Search

The discovery of pulsations in (at least) six ultraluminous X-ray sources (ULXs) has shown that neutron stars can accrete at (highly) super-Eddington rates, challenging the standard accretion theories. M51 ULX-7, with a spin signal of $P\simeq2.8$ s, is the pulsating ULX (PULX) with the shortest known orbital period ($P_\mathrm{orb}\simeq2$ d) and has been observed multiple times by XMM-Newton, Chandra, and NuSTAR. We report on the timing and spectral analyses of three XMM-Newton observations of M51 ULX-7 performed between the end of 2021 and the beginning of 2022, together with a timing re-analysis of XMM-Newton, Chandra, and NuSTAR archival observations. We investigated the spin signal by applying accelerated search techniques and studied the power spectrum through the fast Fourier transform, looking for (a)periodic variability in the source flux. We analysed the energy spectra of the 2021-2022 observations and compared them to the older ones. We report the discovery of a recurrent, significant ($>$3$\sigma$) broad complex at mHz frequencies in the power spectra of M51 ULX-7. We did not detect the spin signal, setting a 3$\sigma$ upper limit on the pulsed fraction of $\lesssim10\%$ for the single observation. The complex is significantly detected also in five Chandra observations performed in 2012. M51 ULX-7 represents the second PULX for which we have a significant detection of mHz-QPOs at super-Eddington luminosities. These findings suggest that one should avoid using the observed QPO frequency to infer the mass of the accretor in a ULX. The absence of spin pulsations when the broad complex is detected suggests that the mechanism responsible for the aperiodic modulation also dampens the spin signal's pulsed fraction. If true, this represents an additional obstacle in the detection of new PULXs, suggesting an even larger occurrence of PULXs among ULXs., Comment: 14 pages (12 main text + Appendix), 5 figures. Accepted for publication on A&A

We report on the multi-year evolution of the population of X-ray sources in the nuclear region of NGC 3621 based on Chandra, XMM-Newton and Swift observations. Among these, two sources, X1 and X5, after their first detection in 2008, seem to have faded below the detectability threshold, a most interesting fact as X1 is associated with the AGN of the galaxy. Two other sources, X3 and X6 are presented for the first time, the former showing a peculiar short-term variability in the latest available dataset, suggesting an egress from eclipse, hence belonging to the handful of known eclipsing ultra-luminous X-ray sources. One source, X4, previously known for its "heart-beat", i.e. a characteristic modulation in its signal with a period of $\approx1$ h, shows a steady behaviour in the latest observation. Finally, the brightest X-ray source in NGC 3621, here labelled X2, shows steady levels of flux across all the available datasets but a change in its spectral shape, reminiscent of the behaviours of Galactic disk-fed X-ray binaries., Comment: 11 pages, 10 figures, 3 tables, 2 appendices available online. Accepted for publication on A&A

In this study, we propose a novel surrogate modelling approach to efficiently and accurately approximate the response of complex dynamical systems driven by time-varying Recently, there has been increased interest in assessing the seismic fragility of industrial plants and process equipment. This is reflected in the growing number of studies, community-funded research projects and experimental campaigns on the matter.Nonetheless, the complexity of the problem and its inherent modelling, coupled with a general scarcity of available data on process equipment, has limited the development of risk assessment methods. In fact, these limitations have led to the creation of simplified and quick-to-run models. In this context, we propose an innovative framework for developing state-dependent fragility functions. This new methodology combines limited data with the power of metamodelling and statistical techniques, namely polynomial chaos expansions (PCE) and bootstrapping. Therefore, we validated the framework on a simplified and inexpensive-to-run MDoF system endowed with Bouc-Wen hysteresis.Then, we tested it on a real nonstructural industrial process component. Specifically, we applied the state-dependent fragility framework to a critical vertical tank of a multicomponent full-scale 3D steel braced frame (BF). The seismic performance of the BF endowed with process components was captured by means of shake table campaign within the European SPIF project. Finally, we derived state-dependent fragility functions based on the combination of PCE and bootstrap at a greatly reduced computational cost.

