIM performance based on industry IPs. H4H-NAS searches hybrid CNN/ViT models with fine granularity and achieves significant (up to 1.34%) top-1 accuracy improvement on ImageNet dataset. Moreover, results from our Algo/HW co-design reveal up to 56.08% overall latency and 41.72% energy improvements by introducing such heterogeneous computing over baseline solutions. The framework guides the design of hybrid network architectures and system architectures of NPU+CIM heterogeneous systems.

We apply the linked cluster expansion as well as the quench action approach to study the time evolution of one-point functions after a quantum quench in integrable field theories. We argue that the relaxation towards the stationary value fundamentally differs depending on the locality properties of the considered observable: while for local operators both exponential and power-law decaying terms are present in the dynamics, for semi-local operators the latter are absent. We explicitly confirm this for the Ising field theory, the sinh-Gordon model, and the repulsive sine-Gordon model., Comment: 43 pages

This study introduces unORANIC+, a novel method that integrates unsupervised feature orthogonalization with the ability of a Vision Transformer to capture both local and global relationships for improved robustness and generalizability. The streamlined architecture of unORANIC+ effectively separates anatomical and image-specific attributes, resulting in robust and unbiased latent representations that allow the model to demonstrate excellent performance across various medical image analysis tasks and diverse datasets. Extensive experimentation demonstrates unORANIC+'s reconstruction proficiency, corruption resilience, as well as capability to revise existing image distortions. Additionally, the model exhibits notable aptitude in downstream tasks such as disease classification and corruption detection. We confirm its adaptability to diverse datasets of varying image sources and sample sizes which positions the method as a promising algorithm for advanced medical image analysis, particularly in resource-constrained environments lacking large, tailored datasets. The source code is available at https://github.com/sdoerrich97/unoranic-plus ., Comment: Accepted at RROW@BMVC 2024 (Workshop on Robust Recognition in the Open World at the British Machine Vision Conference)

We propose a novel approach for determining the orbital inclination of low-mass X-ray binary systems by modelling the H$\alpha$ and H$\beta$ line profiles emitted by the accretion disc, with a Newtonian version of diskline. We applied the model to two sample sources, Swift J1357.2-0933 and MAXI J1305-704, which are both transient black hole systems, and analyse two observations that were collected during a quiescent state and one observation of an outburst. The line profile is well described by the diskline model, although we had to add a Gaussian line to describe the deep inner core of the double-peaked profile, which the diskline model was unable to reproduce. The H$\beta$ emission lines in the spectrum of Swift J1357.2-0933 and the H$\alpha$ emission lines in that of MAXI J1305-704 during the quiescent state are consistent with a scenario in which these lines originate from a disc ring between $(9.6-57) \times 10^{3}, \rm{R_{g}}$ and $(1.94-20) \times 10^{4}, \rm{R_{g}}$, respectively. We estimate an inclination angle of $81 \pm 5$ degrees for Swift J1357.2-0933 and an angle of $73 \pm 4$ degrees for MAXI J1305-704. This is entirely consistent with the values reported in the literature. In agreement with the recent literature, our analysis of the outburst spectrum of MAXI J1305-704 revealed that the radius of the emission region deviates from expected values. This outcome implies several potential scenarios, including alternative disc configuration or even a circumbinary disc. We caution that these results were derived from a simplistic model that may not fully describe the complicated physics of accretion discs. Despite these limitations, our results for the inclination angles are remarkably consistent with recent complementary studies, and the proposed description of the emitting region remains entirely plausible., Comment: 11 pages, 7 figures

We revisit the construction of the duality group for $\mathcal N=2$ $\widehat{A}_n$-shaped quivers SCFTs and generalize it to the previously unexplored case of $\widehat{D}_n$-shaped quivers. We then provide a systematic description of non-invertible duality symmetries in both classes. Furthermore, we characterize the $\mathcal N=1$ mass deformations of these theories that preserve such symmetries, thereby identifying a large class of $\mathcal N=1$ SCFTs with non-invertible duality symmetries inherited from their parent $\mathcal N=2$ theories., Comment: 44 pages, 16 figures

