Search

[Abridged] The X-ray broadband spectra of the bare AGN Mrk 110, obtained by simultaneous XMM-Newton and NuSTAR observations (Nov 2019 and April 2020), are characterised by the presence of a prominent and absorption-free smooth soft X-ray excess, moderately broad OVII and Fe Kalpha emission lines, and a lack of a strong Compton hump. While relativistic reflection as the sole emission is ruled out, a simplified combination of soft and hard Comptonisation from a warm and a hot coronae, plus mild relativistic disc reflection reproduces the data very well. We aim to confirm the physical origin of the soft X-ray excess of Mrk 110 and to determine its disc-corona system properties from its energetics using two new sophisticated models: reXcor and relagn, respectively. At both epochs, the inferred high-values of the warm-corona heating from the X-ray broadband spectral analysis using reXcor confirm that the soft X-ray excess originates mainly from a warm corona rather than relativistic reflection. The intrinsic best-fit SED determined at both epochs using relagn show a high X-ray contribution relative to the UV and are very well reproduced by a warm and hot coronae plus mild relativistic reflection. The outer radii of the hot and warm coronae are located at a few 10s and ~100 Rg, respectively. Moreover, combining the inferred low Eddington ratio (~ a few %) from this work, and previous multi-wavelength spectral and timing studies suggests that Mrk 110 could be classified as a moderate changing-state AGN. Our analysis confirms the existence of a warm corona as a significant contribution to the soft X-ray excess and UV emission in Mrk 110, adding to growing evidence that AGN accretion deviates from standard disc theory. This strengthens the importance of long-term multi-wavelength monitoring on both single targets and large AGN surveys to reveal the real nature of disc-corona system in AGN., Comment: 12 pages, 7 figures, accepted for publication in A&A

Introduction: Photogrammetric surface scans provide a radiation-free option to assess and classify craniosynostosis. Due to the low prevalence of craniosynostosis and high patient restrictions, clinical data is rare. Synthetic data could support or even replace clinical data for the classification of craniosynostosis, but this has never been studied systematically. Methods: We test the combinations of three different synthetic data sources: a statistical shape model (SSM), a generative adversarial network (GAN), and image-based principal component analysis for a convolutional neural network (CNN)-based classification of craniosynostosis. The CNN is trained only on synthetic data, but validated and tested on clinical data. Results: The combination of a SSM and a GAN achieved an accuracy of more than 0.96 and a F1-score of more than 0.95 on the unseen test set. The difference to training on clinical data was smaller than 0.01. Including a second image modality improved classification performance for all data sources. Conclusion: Without a single clinical training sample, a CNN was able to classify head deformities as accurate as if it was trained on clinical data. Using multiple data sources was key for a good classification based on synthetic data alone. Synthetic data might play an important future role in the assessment of craniosynostosis.

We propose a Perceiver-based sequence classifier to detect abnormalities in speech reflective of several neurological disorders. We combine this classifier with a Universal Speech Model (USM) that is trained (unsupervised) on 12 million hours of diverse audio recordings. Our model compresses long sequences into a small set of class-specific latent representations and a factorized projection is used to predict different attributes of the disordered input speech. The benefit of our approach is that it allows us to model different regions of the input for different classes and is at the same time data efficient. We evaluated the proposed model extensively on a curated corpus from the Mayo Clinic. Our model outperforms standard transformer (80.9%) and perceiver (81.8%) models and achieves an average accuracy of 83.1%. With limited task-specific data, we find that pretraining is important and surprisingly pretraining with the unrelated automatic speech recognition (ASR) task is also beneficial. Encodings from the middle layers provide a mix of both acoustic and phonetic information and achieve best prediction results compared to just using the final layer encodings (83.1% vs. 79.6%). The results are promising and with further refinements may help clinicians detect speech abnormalities without needing access to highly specialized speech-language pathologists.

How will generative AI pay for itself? Unless charging users for access, selling advertising is the only alternative. Especially in the multi-billion dollar web search market with ads as the main source of revenue, the introduction of a subscription model seems unlikely. The recent disruption of search by generative large language models could thus ultimately be accompanied by generated ads. Our concern is that the commercialization of generative AI in general and large language models in particular could lead to native advertising in the form of quite subtle brand or product placements. In web search, the evolution of search engine results pages (SERPs) from traditional lists of ``ten blue links'' (lists SERPs) to generated text with web page references (text SERPs) may further blur the line between advertising-based and organic search results, making it difficult for users to distinguish between the two, depending on how advertising is integrated and disclosed. To raise awareness of this potential development, we conduct a pilot study analyzing the capabilities of current large language models to blend ads with organic search results. Although the models still struggle to subtly frame ads in an unrelated context, their potential is evident when integrating ads into related topics which calls for further investigation., Comment: Presented at OSSYM 2023