Reliability analysis aims at estimating the failure probability of an engineering system. It often requires multiple runs of a limit-state function, which usually relies on computationally intensive simulations. Traditionally, these simulations have been considered deterministic, i.e., running them multiple times for a given set of input parameters always produces the same output. However, this assumption does not always hold, as many studies in the literature report non-deterministic computational simulations (also known as noisy models). In such cases, running the simulations multiple times with the same input will result in different outputs. Similarly, data-driven models that rely on real-world data may also be affected by noise. This characteristic poses a challenge when performing reliability analysis, as many classical methods, such as FORM and SORM, are tailored to deterministic models. To bridge this gap, this paper provides a novel methodology to perform reliability analysis on models contaminated by noise. In such cases, noise introduces latent uncertainty into the reliability estimator, leading to an incorrect estimation of the real underlying reliability index, even when using Monte Carlo simulation. To overcome this challenge, we propose the use of denoising regression-based surrogate models within an active learning reliability analysis framework. Specifically, we combine Gaussian process regression with a noise-aware learning function to efficiently estimate the probability of failure of the underlying noise-free model. We showcase the effectiveness of this methodology on standard benchmark functions and a finite element model of a realistic structural frame.

Computer simulations (a.k.a. white-box models) are more indispensable than ever to model intricate engineering systems. However, computational models alone often fail to fully capture the complexities of reality. When physical experiments are accessible though, it is of interest to enhance the incomplete information offered by computational models. Gray-box modeling is concerned with the problem of merging information from data-driven (a.k.a. black-box) models and white-box (i.e., physics-based) models. In this paper, we propose to perform this task by using multi-fidelity surrogate models (MFSMs). A MFSM integrates information from models with varying computational fidelity into a new surrogate model. The multi-fidelity surrogate modeling framework we propose handles noise-contaminated data and is able to estimate the underlying noise-free high-fidelity function. Our methodology emphasizes on delivering precise estimates of the uncertainty in its predictions in the form of confidence and prediction intervals, by quantitatively incorporating the different types of uncertainty that affect the problem, arising from measurement noise and from lack of knowledge due to the limited experimental design budget on both the high- and low-fidelity models. Applied to gray-box modeling, our MFSM framework treats noisy experimental data as the high-fidelity and the white-box computational models as their low-fidelity counterparts. The effectiveness of our methodology is showcased through synthetic examples and a wind turbine application.

Survival models can help medical practitioners to evaluate the prognostic importance of clinical variables to patient outcomes such as mortality or hospital readmission and subsequently design personalized treatment regimes. Electronic Health Records (EHRs) hold the promise for large-scale survival analysis based on systematically recorded clinical features for each patient. However, existing survival models either do not scale to high dimensional and multi-modal EHR data or are difficult to interpret. In this study, we present a supervised topic model called MixEHR-SurG to simultaneously integrate heterogeneous EHR data and model survival hazard. Our contributions are three-folds: (1) integrating EHR topic inference with Cox proportional hazards likelihood; (2) integrating patient-specific topic hyperparameters using the PheCode concepts such that each topic can be identified with exactly one PheCode-associated phenotype; (3) multi-modal survival topic inference. This leads to a highly interpretable survival topic model that can infer PheCode-specific phenotype topics associated with patient mortality. We evaluated MixEHR-SurG using a simulated dataset and two real-world EHR datasets: the Quebec Congenital Heart Disease (CHD) data consisting of 8,211 subjects with 75,187 outpatient claim records of 1,767 unique ICD codes; the MIMIC-III consisting of 1,458 subjects with multi-modal EHR records. Compared to the baselines, MixEHR-SurG achieved a superior dynamic AUROC for mortality prediction, with a mean AUROC score of 0.89 in the simulation dataset and a mean AUROC of 0.645 on the CHD dataset. Qualitatively, MixEHR-SurG associates severe cardiac conditions with high mortality risk among the CHD patients after the first heart failure hospitalization and critical brain injuries with increased mortality among the MIMIC-III patients after their ICU discharge., Comment: 45 pages total, 19 pages main text, 6 main figures, 10 supplementary figures

The probabilistic design of offshore wind turbines aims to ensure structural safety in a cost-effective way. This involves conducting structural reliability assessments for different design options and considering different structural responses. There are several structural reliability methods, and this paper will apply and compare different approaches in some simplified case studies. In particular, the well known environmental contour method will be compared to a more novel approach based on sequential sampling and Gaussian processes regression for an ultimate limit state case study. For one of the case studies, results will also be compared to results from a brute force simulation approach. Interestingly, the comparison is very different from the two case studies. In one of the cases the environmental contours method agrees well with the sequential sampling method but in the other, results vary considerably. Probably, this can be explained by the violation of some of the assumptions associated with the environmental contour approach, i.e. that the short-term variability of the response is large compared to the long-term variability of the environmental conditions. Results from this simple comparison study suggests that the sequential sampling method can be a robust and computationally effective approach for structural reliability assessment.