HERMES (High Energy Rapid Modular Ensemble of Satellites) Pathfinder is a space-borne mission based on a constellation of six nano-satellites flying in a low-Earth orbit (LEO). The 3U CubeSats, to be launched in early 2025, host miniaturized instruments with a hybrid Silicon Drift Detector/GAGG:Ce scintillator photodetector system, sensitive to X-rays and gamma-rays in a large energy band. HERMES will operate in conjunction with Australian Space Industry Responsive Intelligent Thermal (SpIRIT) 6U CubeSat, launched in December 2023. HERMES will probe the temporal emission of bright high-energy transients such as Gamma-Ray Bursts (GRBs), ensuring a fast transient localization in a field of view of several steradians exploiting the triangulation technique. HERMES intrinsically modular transient monitoring experiment represents a keystone capability to complement the next generation of gravitational wave experiments. In this paper we outline the scientific case, development and programmatic status of the mission, Comment: 13 pages, 16 figures. Proceedings of SPIE "Astronomical Telescopes and Instrumentation" 2024. arXiv admin note: text overlap with arXiv:2210.13866

The performance of deep neural networks scales with dataset size and label quality, rendering the efficient mitigation of low-quality data annotations crucial for building robust and cost-effective systems. Existing strategies to address label noise exhibit severe limitations due to computational complexity and application dependency. In this work, we propose WANN, a Weighted Adaptive Nearest Neighbor approach that builds on self-supervised feature representations obtained from foundation models. To guide the weighted voting scheme, we introduce a reliability score, which measures the likelihood of a data label being correct. WANN outperforms reference methods, including a linear layer trained with robust loss functions, on diverse datasets of varying size and under various noise types and severities. WANN also exhibits superior generalization on imbalanced data compared to both Adaptive-NNs (ANN) and fixed k-NNs. Furthermore, the proposed weighting scheme enhances supervised dimensionality reduction under noisy labels. This yields a significant boost in classification performance with 10x and 100x smaller image embeddings, minimizing latency and storage requirements. Our approach, emphasizing efficiency and explainability, emerges as a simple, robust solution to overcome inherent limitations of deep neural network training. The code is available at https://github.com/francescodisalvo05/wann-noisy-labels ., Comment: Preprint - Under Review

We present a comprehensive study of the X-ray spectral properties of the accreting millisecond pulsar IGR J17498$-$2921 during its 2023 outburst. Similar to other accreting millisecond pulsars, the broad-band spectral emission observed quasi-simultaneously by NICER and NuSTAR is well described by an absorbed Comptonized emission with an electron temperature of $\sim$17 keV plus a disk reflection component. The broadening of the disk reflection spectral features, such as a prominent iron emission line at 6.4-6.7 keV, is consistent with the relativistic motion of matter in a disk truncated at $\sim$$21 \, \mathrm{R_g}$ from the source, near the Keplerian co-rotation radius. From the high-cadence monitoring data obtained with NICER, we observe that the evolution of the photon index and the temperature of seed photons tracks variations in the X-ray flux. This is particularly evident close to a sudden $\sim$-0.25 cycles jump in the pulse phase, which occurs immediately following an X-ray flux flare and a drop in the pulse amplitude below the $3\sigma$ detection threshold. We also report on the non-detection of optical pulsations with TNG/SiFAP2 from the highly absorbed optical counterpart., Comment: 14 pages, 5 figures. Submitted to A&A

The interacting SSH model provides an ideal ground to study the interplay between topologically insulating phases and electron-electron interactions. We study the polarization density as a topological invariant and provide an analytic treatment of its behavior in the low-energy sector of the one-dimensional interacting SSH model. By formulating the topological invariant in terms of Green's functions, we use the low-energy field theory of the Thirring model to derive the behavior of the polarization density. We show that the polarization density in the continuum theory describes the usual topological insulating phases. Still, it contains an extra factor from the fields' scaling dimensions in the low-energy quantum field theory. We interpret this as a measure of the modified charge of the new excitations in the system. We find two distinct contributions: a renormalization of the electronic charge $e$ of a Fermi liquid because of quasiparticle smearing and an additional contribution coming from the topological charge of the soliton arising in the bosonized version of the Thirring model, the sine-Gordon model., Comment: 18 pages, 6 figures to be submitted to PRB. Comments are welcome