This article focuses on developing classical surrogates for parametrized quantum circuits using interpolation via (trigonometric) polynomials. We develop two algorithms for the construction of such surrogates and prove performance guarantees. The constructions are based on circuit evaluations which are blackbox in the sense that no structural specifics of the circuits are exploited. While acknowledging the limitations of the blackbox approach compared to whitebox evaluations, which exploit specific circuit properties, we demonstrate scenarios in which the blackbox approach might prove beneficial. Sample applications include but are not restricted to the approximation of VQEs and the alleviaton of the barren plateau problem., Comment: 27 pages, 3 figures, 1 table

We present calculations of the $^{40}$Ca transverse response function obtained from coupled-cluster theory used in conjunction with the Lorentz integral transform method. We employ nuclear forces derived at next-to-next-to leading order in chiral effective field theory with and without $\Delta$ degrees of freedom. We first benchmark this approach on the $^4$He nucleus and compare both the transverse sum rule and the response function to earlier calculations based on different methods. As expected from the power counting of the chiral expansion of electromagnetic currents and from previous studies, our results retaining only one-body term underestimate the experimental data for $^4$He by about $20\%$. However, when the method is applied to $^{40}$Ca at the same order of the expansion, response functions do not lack strength and agree well with the world electron scattering data. We discuss various sources of theoretical uncertainties and comment on the comparison of our results with the available experiments., Comment: 9 pages, 9 figures

Virtual and augmented realities are increasingly popular tools in many domains such as architecture, production, training and education, (psycho)therapy, gaming, and others. For a convincing rendering of sound in virtual and augmented environments, audio signals must be convolved in real-time with impulse responses that change from one moment in time to another. Key requirements for the implementation of such time-variant real-time convolution algorithms are short latencies, moderate computational cost and memory footprint, and no perceptible switching artifacts. In this engineering report, we introduce a partitioned convolution algorithm that is able to quickly switch between impulse responses without introducing perceptible artifacts, while maintaining a constant computational load and low memory usage. Implementations in several popular programming languages are freely available via GitHub.

We present a joint Speech and Language Model (SLM), a multitask, multilingual, and dual-modal model that takes advantage of pretrained foundational speech and language models. SLM freezes the pretrained foundation models to maximally preserves their capabilities, and only trains a simple adapter with just 1\% (156M) of the foundation models' parameters. This adaptation not only leads SLM to achieve strong performance on conventional tasks such as speech recognition (ASR) and speech translation (AST), but also introduces the novel capability of zero-shot instruction-following for more diverse tasks: given a speech input and a text instruction, SLM is able to perform unseen generation tasks including contextual biasing ASR using real-time context, dialog generation, speech continuation, and question answering, etc. Our approach demonstrates that the representational gap between pretrained speech and language models might be narrower than one would expect, and can be bridged by a simple adaptation mechanism. As a result, SLM is not only efficient to train, but also inherits strong capabilities already acquired in foundation models of different modalities.

We show that quasi-isometries of (well-behaved) hierarchically hyperbolic groups descend to quasi-isometries of their maximal hyperbolic space. This has two applications, one relating to quasi-isometry invariance of acylindrical hyperbolicity, and the other a linear progress result for Markov chains. The appendix, by Jacob Russell, contains a partial converse under the (necessary) condition that the maximal hyperbolic space is one-ended., Comment: 38 pages, 1 figure. Main paper by A. Goldsborough, M. Hagen, H. Petyt and A. Sisto; appendix by J. Russell. To appear in Groups, Geometry, and Dynamics

Charge radii of neutron deficient 40Sc and 41Sc nuclei were determined using collinear laser spectroscopy. With the new data, the chain of Sc charge radii extends below the neutron magic number N=20 and shows a pronounced kink, generally taken as a signature of a shell closure, but one notably absent in the neighboring Ca, K and Ar isotopic chains. Theoretical models that explain the trend at N=20 for the Ca isotopes cannot reproduce this puzzling behavior.