Large language models (LLMs) are being increasingly incorporated into scientific workflows. However, we have yet to fully grasp the implications of this integration. How should the advent of large language models affect the practice of science? For this opinion piece, we have invited four diverse groups of scientists to reflect on this query, sharing their perspectives and engaging in debate. Schulz et al. make the argument that working with LLMs is not fundamentally different from working with human collaborators, while Bender et al. argue that LLMs are often misused and over-hyped, and that their limitations warrant a focus on more specialized, easily interpretable tools. Marelli et al. emphasize the importance of transparent attribution and responsible use of LLMs. Finally, Botvinick and Gershman advocate that humans should retain responsibility for determining the scientific roadmap. To facilitate the discussion, the four perspectives are complemented with a response from each group. By putting these different perspectives in conversation, we aim to bring attention to important considerations within the academic community regarding the adoption of LLMs and their impact on both current and future scientific practices.

Humanitarian organizations provide aid to people in need. To use their limited budget efficiently, their distribution processes must ensure that legitimate recipients cannot receive more aid than they are entitled to. Thus, it is essential that recipients can register at most once per aid program. Taking the International Committee of the Red Cross's aid distribution registration process as a use case, we identify the requirements to detect double registration without creating new risks for aid recipients. We then design Janus, which combines privacy-enhancing technologies with biometrics to prevent double registration in a safe manner. Janus does not create plaintext biometric databases and reveals only one bit of information at registration time (whether the user registering is present in the database or not). We implement and evaluate three instantiations of Janus based on secure multiparty computation, somewhat homomorphic encryption, and trusted execution environments. We demonstrate that they support the privacy, accuracy, and performance needs of humanitarian organizations. We compare Janus with existing alternatives and show it is the first system that provides the accuracy our scenario requires while providing strong protection.

Context. We have investigated the multiwavelength emission of PSR J2021+4026, the only isolated gamma-ray pulsar known to be variable, which in October 2011 underwent a simultaneous change in gamma-ray flux and spin-down rate, followed by a second mode change in February 2018. Multiwavelength monitoring is crucial to understand the physics behind these events and how they may have affected the structure of the magnetosphere. Aims.The monitoring of pulse profile alignment is a powerful diagnostic tool for constraining magnetospheric reconfiguration. We aim to investigate timing or flux changes related to the variability of PSR J2021+4026 via multiwavelength observations, including gamma-ray observations from Fermi-LAT, X-ray observations from XMM-Newton, and a deep optical observation with the Gran Telescopio Canarias.Methods. We performed a detailed comparison of the timing features of the pulsar in gamma and X-rays and searched for any change in phase lag between the phaseogram peaks in these two energy bands. Although previous observations did not detect a counterpart in visible light, we also searched for optical emission that might have increased due to the mode change, making this pulsar detectable in the optical. Results.We have found a change in the gamma-to X-ray pulse profile alignment by 0.21$\pm$0.02 in phase, which indicates that the first mode change affected different regions of the pulsar magnetosphere. No optical counterpart was detected down to g'=26.1 and r'=25.3. Conclusions.We suggest that the observed phase shift could be related to a reconfiguration of the connection between the quadrupole magnetic field near the stellar surface and the dipole field that dominates at larger distances. This is consistent with the picture of X-ray emission coming from the heated polar cap and with the simultaneous flux and frequency derivative change observed during the mode changes., Comment: 9 pages, 7 figures, 1 Table. Accepted for publication in Astronomy and Astrophysics (A&A)