Large and well-annotated datasets are essential for advancing deep learning applications, however often costly or impossible to obtain by a single entity. In many areas, including the medical domain, approaches relying on data sharing have become critical to address those challenges. While effective in increasing dataset size and diversity, data sharing raises significant privacy concerns. Commonly employed anonymization methods based on the k-anonymity paradigm often fail to preserve data diversity, affecting model robustness. This work introduces a novel approach using Conditional Variational Autoencoders (CVAEs) trained on feature vectors extracted from large pre-trained vision foundation models. Foundation models effectively detect and represent complex patterns across diverse domains, allowing the CVAE to faithfully capture the embedding space of a given data distribution to generate (sample) a diverse, privacy-respecting, and potentially unbounded set of synthetic feature vectors. Our method notably outperforms traditional approaches in both medical and natural image domains, exhibiting greater dataset diversity and higher robustness against perturbations while preserving sample privacy. These results underscore the potential of generative models to significantly impact deep learning applications in data-scarce and privacy-sensitive environments. The source code is available at https://github.com/francescodisalvo05/cvae-anonymization ., Comment: Accepted at BMVC 2024

We report on the discovery of polarized X-ray emission from an accreting millisecond pulsar. During a 10-day-long coverage of the February 2024 outburst of SRGA J144459.2-604207, the Imaging X-ray Polarimetry Explorer (IXPE) detected an average polarization degree of the 2-8 keV emission of 2.3% +/- 0.4% at an angle of 59{\deg} +/- 6{\deg} (East of North; uncertainties quoted at the 1${\sigma}$ confidence level). The polarized signal shows a significant energy dependence with a degree of 4.0% +/- 0.5% between 3 and 6 keV and < 1.5% (90% c.l.) in the 2-3 keV range. We used NICER, XMM-Newton, and NuSTAR observations to obtain an accurate pulse timing solution and perform a phase-resolved polarimetric analysis of IXPE data. We did not detect any significant variability of the Stokes parameters Q and U with the spin and the orbital phases. We used the relativistic rotating vector model to show that a moderately fan-beam emission from two point-like spots at a small magnetic obliquity ($\simeq$ 10{\deg}) is compatible with the observed pulse profile and polarization properties. IXPE also detected 52 type-I X-ray bursts, with a recurrence time $\Delta t_{rec}$ increasing from 2 to 8 h as a function of the observed count rate $C$ as as $\Delta t_{rec} \simeq C^{-0.8}$ We stacked the emission observed during all the bursts and obtained an upper limit on the polarization degree of 8.5% (90% c.l.)., Comment: 10 pages, 10 figures, accepted for publication by A&A

Despite notable advancements, the integration of deep learning (DL) techniques into impactful clinical applications, particularly in the realm of digital histopathology, has been hindered by challenges associated with achieving robust generalization across diverse imaging domains and characteristics. Traditional mitigation strategies in this field such as data augmentation and stain color normalization have proven insufficient in addressing this limitation, necessitating the exploration of alternative methodologies. To this end, we propose a novel generative method for domain generalization in histopathology images. Our method employs a generative, self-supervised Vision Transformer to dynamically extract characteristics of image patches and seamlessly infuse them into the original images, thereby creating novel, synthetic images with diverse attributes. By enriching the dataset with such synthesized images, we aim to enhance its holistic nature, facilitating improved generalization of DL models to unseen domains. Extensive experiments conducted on two distinct histopathology datasets demonstrate the effectiveness of our proposed approach, outperforming the state of the art substantially, on the Camelyon17-wilds challenge dataset (+2%) and on a second epithelium-stroma dataset (+26%). Furthermore, we emphasize our method's ability to readily scale with increasingly available unlabeled data samples and more complex, higher parametric architectures. Source code is available at https://github.com/sdoerrich97/vits-are-generative-models ., Comment: Accepted at MICCAI 2024. This is the submitted manuscript with added link to github repo and funding acknowledgements. No further post submission improvements or corrections were integrated. Final version not published yet

