Search

Self-attention in transformer models is an incremental associative memory that maps key vectors to value vectors. One way to speed up self-attention is to employ GPU-compliant vector search algorithms, yet the standard partitioning methods yield poor results in this context, because (1) keys and queries follow different distributions and (2) the effect of RoPE positional encoding. In this paper, we introduce SAAP (Self-Attention with Asymmetric Partitions), which overcomes these problems. It is an asymmetrical indexing technique that employs distinct partitions for keys and queries, thereby approximating self-attention with a data-adaptive sparsity pattern. It works on pretrained language models without finetuning, as it only requires to train (offline) a small query classifier. On a long context Llama 3.1-8b model, with sequences ranging from 100k to 500k tokens, our method typically reduces by a factor 20 the fraction of memory that needs to be looked-up, which translates to a time saving of 60\% when compared to FlashAttention-v2.

Change-point models are frequently considered when modeling phenomena where a regime shift occurs at an unknown time. In ageing research, these models are commonly adopted to estimate of the onset of cognitive decline. Yet commonly used models present several limitations. Here, we present a Bayesian non-linear mixed-effects model based on a differential equation designed for longitudinal studies to overcome some limitations of classical change point models used in ageing research. We demonstrate the ability of the proposed model to avoid biases in estimates of the onset of cognitive impairment in a simulated study. Finally, the methodology presented in this work is illustrated by analysing results from memory tests from older adults who participated in the English Longitudinal Study of Ageing.

The old G3V star Kepler-10 is known to host two transiting planets, the ultra-short-period super-Earth Kepler-10b ($P=0.837$ d; $R_{\rm p}=1.47~\rm R_\oplus$) and the long-period sub-Neptune Kepler-10c ($P=45.294$ d; $R_{\rm p}=2.35~\rm R_\oplus$), and a non-transiting planet causing variations in the Kepler-10c transit times. Measurements of the mass of Kepler-10c in the literature have shown disagreement, depending on the radial-velocity dataset and/or the modeling technique used. Here we report on the analysis of almost 300 high-precision radial velocities gathered with the HARPS-N spectrograph at the Telescopio Nazionale Galileo over $\sim11$~years, and extracted with the YARARA-v2 tool correcting for possible systematics and/or low-level activity variations at the spectrum level. To model these radial velocities, we used three different noise models and various numerical techniques, which all converged to the solution: $M_{\rm p, b}=3.24 \pm 0.32~\rm M_\oplus$ (10$\sigma$) and $\rho_{\rm p, b}=5.54 \pm 0.64~\rm g\;cm^{-3}$ for planet b; $M_{\rm p, c}=11.29 \pm 1.24~\rm M_\oplus$ (9$\sigma$) and $\rho_{\rm p, c}=4.75 \pm 0.53~\rm g\;cm^{-3}$ for planet c; and $M_{\rm p, d}\sin{i}=12.00 \pm 2.15~\rm M_\oplus$ (6$\sigma$) and $P=151.06 \pm 0.48$ d for the non-transiting planet Kepler-10d. This solution is further supported by the analysis of the Kepler-10c transit timing variations and their simultaneous modeling with the HARPS-N radial velocities. While Kepler-10b is consistent with a rocky composition and a small or no iron core, Kepler-10c may be a water world that formed beyond the water snowline and subsequently migrated inward., Comment: 15 pages, 12 figures, 5 tables, accepted for publication in Astronomy and Astrophysics

A major open issue concerning the active Sun is the effectiveness with which magnetic reconnection accelerates electrons in flares. A paper published by {\em{Nature}} in 2022 used microwave observations to conclude that the Sun is an almost ideal accelerator, energizing nearly all electrons within a coronal volume to nonthermal energies. Shortly thereafter, a paper published in {\em{Astrophysical Journal Letters}} used hard X-ray measurements \emph{of the same event} to reach the contradictory conclusion that less than 1\% of the available electrons were accelerated. Here we address this controversy by using spatially resolved observations of hard X-ray emission and a spectral inversion method to determine the evolution of the electron spectrum throughout the flare. So we estimated the density of the medium where electrons accelerate and, from this, the ratio of accelerated to ambient electron densities. Results show that this ratio never exceeds a percent or so in the cases analyzed.