We present 294 pulsars found in GeV data from the Large Area Telescope (LAT) on the Fermi Gamma-ray Space Telescope. Another 33 millisecond pulsars (MSPs) discovered in deep radio searches of LAT sources will likely reveal pulsations once phase-connected rotation ephemerides are achieved. A further dozen optical and/or X-ray binary systems co-located with LAT sources also likely harbor gamma-ray MSPs. This catalog thus reports roughly 340 gamma-ray pulsars and candidates, 10% of all known pulsars, compared to $\leq 11$ known before Fermi. Half of the gamma-ray pulsars are young. Of these, the half that are undetected in radio have a broader Galactic latitude distribution than the young radio-loud pulsars. The others are MSPs, with 6 undetected in radio. Overall, >235 are bright enough above 50 MeV to fit the pulse profile, the energy spectrum, or both. For the common two-peaked profiles, the gamma-ray peak closest to the magnetic pole crossing generally has a softer spectrum. The spectral energy distributions tend to narrow as the spindown power $\dot E$ decreases to its observed minimum near $10^{33}$ erg s$^{-1}$, approaching the shape for synchrotron radiation from monoenergetic electrons. We calculate gamma-ray luminosities when distances are available. Our all-sky gamma-ray sensitivity map is useful for population syntheses. The electronic catalog version provides gamma-ray pulsar ephemerides, properties and fit results to guide and be compared with modeling results., Comment: 142 pages. Accepted by the Astrophysical Journal Supplement

Optical projection tomography (OPT) is a powerful tool for biomedical studies. It achieves 3D visualization of mesoscopic biological samples with high spatial resolution using conventional tomographic-reconstruction algorithms. However, various artifacts degrade the quality of the reconstructed images due to experimental imperfections in the OPT instruments. While many efforts have been made to characterize and correct for these artifacts, they focus on one specific type of artifacts. This work has two contributions. First, we systematically document a catalog of mechanical artifacts based on a 3D description of the imaging system that uses a set of angular and translational parameters. Then, we introduce a calibration algorithm that recovers the unknown system parameters fed into the final 3D iterative reconstruction algorithm for a distortion-free volumetric image. Simulations with beads data and experimental results on a fluorescent textile fiber confirm that our algorithm successfully removes miscalibration artifacts in the reconstruction., Comment: Presented in ISCS2023. arXiv admin note: substantial text overlap with arXiv:2210.03513

Implementations of Markov chain Monte Carlo (MCMC) methods need to confront two fundamental challenges: accurate representation of prior information and efficient evaluation of likelihoods. Principal component analysis (PCA) and related techniques can in some cases facilitate the definition and sampling of the prior distribution, as well as the training of accurate surrogate models, using for instance, polynomial chaos expansion (PCE). However, complex geological priors with sharp contrasts necessitate more complex dimensionality-reduction techniques, such as, deep generative models (DGMs). By sampling a low-dimensional prior probability distribution defined in the low-dimensional latent space of such a model, it becomes possible to efficiently sample the physical domain at the price of a generator that is typically highly non-linear. Training a surrogate that is capable of capturing intricate non-linear relationships between latent parameters and outputs of forward modeling presents a notable challenge. Indeed, while PCE models provide high accuracy when the input-output relationship can be effectively approximated by relatively low-degree multivariate polynomials, this condition is typically not met when employing latent variables derived from DGMs. In this contribution, we present a strategy combining the excellent reconstruction performances of a variational autoencoder (VAE) with the accuracy of PCA-PCE surrogate modeling in the context of Bayesian ground penetrating radar (GPR) traveltime tomography. Within the MCMC process, the parametrization of the VAE is leveraged for prior exploration and sample proposals. Concurrently, surrogate modeling is conducted using PCE, which operates on either globally or locally defined principal components of the VAE samples under examination., Comment: 25 pages, 15 figures