The integration of neural-network-based systems into clinical practice is limited by challenges related to domain generalization and robustness. The computer vision community established benchmarks such as ImageNet-C as a fundamental prerequisite to measure progress towards those challenges. Similar datasets are largely absent in the medical imaging community which lacks a comprehensive benchmark that spans across imaging modalities and applications. To address this gap, we create and open-source MedMNIST-C, a benchmark dataset based on the MedMNIST+ collection covering 12 datasets and 9 imaging modalities. We simulate task and modality-specific image corruptions of varying severity to comprehensively evaluate the robustness of established algorithms against real-world artifacts and distribution shifts. We further provide quantitative evidence that our simple-to-use artificial corruptions allow for highly performant, lightweight data augmentation to enhance model robustness. Unlike traditional, generic augmentation strategies, our approach leverages domain knowledge, exhibiting significantly higher robustness when compared to widely adopted methods. By introducing MedMNIST-C and open-sourcing the corresponding library allowing for targeted data augmentations, we contribute to the development of increasingly robust methods tailored to the challenges of medical imaging. The code is available at https://github.com/francescodisalvo05/medmnistc-api ., Comment: Accepted at MICCAI Workshop on Advancing Data Solutions in Medical Imaging AI (ADSMI @ MICCAI 2024)

We compare the features of the semi-leptonic decays $B\to D^{(*)} l \bar{\nu}_l$ and $\Lambda_b\to \Lambda_c l \bar{\nu}_l$, with $l$ = $e$, $\mu$ and $\tau$. Firstly, we study such processes according to the $\ell-s$ scheme. Then, we write the angular distributions of those decays in the helicity formalism; we focus especially on the case of $l$ = $\tau$ and suggest for each decay some tests for distinguishing among the various Dirac operators which may explain the tension with the standard model. In particular, as regards the $\Lambda_b$ semi-leptonic decay, we single out five observables, which assume different values for each new-physics interaction; one such observable is related to the longitudinal polarization of the $\tau-\bar{\nu}_\tau$ system, which can be compared with the $D^*$ longitudinal polarization in the $B\to D^* l \bar{\nu}_l$ decay. Moreover, we perform a phenomenological analysis of the decays, particularly of $\Lambda_b$, starting from an assumption which agrees with previous theoretical calculations and has a predictive power about the total widths of several decays; we give also a theoretical explanation to some results that were found in preceding analyses about the above mentioned $B$-decays., Comment: 31 pages

The differential cross section for $\gamma p\rightarrow K^+\Lambda(1520)$ was measured from threshold to a centre-of-mass energy of 2090\,MeV at forward angles at the BGOOD experiment. The high statistical precision and resolution in centre-of-mass energy and angle allows a detailed characterisation of this low-momentum transfer kinematic region. The data agree with a previous LEPS measurement and support effective Lagrangian models that indicate that the contact term dominates the cross section near threshold., Comment: 6 pages, 7 figures

Gamma Ray Bursts (GRBs) bridge relativistic astrophysics and multi-messenger astronomy. Space-based gamma/X-ray wide field detectors have proven essential to detect and localize the highly variable GRB prompt emission, which is also a counterpart of gravitational wave events. We study the capabilities to detect long and short GRBs by the High Energy Rapid Modular Ensemble of Satellites (HERMES) Pathfinder (HP) and SpIRIT, namely a swarm of six 3U CubeSats to be launched in early 2025, and a 6U CubeSat launched on December 1st 2023. We also study the capabilities of two advanced configurations of swarms of >8 satellites with improved detector performances (HERMES Constellations). The HERMES detectors, sensitive down to ~2-3 keV, will be able to detect faint/soft GRBs which comprise X-ray flashes and high redshift bursts. By combining state-of-the-art long and short GRB population models with a description of the single module performance, we estimate that HP will detect ~195^{+22}_{-21} long GRBs (3.4^{+0.3}_{-0.8} at redshift z>6) and ~19^{+5}_{-3} short GRBs per year. The larger HERMES Constellations under study can detect between ~1300 and ~3000 long GRBs per year and between ~160 and ~400 short GRBs per year, depending on the chosen configuration, with a rate of long GRBs above z>6 between 30 and 75 per year. Finally, we explore the capabilities of HERMES to detect short GRBs as electromagnetic counterparts of binary neutron star (BNS) mergers detected as gravitational signals by current and future ground-based interferometers. Under the assumption that the GRB jets are structured, we estimate that HP can provide up to 1 (14) yr^{-1} joint detections during the fifth LIGO-Virgo-KAGRA observing run (Einstein Telescope single triangle 10 km arm configuration). These numbers become 4 (100) yr^{-1}, respectively, for the HERMES Constellation configuration., Comment: 13 pages, 6 figures, 4 tabels. Accepted for publication by Astronomy & Astrophysics