The RPCs employed at the LHC experiments are currently operated in avalanche mode, with a mixture containing a large fraction of C$_{2}$H$_{2}$F$_{4}$ ($\approx$90\% or more) with the addition of i-C$_{4}$H$_{10}$ and SF$_{6}$ in different concentrations. However, C$_{2}$H$_{2}$F$_{4}$ and SF$_{6}$ are fluorinated greenhouse gases (F-gases) with Global Warming Potential (GWP) of $\approx$1400 and $\approx$22800, respectively. EU regulations imposed a progressive phase-down of C$_{2}$H$_{2}$F$_{4}$ production and consumption, aiming at strongly reducing its emission. This is already resulting in an increase of its price and reduction in availability. The most desirable long-term solution to this problem is to find an alternative, F-gases-free gas mixture, able to maintain similar detector performance. To address this challenge, the RPC ECOGasas@GIF++ collaboration (including RPC experts of ALICE, ATLAS, CMS, SHiP/LHCb, and the CERN EP-DT group) was created in 2019. The collaboration is currently studying a gas from the olefine family, the C$_{3}$H$_{2}$F$_{4}$ (or simply HFO, with GWP $\approx$6), to be used, in combination with CO$_{2}$, as a substitute for C$_{2}$H$_{2}$F$_{4}$. This contribution will focus on the signal shape studies that have been carried out by the collaboration during dedicated beam test periods. The methodology used in the data analysis will be presented, together with the results obtained with several HFO-based gas mixtures, and with the currently employed one. Furthermore, results on the counting-rate dependence of the RPC performance, obtained by combining the muon beam with the GIF++ $^{137}$Cs source with different attenuation factors, will also be presented.

We study the flexible piecewise exponential model in a high dimensional setting where the number of covariates $p$ grows proportionally to the number of observations $n$ and under the hypothesis of random uncorrelated Gaussian designs. We prove rigorously that the optimal ridge penalized log-likelihood of the model converges in probability to the saddle point of a surrogate objective function. The technique of proof is the Convex Gaussian Min-Max theorem of Thrampoulidis, Oymak and Hassibi. An important consequence of this result, is that we can study the impact of the ridge regularization on the estimates of the parameter of the model and the prediction error as a function of the ratio $p/n > 0$. Furthermore, these results represent a first step toward rigorously proving the (conjectured) correctness of several results obtained with the heuristic replica method for the Cox semi-parametric model.

Research into the polychromy of Greek and Roman sculptures has surged to explore the hypothesis that ancient sculptures were originally not pristine white but adorned with colors. Multispectral and multimodal imaging techniques have been crucial in studying painted surfaces, revealing polychromies even in traces. In fact, imaging techniques, such as reflectance and fluorescence, can identify different materials and map inhomogeneities, guiding further investigations such as Raman, XRays Fluorescence, and Fourier Transform InfraRed Spectroscopy (FTIR) to investigate residual colors. However, this approach may underestimate the original polychromies' extent over the complex articulation of a sculptured surface. This study proposes a methodology to analyze the original appearance of ancient sculptures using reality-based 3D models with textures not limited to those visible to the naked eye. We employ Visible Reflected Imaging (VIS) and Ultraviolet-induced Fluorescence Imaging (UVF). From the UVF and VIS datasets, the underlying 3D model is built by means of photogrammetry. Through raw data processing, images taken with different illuminating sources are successfully aligned and processed, creating a single 3D model with multiple textures mapped onto the same bi-dimensional space. The pixel-to-pixel correspondence of different textures allows for the implementation of a classification algorithm that can directly map its outcome onto the 3D model surface. This enables conservators to deepen their understanding of artifact preservation, observe mate-rial distribution in detail, and correlate this with 3D geometrical data. In this study, we experiment with this approach on an ancient Roman sculpture of Artemis, conserved at the Archeological and Art Museum of Maremma (MAAM) in Grosseto, Italy., Comment: 14 pages, 5 figures, to be published in the proceedings of "Heri-Tech - The Future of Heritage Science And Technologies" Conference by Springer, 29-30 April 2024, Florence, Italy (https://www.florenceheritech.com/)

The Cloud-Edge continuum enhances application performance by bringing computation closer to data sources. However, it presents considerable challenges in managing resources and determining service placement, as these tasks require navigating diverse, dynamic environments characterised by fluctuating network conditions. Addressing these challenges calls for tools combining simulation and emulation of Cloud-Edge systems to rigorously assess novel application and resource management strategies. In this paper, we introduce ECLYPSE, a Python-based framework that enables the simulation and emulation of the Cloud-Edge continuum via adaptable resource allocation and service placement models. ECLYPSE features an event-driven architecture for dynamically adapting network configurations and resources. It also supports seamless transitions between simulated and emulated setups. In this work, ECLYPSE capabilities are illustrated over three use cases, showing how the framework supports rapid prototyping across diverse experimental settings.