We propose a novel surrogate modelling approach to efficiently and accurately approximate the response of complex dynamical systems driven by time-varying exogenous excitations over extended time periods. Our approach, namely manifold nonlinear autoregressive modelling with exogenous input (mNARX), involves constructing a problem-specific exogenous input manifold that is optimal for constructing autoregressive surrogates. The manifold, which forms the core of mNARX, is constructed incrementally by incorporating the physics of the system, as well as prior expert- and domain- knowledge. Because mNARX decomposes the full problem into a series of smaller sub-problems, each with a lower complexity than the original, it scales well with the complexity of the problem, both in terms of training and evaluation costs of the final surrogate. Furthermore, mNARX synergizes well with traditional dimensionality reduction techniques, making it highly suitable for modelling dynamical systems with high-dimensional exogenous inputs, a class of problems that is typically challenging to solve. Since domain knowledge is particularly abundant in physical systems, such as those found in civil and mechanical engineering, mNARX is well suited for these applications. We demonstrate that mNARX outperforms traditional autoregressive surrogates in predicting the response of a classical coupled spring-mass system excited by a one-dimensional random excitation. Additionally, we show that mNARX is well suited for emulating very high-dimensional time- and state-dependent systems, even when affected by active controllers, by surrogating the dynamics of a realistic aero-servo-elastic onshore wind turbine simulator. In general, our results demonstrate that mNARX offers promising prospects for modelling complex dynamical systems, in terms of accuracy and efficiency.

System reliability analysis aims at computing the probability of failure of an engineering system given a set of uncertain inputs and limit state functions. Active-learning solution schemes have been shown to be a viable tool but as of yet they are not as efficient as in the context of component reliability analysis. This is due to some peculiarities of system problems, such as the presence of multiple failure modes and their uneven contribution to failure, or the dependence on the system configuration (e.g., series or parallel). In this work, we propose a novel active learning strategy designed for solving general system reliability problems. This algorithm combines subset simulation and Kriging/PC-Kriging, and relies on an enrichment scheme tailored to specifically address the weaknesses of this class of methods. More specifically, it relies on three components: (i) a new learning function that does not require the specification of the system configuration, (ii) a density-based clustering technique that allows one to automatically detect the different failure modes, and (iii) sensitivity analysis to estimate the contribution of each limit state to system failure so as to select only the most relevant ones for enrichment. The proposed method is validated on two analytical examples and compared against results gathered in the literature. Finally, a complex engineering problem related to power transmission is solved, thereby showcasing the efficiency of the proposed method in a real-case scenario.

Over the last years, advancements in deep learning models for computer vision have led to a dramatic improvement in their image classification accuracy. However, models with a higher accuracy in the task they were trained on do not necessarily develop better image representations that allow them to also perform better in other tasks they were not trained on. In order to investigate the representation learning capabilities of prominent high-performing computer vision models, we investigated how well they capture various indices of perceptual similarity from large-scale behavioral datasets. We find that higher image classification accuracy rates are not associated with a better performance on these datasets, and in fact we observe no improvement in performance since GoogLeNet (released 2015) and VGG-M (released 2014). We speculate that more accurate classification may result from hyper-engineering towards very fine-grained distinctions between highly similar classes, which does not incentivize the models to capture overall perceptual similarities.

Time-domain analysis of an archival XMM-Newton observation unveiled a very unusual variability pattern in the soft X-ray emission of PSR J1311-3430, a black widow millisecond pulsar in a tight binary (P_B=93.8 min) with a very low-mass (M~0.01 Msun) He companion star, known to show flaring emission in the optical and in the X-rays. A series of six pulses with a regular recurrence time of ~124 min is apparent in the 0.2-10 keV light curve of the system, also featuring an initial, bright flare and a quiescent phase lasting several hours. The X-ray spectrum does not change when the pulses are seen and is consistent with a power law with photon index Gamma~1.6, also describing the quiescent emission. The peak luminosity of the pulses is of several 10^32 erg/s. Simultaneous observations in the U band with the Optical Monitor onboard XMM and in the g' band from the Las Cumbres Observatory do not show any apparent counterpart of the pulses and only display the well-known orbital modulation of the system. We consider different hypotheses to explain the recurrent pulses: we investigate their possible analogy with other phenomena already observed in this pulsar and in similar systems and we also study possible explanations related to the interaction of the energetic pulsar wind with intrabinary material, but we found none of these pictures to be convincing. We identify simultaneous X-ray observations and optical spectroscopy as a possible way to constrain the nature of the phenomenon., Comment: Accepted for publication in Astronomy & Astrophysics as a Letter