The coherent reaction, $\gamma d \rightarrow \pi^0\eta d$ was studied with the BGOOD experiment at ELSA from threshold to a centre-of-mass energy of 3200\,MeV. A full kinematic reconstruction was made, with final state deuterons identified in the forward spectrometer and $\pi^0$ and $\eta$ decays in the central BGO Rugby Ball. The strength of the differential cross section exceeds what can be described by models of coherent photoproduction at forward angles by orders of magnitude. The distribution of the differential cross section has an excellent agreement with a model including quasi-free $\Delta \pi$ photoproduction, pion re-scattering and $N(1535)$ formation and subsequent nucleon coalescence to the deuteron. This also gives a reasonable description of the two-body invariant mass distributions and naturally explains the similar magnitudes of this channel and $\pi^0\pi^0 d$ coherent photoproduction., Comment: 7 pages, 5 figures

The integration of deep learning based systems in clinical practice is often impeded by challenges rooted in limited and heterogeneous medical datasets. In addition, prioritization of marginal performance improvements on a few, narrowly scoped benchmarks over clinical applicability has slowed down meaningful algorithmic progress. This trend often results in excessive fine-tuning of existing methods to achieve state-of-the-art performance on selected datasets rather than fostering clinically relevant innovations. In response, this work presents a comprehensive benchmark for the MedMNIST+ database to diversify the evaluation landscape and conduct a thorough analysis of common convolutional neural networks (CNNs) and Transformer-based architectures, for medical image classification. Our evaluation encompasses various medical datasets, training methodologies, and input resolutions, aiming to reassess the strengths and limitations of widely used model variants. Our findings suggest that computationally efficient training schemes and modern foundation models hold promise in bridging the gap between expensive end-to-end training and more resource-refined approaches. Additionally, contrary to prevailing assumptions, we observe that higher resolutions may not consistently improve performance beyond a certain threshold, advocating for the use of lower resolutions, particularly in prototyping stages, to expedite processing. Notably, our analysis reaffirms the competitiveness of convolutional models compared to ViT-based architectures emphasizing the importance of comprehending the intrinsic capabilities of different model architectures. Moreover, we hope that our standardized evaluation framework will help enhance transparency, reproducibility, and comparability on the MedMNIST+ dataset collection as well as future research within the field. Code is available at https://github.com/sdoerrich97 .