The fragmentation cross sections of carbon ion beams with kinetic energies of $115 - 353~\text{MeV/u}$ impinging on thin targets of graphite (C), polyvinyl-toluene (C$_9$H$_{10}$) and PMMA (C$_2$O$_5$H$_8$) have been measured at 90$^{\text{o}}$ and 60$^{\text{o}}$ at the CNAO particle therapy center (Pavia, Italy). The presented measurements are a complete reanalysis by the FOOT collaboration of already published elemental cross section on composite targets, in order to refine the analysis, improve the systematic uncertainties and show the comparison with the FLUKA Monte Carlo code calculations. In this work, the kinetic energy at production of measured fragments has been completely redefined, together with the efficiencies computation. The new analysis strategy has been successfully validated against the Monte Carlo cross sections. Two detection arms were positioned at two different angles to perform the measurement at 90$^{\text{o}}$ and 60$^{\text{o}}$. The fragment species have been identified in charge (Z$_{id}$ = H) and mass (M$_{id}$ = $^1$H, $^2$H, $^3$H) combining the information of the deposited energy in thin plastic scintillators, of the deposited energy in a thick LYSO crystal and of the fragments Time of Flight (ToF) measurement. The ToF was also used to compute the fragments measured kinetic energy. The cross sections are presented as a function of the fragments kinetic energy at production thanks to an unfolding technique applied to the data., Comment: Corresponding Author: Marco Toppi, marco.toppi@uniroma1.it

This paper presents the measurements of the angular differential cross sections for the forward production of He, Li, Be, B, C and N nuclei in the fragmentation process of a 400$\text{MeV/nucleon}$ $^{16}\text{O}$ beam interacting with a graphite target. Due to the limited data available in this energy regime, these measurements of nuclear fragmentation cross sections are relevant to improve nuclear interaction models for Particle Therapy and space radioprotection applications. The data analyzed in this paper were collected during a measurement campaign carried out at the GSI Helmholtz Center for Heavy Ion Research facility in Darmstadt (Germany) in 2021 by the FOOT collaboration. The results are compared with similar results found in the literature and with a previous FOOT measurement of the same process, using the same setup, from a previous pilot run performed at GSI in 2019. The pilot run data, however, had limited statistics and only allowed for the measurement of elemental fragmentation cross sections integrated in the setup acceptance. The 2021 data set, with statistics more than 100 times larger compared to the data collected in 2019, enabled the measurement of angular differential cross sections, fully exploiting the granularity of the FOOT $\Delta \text{E}$-TOF system. Furthermore, a better comprehension of the FOOT apparatus allowed to improve the analysis techniques, leading to a reduction in the final systematic uncertainties.

The synthesis of thinned isophoric arrays (TIAs) radiating mask-constrained patterns is addressed. By leveraging on the recently-introduced formulation of the design of antenna arrays in the autocorrelation-domain (AD), the TIA synthesis is recast as the matching of a target autocorrelation function derived from the user-defined guidelines and objectives. By exploiting the autocorrelation invariance of cyclic binary sequences, the AD solution space is significantly reduced and it is efficiently sampled by means of a discrete hybrid optimization approach. Two possible implementations of the AD-based TIA formulation are discussed and assessed in a set of representative numerical examples concerned with both ideal and real radiators, which are full-wave modeled to account for the mutual coupling effects. Comparisons with traditional pattern-domain (PD) synthesis methods are also considered to point out the features and the advantages of AD-based approaches.

Distributed training of large models consumes enormous computation resources and requires substantial engineering efforts to compose various training techniques. This paper presents SimpleFSDP, a PyTorch-native compiler-based Fully Sharded Data Parallel (FSDP) framework, which has a simple implementation for maintenance and composability, allows full computation-communication graph tracing, and brings performance enhancement via compiler backend optimizations. SimpleFSDP's novelty lies in its unique $torch.compile$-friendly implementation of collective communications using existing PyTorch primitives, namely parametrizations, selective activation checkpointing, and DTensor. It also features the first-of-its-kind intermediate representation (IR) nodes bucketing and reordering in the TorchInductor backend for effective computation-communication overlapping. As a result, users can employ the aforementioned optimizations to automatically or manually wrap model components for minimal communication exposure. Extensive evaluations of SimpleFSDP on Llama 3 models (including the ultra-large 405B) using TorchTitan demonstrate up to 28.54% memory reduction and 68.67% throughput improvement compared to the most widely adopted FSDP2 eager framework, when composed with other distributed training techniques.