The combinatorial hierarchical hyperbolicity criterion is a very useful way of constructing new hierarchically hyperbolic spaces (HHSs). We show that, conversely, HHSs satisfying natural assumptions (satisfied, for example, by mapping class groups) admit a combinatorial HHS structure. This can be useful in constructions of new HHSs, and also our construction clarifies how to apply the combinatorial HHS criterion to suspected examples. We also uncover connections between HHS notions and lattice theory notions., Comment: 64 pages, 10 figures

The development of tissue-engineered cardiovascular implants can improve the lives of large segments of our society who suffer from cardiovascular diseases. Regenerative tissues are fabricated using a process called tissue maturation. Furthermore, it is highly challenging to produce cardiovascular regenerative implants with sufficient mechanical strength to withstand the loading conditions within the human body. Therefore, biohybrid implants for which the regenerative tissue is reinforced by standard reinforcement material (e.g. textile or 3d printed scaffold) can be an interesting solution. In silico models can significantly contribute to characterizing, designing, and optimizing biohybrid implants. The first step towards this goal is to develop a computational model for the maturation process of tissue-engineered implants. This paper focuses on the mechanical modeling of textile-reinforced tissue-engineered cardiovascular implants. First, we propose an energy-based approach to compute the collagen evolution during the maturation process. Then, we apply the concept of structural tensors to model the anisotropic behavior of the extracellular matrix and the textile scaffold. Next, the newly developed material model is embedded into a special solid-shell finite element formulation with reduced integration. Finally, we use our framework to compute two structural problems: a pressurized shell construct and a tubular-shaped heart valve. The results show the ability of the model to predict collagen growth in response to the boundary conditions applied during the maturation process. Consequently, we can predict the implant's mechanical response, such as the deformation and stresses of the implant., Comment: Preprint submitted to Elsevier

Gamma decays were observed in $^{56}$Ca and $^{58}$Ca following quasi-free one-proton knockout reactions from $^{57,59}$Sc beams at $\approx 200$ MeV/nucleon. For $^{56}$Ca, a $\gamma$ ray transition was measured to be 1456(12) keV, while for $^{58}$Ca an indication for a transition was observed at 1115(34) keV. Both transitions were tentatively assigned as the $2^+_1 \rightarrow 0^+_{gs}$ decays, and were compared to results from ab initio and conventional shell-model approaches. A shell-model calculation in a wide model space with a marginally modified effective nucleon-nucleon interaction depicts excellent agreement with experiment for $2^+_1$ level energies, two-neutron separation energies, and reaction cross sections, corroborating the formation of a new nuclear shell above the $N$ = 34 shell. Its constituents, the $0f_{5/2}$ and $0g_{9/2}$ orbitals, are almost degenerate. This degeneracy precludes the possibility for a doubly magic $^{60}$Ca and potentially drives the dripline of Ca isotopes to $^{70}$Ca or even beyond., Comment: 8 pages, 3 figures

X-ray dark-field or ultra-small angle scatter imaging has become increasingly important since the introduction of phase-based x-ray imaging and is having transformative impact in fields such as in vivo lung imaging and explosives detection. Here we show that dark-field images acquired with the edge-illumination method (either in its traditional double mask or simplified single mask implementation) provide a direct measurement of the scattering function, which is unaffected by system-specific parameters such as the autocorrelation length. We show that this is a consequence both of the specific measurement setup and of the mathematical approach followed to retrieve the dark-field images. We show agreement with theoretical models for datasets acquired both with synchrotron and laboratory x-ray sources. We also introduce a new contrast mechanism, the variance of refraction, which is extracted from the same dataset and provides a direct link with the size of the scattering centres. We show that this can also be described by the same theoretical models. We study the behaviour of both signals vs. key parameters such as x-ray energy and scatterer radius. We find this allows quantitative, direct, multi-scale scattering measurements during imaging, with implications in all fields where dark-field imaging is used., Comment: 18 pages, 5 main figures, 6 extended data figures

We report on the curation of several publicly available datasets for age and gender prediction. Furthermore, we present experiments to predict age and gender with models based on a pre-trained wav2vec 2.0. Depending on the dataset, we achieve an MAE between 7.1 years and 10.8 years for age, and at least 91.1% ACC for gender (female, male, child). Compared to a modelling approach built on handcrafted features, our proposed system shows an improvement of 9% UAR for age and 4% UAR for gender. To make our findings reproducible, we release the best performing model to the community as well as the sample lists of the data splits., Comment: 5 pages, submitted to 15th ITG Conference on Speech Communication

Large pre-trained speech models are widely used as the de-facto paradigm, especially in scenarios when there is a limited amount of labeled data available. However, finetuning all parameters from the self-supervised learned model can be computationally expensive, and becomes infeasiable as the size of the model and the number of downstream tasks scales. In this paper, we propose a novel approach called Two Parallel Adapter (TPA) that is inserted into the conformer-based model pre-trained model instead. TPA is based on systematic studies of the residual adapter, a popular approach for finetuning a subset of parameters. We evaluate TPA on various public benchmarks and experiment results demonstrates its superior performance, which is close to the full finetuning on different datasets and speech tasks. These results show that TPA is an effective and efficient approach for serving large pre-trained speech models. Ablation studies show that TPA can also be pruned, especially for lower blocks.