XTE J1710-281 is a transient eclipsing binary system with a period close to 3.28 hours, hosting a neutron star. The average eclipse duration is 420 seconds, and eclipse arrival times reported in the literature span from 1999 to 2017. A previous analysis of the eclipse arrival times using the eclipse timing technique revealed a complex pattern of delays, indicating the presence of three orbital glitches. These glitches correspond to sudden variations in the orbital period, allowing for the identification of four distinct epochs. We have re-analyzed the 78 eclipse arrival times spanning 18 years utilizing the eclipse timing technique to derive the corresponding delays as a function of time. We find that the observed delays align well with a fitting model employing an eccentric sine function characterized by an amplitude of $6.1 \pm 0.5$ s, eccentricity of $0.38 \pm 0.17$, and a period of $17.1 \pm 1.5$ years. Additionally, we identified the orbital period as 3.28106345(13) hours, with a reference epoch of $T_0=54112.83200(2)$ Modified Julian Date (MJD). We obtained an upper limit of the orbital period derivative of $3.6 \times 10^{-13}$ s~s$^{-1}$. From the average value of the eclipse duration, we estimate that the companion star has a mass of 0.22~\Msun for a neutron star mass of 1.4~\Msun, and the inclination of the source is $78.1^{+1.5}_{-1.2}$ degrees. The companion star is in thermal equilibrium. The orbital period derivative is consistent with a conservative mass transfer scenario, where the angular momentum loss due to magnetic braking dominates over gravitational radiation angular momentum loss if the former is present. The eccentric modulation can be explained by a third body with a mass of 2.7 Jovian masses, orbiting with a revolution period close to 17 years and an eccentricity of 0.38. (abridged abstract), Comment: 7 pages, 10 figures, accepted for the publication by A&A on 12/04/2024

Machine learning plays an important role in the operation of current wind energy production systems. One central application is predictive maintenance to increase efficiency and lower electricity costs by reducing downtimes. Integrating physics-based knowledge in neural networks to enforce their physical plausibilty is a promising method to improve current approaches, but incomplete system information often impedes their application in real world scenarios. We describe a simple and efficient way for physics-constrained deep learning-based predictive maintenance for wind turbine gearbox bearings with partial system knowledge. The approach is based on temperature nowcasting constrained by physics, where unknown system coefficients are treated as learnable neural network parameters. Results show improved generalization performance to unseen environments compared to a baseline neural network, which is especially important in low data scenarios often encountered in real-world applications., Comment: Published at ICLR 2024 Tackling Climate Change with Machine Learning Workshop

The nuclear region of the Milky Way is believed to host a nuclear stellar disk, co-spatial with the gaseous central molecular zone. Previous kinematical studies detected faster rotation for the stars belonging to the nuclear stellar disk, compared to the surrounding regions. We analyze the rotation velocity of stars at the nuclear stellar disk, and compare them with its analog in a few control fields just outside this region. We limit our analysis to stars in the red clump of the color magnitude diagram, in order to be able to relate their mean de-reddened luminosity with distance along the line of sight. We used a proper motion catalog, obtained from point spread function photometry on VVV images, to construct maps of the transverse velocity for these stars. We complemented our analysis with radial velocities from the 17th data release of the APOGEE survey. We find that the main difference between the nuclear stellar disk region and its surroundings is that at the former we see only stars moving eastward, which we believe are located in front of the Galactic center. On the contrary, in every other direction, we see the brightest red clump stars moving eastward, and the faintest ones moving westward, as expected for a rotating disk. We interpret these observations as being produced by the central molecular zone, hiding stars behind itself. What we observe is compatible with being produced by just the absence of the component at the back, without requiring the presence of a cold, fast rotating disk. This component is also not clearly detected in the newest release of the APOGEE catalog. In other words, we find no clear signature of the nuclear stellar disk as a distinct kinematical component., Comment: 12 pages, 14 figures, accepted for publication in Astronomy and Astrophysics

Traditional deep learning models implicity encode knowledge limiting their transparency and ability to adapt to data changes. Yet, this adaptability is vital for addressing user data privacy concerns. We address this limitation by storing embeddings of the underlying training data independently of the model weights, enabling dynamic data modifications without retraining. Specifically, our approach integrates the $k$-Nearest Neighbor ($k$-NN) classifier with a vision-based foundation model, pre-trained self-supervised on natural images, enhancing interpretability and adaptability. We share open-source implementations of a previously unpublished baseline method as well as our performance-improving contributions. Quantitative experiments confirm improved classification across established benchmark datasets and the method's applicability to distinct medical image classification tasks. Additionally, we assess the method's robustness in continual learning and data removal scenarios. The approach exhibits great promise for bridging the gap between foundation models' performance and challenges tied to data privacy. The source code is available at https://github.com/TobArc/privacy-aware-image-classification-with-kNN., Comment: Accepted at 21st IEEE International Symposium on Biomedical Imaging (IEEE ISBI 2024)