This paper aims to investigate a reaction-diffusion model which describes in-host infection for Mycobacterium tuberculosis (Mtb) allowing random motion (i.e. linear diffusion) and chemotaxis (i.e. non-linear diffusion) of macrophages and bacteria populations. In particular, chemotaxis-driven aggregation of macrophages plays a fundamental role in the development of the Mtb infection and the production of chemokine - located in the infection site - represents an attractant for the uninfected macrophages, therefore we consider chemotaxis between infected macrophages and uninfected macrophages. The linear stability of the endemic equilibria is investigated. In particular, by looking for conditions guaranteeing that an equilibrium, stable in the absence of diffusion, becomes unstable when diffusion is allowed, the formation of Turing patterns - that biologically represent the formation of granuloma initiated by the immune cells - is investigated. The weakly nonlinear analysis is performed to deeply explore the patterns amplitude. Furthermore, via numerical simulations, the formation of Turing patterns is analysed.

Speaker verification systems have seen significant advancements with the introduction of Multi-scale Feature Aggregation (MFA) architectures, such as MFA-Conformer and ECAPA-TDNN. These models leverage information from various network depths by concatenating intermediate feature maps before the pooling and projection layers, demonstrating that even shallower feature maps encode valuable speaker-specific information. Building upon this foundation, we propose a Multi-scale Feature Contrastive (MFCon) loss that directly enhances the quality of these intermediate representations. Our MFCon loss applies contrastive learning to all feature maps within the network, encouraging the model to learn more discriminative representations at the intermediate stage itself. By enforcing better feature map learning, we show that the resulting speaker embeddings exhibit increased discriminative power. Our method achieves a 9.05% improvement in equal error rate (EER) compared to the standard MFA-Conformer on the VoxCeleb-1O test set.

Generative artificial intelligence (AI) technology is revolutionizing the computing industry. Not only its applications have broadened to various sectors but also poses new system design and optimization opportunities. The technology is capable of understanding and responding in multiple modalities. However, the advanced capability currently comes with significant system resource demands. To sustainably scale generative AI capabilities to billions of users in the world, inference must be fast and efficient. This paper pinpoints key system design and optimization opportunities by characterizing a family of emerging multi-modal generation models on real systems. Auto-regressive token generation is a critical latency performance bottleneck, typically dominated by GPU idle time. In addition to memory-intensive attention across the generative AI models, linear operations constitute significant inference latency due to the feed forward networks in Transformer-based models. We demonstrate that state-of-the-art optimization levers, spanning from applications to system software and hardware, set a 3.88x better baseline., Comment: 13 pages including references. 8 Figures. Under review to HPCA 2025 Industry Track

Neural codecs have become crucial to recent speech and audio generation research. In addition to signal compression capabilities, discrete codecs have also been found to enhance downstream training efficiency and compatibility with autoregressive language models. However, as extensive downstream applications are investigated, challenges have arisen in ensuring fair comparisons across diverse applications. To address these issues, we present a new open-source platform ESPnet-Codec, which is built on ESPnet and focuses on neural codec training and evaluation. ESPnet-Codec offers various recipes in audio, music, and speech for training and evaluation using several widely adopted codec models. Together with ESPnet-Codec, we present VERSA, a standalone evaluation toolkit, which provides a comprehensive evaluation of codec performance over 20 audio evaluation metrics. Notably, we demonstrate that ESPnet-Codec can be integrated into six ESPnet tasks, supporting diverse applications., Comment: Accepted by SLT

Speaker verification systems are crucial for authenticating identity through voice. Traditionally, these systems focus on comparing feature vectors, overlooking the speech's content. However, this paper challenges this by highlighting the importance of phonetic dominance, a measure of the frequency or duration of phonemes, as a crucial cue in speaker verification. A novel Phoneme Debiasing Attention Framework (PDAF) is introduced, integrating with existing attention frameworks to mitigate biases caused by phonetic dominance. PDAF adjusts the weighting for each phoneme and influences feature extraction, allowing for a more nuanced analysis of speech. This approach paves the way for more accurate and reliable identity authentication through voice. Furthermore, by employing various weighting strategies, we evaluate the influence of phonetic features on the efficacy of the speaker verification system., Comment: Accepted to SLT