Optical projection tomography (OPT) is a powerful tool for biomedical studies. It achieves 3D visualization of mesoscopic biological samples with high spatial resolution using conventional tomographic-reconstruction algorithms. However, various artifacts degrade the quality of the reconstructed images due to experimental imperfections in the OPT instruments. While many efforts have been made to characterize and correct for these artifacts, they focus on one specific type of artifacts, whereas a comprehensive catalog of all sorts of mechanical artifacts does not currently exist. In this work, we systematically document many mechanical artifacts. We rely on a 3D description of the imaging system that uses a set of angular and translational parameters. We provide a catalog of artifacts. It lists their cause, resulting effects, and existing correction methods. Then, we introduce an automatic calibration algorithm that is able to recover the unknown system parameters fed into the final 3D iterative reconstruction algorithm for a distortion-free volumetric image. Simulations with beads data and experimental results on a fluorescent textile fiber confirm that our algorithm successfully removes miscalibration artifacts in the reconstruction.

The Athena X-ray Integral Unit (X-IFU) is the high resolution X-ray spectrometer, studied since 2015 for flying in the mid-30s on the Athena space X-ray Observatory, a versatile observatory designed to address the Hot and Energetic Universe science theme, selected in November 2013 by the Survey Science Committee. Based on a large format array of Transition Edge Sensors (TES), it aims to provide spatially resolved X-ray spectroscopy, with a spectral resolution of 2.5 eV (up to 7 keV) over an hexagonal field of view of 5 arc minutes (equivalent diameter). The X-IFU entered its System Requirement Review (SRR) in June 2022, at about the same time when ESA called for an overall X-IFU redesign (including the X-IFU cryostat and the cooling chain), due to an unanticipated cost overrun of Athena. In this paper, after illustrating the breakthrough capabilities of the X-IFU, we describe the instrument as presented at its SRR, browsing through all the subsystems and associated requirements. We then show the instrument budgets, with a particular emphasis on the anticipated budgets of some of its key performance parameters. Finally we briefly discuss on the ongoing key technology demonstration activities, the calibration and the activities foreseen in the X-IFU Instrument Science Center, and touch on communication and outreach activities, the consortium organisation, and finally on the life cycle assessment of X-IFU aiming at minimising the environmental footprint, associated with the development of the instrument. Thanks to the studies conducted so far on X-IFU, it is expected that along the design-to-cost exercise requested by ESA, the X-IFU will maintain flagship capabilities in spatially resolved high resolution X-ray spectroscopy, enabling most of the original X-IFU related scientific objectives of the Athena mission to be retained. (abridged)., Comment: 48 pages, 29 figures, Accepted for publication in Experimental Astronomy with minor editing

Stochastic simulators are non-deterministic computer models which provide a different response each time they are run, even when the input parameters are held at fixed values. They arise when additional sources of uncertainty are affecting the computer model, which are not explicitly modeled as input parameters. The uncertainty analysis of stochastic simulators requires their repeated evaluation for different values of the input variables, as well as for different realizations of the underlying latent stochasticity. The computational cost of such analyses can be considerable, which motivates the construction of surrogate models that can approximate the original model and its stochastic response, but can be evaluated at much lower cost. We propose a surrogate model for stochastic simulators based on spectral expansions. Considering a certain class of stochastic simulators that can be repeatedly evaluated for the same underlying random event, we view the simulator as a random field indexed by the input parameter space. For a fixed realization of the latent stochasticity, the response of the simulator is a deterministic function, called trajectory. Based on samples from several such trajectories, we approximate the latter by sparse polynomial chaos expansion and compute analytically an extended Karhunen-Lo\`eve expansion (KLE) to reduce its dimensionality. The uncorrelated but dependent random variables of the KLE are modeled by advanced statistical techniques such as parametric inference, vine copula modeling, and kernel density estimation. The resulting surrogate model approximates the marginals and the covariance function, and allows to obtain new realizations at low computational cost. We observe that in our numerical examples, the first mode of the KLE is by far the most important, and investigate this phenomenon and its implications.