The Spectrometer/Telescope for Imaging X-rays (STIX) on-board the ESA Solar Orbiter mission retrieves the coordinates of solar flare locations by means of a specific sub-collimator, named the Coarse Flare Locator (CFL). When a solar flare occurs on the Sun, the emitted X-ray radiation casts the shadow of a peculiar "H-shaped" tungsten grid over the CFL X-ray detector. From measurements of the areas of the detector that are illuminated by the X-ray radiation, it is possible to retrieve the $(x,y)$ coordinates of the flare location on the solar disk. In this paper, we train a neural network on a dataset of real CFL observations to estimate the coordinates of solar flare locations. Further, we apply a post-training quantization technique specifically tailored to the adopted model architecture. This technique allows all computations to be in integer arithmetic at inference time, making the model compatible with the STIX computational requirements. We show that our model outperforms the currently adopted algorithm for estimating the flare locations from CFL data regarding prediction accuracy while requiring fewer parameters. We finally discuss possible future applications of the proposed model on-board STIX., Comment: Conference paper accepted for an poster presentation to the ESA SPAICE CONFERENCE 17 19 September 2024

A new methodology for the improvement of the performance of inexpensive static passive electromagnetic skins (SP-EMSs) is presented. The proposed approach leverages on the non-uniqueness of the inverse source problem associated to the SP-EMS design by decomposing the induced surface current into pre-image (PI) and null-space (NS) components. Successively, the unknown EMS layout and NS expansion coefficients are determined by means of an alternate minimization of a suitable cost function. This latter quantifies the mismatch between the ideal surface current, which radiates the user-defined target field, and that actually induced on the EMS layout. Results from a representative set of numerical experiments, concerned with the design of EMSs reflecting pencil-beam as well as contoured target patterns, are reported to assess the feasibility and the effectiveness of the proposed method in improving the performance of inexpensive EMS realizations. The measurements on an EMS prototype, featuring a conductive ink pattern printed on a standard paper substrate, are also shown to prove the reliability of the synthesis process.

A new methodology for the synthesis of wide-angle scanning arrays is proposed. It is based on the formulation of the array design problem as a multi-objective one where, for each scan angle, both the radiated power density in the scan direction and the total reflected power are accounted for. A set of numerical results from full-wave simulated examples - dealing with different radiators, arrangements, frequencies, and number of elements - is reported to show the features of the proposed approach as well as to assess its potentialities.

Understanding and accurately predicting hydrogen diffusion in materials is challenging due to the complex interactions between hydrogen defects and the crystal lattice. These interactions span large length and time scales, making them difficult to address with standard ab initio techniques. This work addresses this challenge by employing accelerated machine learning (ML) molecular dynamics simulations through active learning. We conduct a comparative study of different ML-based interatomic potential schemes, including VASP, MACE, and CHGNet, utilizing various training strategies such as on-the-fly learning, pre-trained universal models, and fine-tuning.By considering different temperatures and concentration regimes, we obtain hydrogen diffusion coefficients and activation energy values which align remarkably well with experimental results, underlining the efficacy and accuracy of ML-assisted methodologies in the context of diffusive dynamics. Particularly, our procedure significantly reduces the computational effort associated with traditional transition state calculations or ad-hoc designed interatomic potentials. The results highlight the limitations of pre-trained universal solutions for defective materials and how they can be improved by fine-tuning. Specifically, fine-tuning the models on a database produced during on-the-fly training of VASP ML force-field allows the retrieving of DFT-level accuracy at a fraction of the computational cost.

We describe the planned near-term and potential longer-term upgrades of the Belle II detector at the SuperKEKB electron-positron collider operating at the KEK laboratory in Tsukuba, Japan. These upgrades will allow increasingly sensitive searches for possible new physics beyond the Standard Model in flavor, tau, electroweak and dark sector physics that are both complementary to and competitive with the LHC and other experiments., Comment: Editor: F. Forti 170 pages

A novel method to reconstruct the energy of hadronic showers in the CMS High Granularity Calorimeter (HGCAL) is presented. The HGCAL is a sampling calorimeter with very fine transverse and longitudinal granularity. The active media are silicon sensors and scintillator tiles readout by SiPMs and the absorbers are a combination of lead and Cu/CuW in the electromagnetic section, and steel in the hadronic section. The shower reconstruction method is based on graph neural networks and it makes use of a dynamic reduction network architecture. It is shown that the algorithm is able to capture and mitigate the main effects that normally hinder the reconstruction of hadronic showers using classical reconstruction methods, by compensating for fluctuations in the multiplicity, energy, and spatial distributions of the shower's constituents. The performance of the algorithm is evaluated using test beam data collected in 2018 prototype of the CMS HGCAL accompanied by a section of the CALICE AHCAL prototype. The capability of the method to mitigate the impact of energy leakage from the calorimeter is also demonstrated.