Language models demonstrate both quantitative improvement and new qualitative capabilities with increasing scale. Despite their potentially transformative impact, these new capabilities are as yet poorly characterized. In order to inform future research, prepare for disruptive new model capabilities, and ameliorate socially harmful effects, it is vital that we understand the present and near-future capabilities and limitations of language models. To address this challenge, we introduce the Beyond the Imitation Game benchmark (BIG-bench). BIG-bench currently consists of 204 tasks, contributed by 450 authors across 132 institutions. Task topics are diverse, drawing problems from linguistics, childhood development, math, common-sense reasoning, biology, physics, social bias, software development, and beyond. BIG-bench focuses on tasks that are believed to be beyond the capabilities of current language models. We evaluate the behavior of OpenAI's GPT models, Google-internal dense transformer architectures, and Switch-style sparse transformers on BIG-bench, across model sizes spanning millions to hundreds of billions of parameters. In addition, a team of human expert raters performed all tasks in order to provide a strong baseline. Findings include: model performance and calibration both improve with scale, but are poor in absolute terms (and when compared with rater performance); performance is remarkably similar across model classes, though with benefits from sparsity; tasks that improve gradually and predictably commonly involve a large knowledge or memorization component, whereas tasks that exhibit "breakthrough" behavior at a critical scale often involve multiple steps or components, or brittle metrics; social bias typically increases with scale in settings with ambiguous context, but this can be improved with prompting., Comment: 27 pages, 17 figures + references and appendices, repo: https://github.com/google/BIG-bench

We investigate the linear quadratic Gaussian Stackelberg game under a class of nested observation information pattern. Two decision makers implement control strategies relying on different information sets: The follower uses its observation data to design its strategy, whereas the leader implements its strategy using global observation data. We show that the solution requires solving a new type of forward-backward stochastic differential equations whose drift terms contain two types of conditional expectation terms associated to the adjoint variables. We then propose a method to find the functional relations between each adjoint pair, i.e., each pair formed by an adjoint variable and the conditional expectation of its associated state. The proposed method follows a layered pattern. More precisely, in the inner layer, we seek the functional relation for the adjoint pair under the sigma-sub-algebra generated by follower's observation information; and in the outer layer, we look for the functional relation for the adjoint pair under the sigma-sub-algebra generated by leader's observation information. Our result shows that the optimal open-loop solution admits an explicit feedback type representation. More precisely, the feedback coefficient matrices satisfy tuples of coupled forward-backward differential Riccati equations, and feedback variables are computed by Kalman-Bucy filtering.

Transformer-based architectures are the model of choice for natural language understanding, but they come at a significant cost, as they have quadratic complexity in the input length, require a lot of training data, and can be difficult to tune. In the pursuit of lower costs, we investigate simple MLP-based architectures. We find that existing architectures such as MLPMixer, which achieves token mixing through a static MLP applied to each feature independently, are too detached from the inductive biases required for natural language understanding. In this paper, we propose a simple variant, HyperMixer, which forms the token mixing MLP dynamically using hypernetworks. Empirically, we demonstrate that our model performs better than alternative MLP-based models, and on par with Transformers. In contrast to Transformers, HyperMixer achieves these results at substantially lower costs in terms of processing time, training data, and hyperparameter tuning., Comment: Published at ACL 2023

XMM-Newton provides unprecedented insight into the X-ray Universe, recording variability information for hundreds of thousands of sources. Manually searching for interesting patterns in light curves is impractical, requiring an automated data-mining approach for the characterization of sources. Straightforward fitting of temporal models to light curves is not a sure way to identify them, especially with noisy data. We used unsupervised machine learning to distill a large data set of light-curve parameters, revealing its clustering structure in preparation for anomaly detection and subsequent searches for specific source behaviors (e.g., flares, eclipses). Self-organizing maps (SOMs) achieve dimensionality reduction and clustering within a single framework. They are a type of artificial neural network trained to approximate the data with a two-dimensional grid of discrete interconnected units, which can later be visualized on the plane. We trained our SOM on temporal-only parameters computed from more than 100,000 detections from the EXTraS catalog. The resulting map reveals that about 2500 most variable sources are clustered based on temporal characteristics. We find distinctive regions of the SOM map associated with flares, eclipses, dips, linear light curves, and others. Each group contains sources that appear similar by eye. We single out a handful of interesting sources for further study. The condensed view of our dataset provided by SOMs allowed us to identify groups of similar sources, speeding up manual characterization by orders of magnitude. Our method also highlights problems with fitting simple temporal models to light curves and can be used to mitigate them to an extent. This will be crucial for fully exploiting the high data volume expected from upcoming X-ray surveys, and may also help with interpreting supervised classification models., Comment: Accepted in Astronomy & Astrophysics (A&A). Main part: 12 pages, 12 figures, 1 video (appendix: 5 pages and 4 figures)

We report on the systematic analysis of the X-ray observations of the ultra-luminous X-ray source XMMU J122939.7+075333 located in the globular cluster RZ 2109 in the Virgo galaxy NGC 4472. The inclusion of observations and time intervals ignored in previous works and the careful selection of extraction regions and energy bands have allowed us to identify new flaring episodes, in addition to the ones that made it one of the best black hole candidates in globular clusters. Although most observations are too short and sparse to recognize a regular pattern, the spacing of the three most recent X-ray flares is compatible with a ~34 hours recurrence time. If confirmed by future observations, such behavior, together with the soft spectrum of the X-ray flares, would be strikingly similar to the quasi-periodic eruptions recently discovered in galactic nuclei. Following one of the possible interpretations of these systems and of a peculiar class of extra-galactic X-ray transients, we explore the possibility that XMMU J122939.7+075333 might be powered by the partial disruption of a white dwarf by an intermediate mass (M~700 Msun) black hole., Comment: 12 pages, 10 figures, accepted for publication in A&A

We analyzed the unexposed to the sky outFOV region of the MOS2 detector on board XMM-Newton covering 15 years of data amounting to 255 Ms. We show convincing evidence that the origin of the unfocused background in XMM-Newton is due to energetic protons, electrons and hard X-ray photons. Galactic Cosmic Rays are the main contributors as shown by the tight correlation (2.6% of total scatter) with 1 GeV protons data of the SOHO EPHIN detector. Tight correlations are found with a proxy of the Chandra background rate, revealing the common source of background for detectors in similar orbits, and with the data of the EPIC Radiation Monitor (ERM), only when excluding Solar Energetic Particles events (SEPs). The entrance to the outer electron belts is associated to a sudden increase in the outFOV MOS2 rate and a spectral change. These facts support the fact that MeV electrons can generate an unfocused background signal. The correlation between MOS2 outFOV data and the SOHO EPHIN data reveals a term constant in time and isotropic similar to the one found in the study of the pn data. The most plausible origin of this component is hard unfocused X-ray photons of the Cosmic X-ray Background (CXB) Compton-scattering in the detector as supported by the strength of the signal in the two detectors with different thicknesses. Based on this physical understanding a particle radiation monitor on board ATHENA has been proposed and it is currently under study. It will be able to track different species with the necessary accuracy and precision to guarantee the challenging requirement of 2% reproducibility of the background., Comment: 14 pages, 15 figures, accepted for publication in ApJ

We present a Bayesian tomography framework operating with prior-knowledge-based parametrization that is accelerated by surrogate models. Standard high-fidelity forward solvers solve wave equations with natural spatial parametrizations based on fine discretization. Similar parametrizations, typically involving tens of thousand of variables, are usually employed to parameterize the subsurface in tomography applications. When the data do not allow to resolve details at such finely parameterized scales, it is often beneficial to instead rely on a prior-knowledge-based parametrization defined on a lower dimension domain (or manifold). Due to the increased identifiability in the reduced domain, the concomitant inversion is better constrained and generally faster. We illustrate the potential of a prior-knowledge-based approach by considering ground penetrating radar (GPR) travel-time tomography in a crosshole configuration. An effective parametrization of the input (i.e., the permittivity distributions) and compression in the output (i.e., the travel-time gathers) spaces are achieved via data-driven principal component decomposition based on random realizations of the prior Gaussian-process model with a truncation determined by the performances of the standard solver on the full and reduced model domains. To accelerate the inversion process, we employ a high-fidelity polynomial chaos expansion (PCE) surrogate model. We show that a few hundreds design data sets is sufficient to provide reliable Markov chain Monte Carlo inversion. Appropriate uncertainty quantification is achieved by reintroducing the truncated higher-order principle components in the original model space after inversion on the manifold and by adapting a likelihood function that accounts for the fact that the truncated higher-order components are not completely located in the null-space., Comment: 27 pages, 14 images