UV wind line variability in OB stars appears to be universal. To quantify this variation and to estimate its effect on a mass loss rate determined from a single observation, we use the IUE archive to identify non-peculiar OB stars with well developed but unsaturated Si IV 1400 doublets and at least 10 independent observations. This resulted in 1699 spectra of 25 stars. We use SEI modelling to translate the profile variations into optical depth variations and, hence, variations in measured mass loss rates. These variations quantify the intrinsic error inherent in any mass loss rate derived from a single observation. The derived rates have an overall RMS variation of about 22%, but this differs with effective temperature, being as small at 8% for the hottest stars and up to 45% for the cooler ones. Furthermore, any single determination can differ from the mean by a factor of 2 or more. Our results also imply that mass loss rates determined from non-simultaneous observations (such as UV and ground based data) need not agree. We also use our results to examine the nature of the wind structures responsible for the variability. Our findings suggest that the optical depth variations result from optically very thick structures occulting more or less of the line of sight to the stellar disk. Further, the smaller optical depth variations in the hottest stars suggests that the responsible structures are disrupted in their more powerful winds., Comment: 23 pages, 8 figures, Accepted by ApJ

The design of low-profile linear microstrip arrays with wide-band spatial filtering capabilities is dealt with. An innovative architecture, leveraging the angular selectivity of offset stacked patch (OSP) radiators, is proposed to implement phased arrays (PAs) with inter-element spacing larger than half-wavelength that feature remarkable grating lobes (GLs) suppression properties and an enhanced gain within a non-negligible down-looking scanning angular range. The PA layout is then obtained by optimizing the optimal micro-scale geometrical descriptors of the radiating elements so that the macro-scale electromagnetic (EM) features of the arising finite-size PA fulfill the user-defined requirements. A set of numerical test cases, concerned with a variation of the array size and its polarization, is presented to assess the capabilities, the flexibility, and the potentialities of the proposed spatial filtering technique (SFT) also in comparison with competitive state-of-the-art alternatives. The performance of a printed circuit board (PCB)-manufactured prototype are experimentally assessed, as well.

Large language models manifest the ability of few-shot adaptation to a sequence of provided examples. This behavior, known as in-context learning, allows for performing nontrivial machine learning tasks during inference only. In this work, we address the question: can we leverage in-context learning to predict out-of-distribution materials properties? However, this would not be possible for structure property prediction tasks unless an effective method is found to pass atomic-level geometric features to the transformer model. To address this problem, we employ a compound model in which GPT-2 acts on the output of geometry-aware graph neural networks to adapt in-context information. To demonstrate our model's capabilities, we partition the QM9 dataset into sequences of molecules that share a common substructure and use them for in-context learning. This approach significantly improves the performance of the model on out-of-distribution examples, surpassing the one of general graph neural network models., Comment: 12 pages, 4 figures

We study the asymptotic behaviour of the Regularized Maximum Partial Likelihood Estimator (RMPLE) in the proportional limit, considering an arbitrary convex regularizer and assuming that the covariates $\mathbf{X}_i\in\mathbb{R}^{p}$ follow a multivariate Gaussian law with covariance $\mathbf{I}_p/p$ for each $i=1, \dots, n$. In order to efficiently compute the estimator under investigation, we propose a modified Approximate Message Passing (AMP) algorithm, that we name COX-AMP, and compare its performance with the Coordinate-wise Descent (CD) algorithm, which is taken as reference. By means of the Replica method, we derive a set of six Replica Symmetric (RS) equations that we show to correctly describe the average behaviour of the estimators when the sample size and the number of covariates is large and commensurate. These equations cannot be solved in practice, as the data generating process (that we are trying to estimate) is not known. However, the update equations of COX-AMP suggest the construction of a local field that can in turn be used to accurately estimate all the RS order parameters of the theory \emph{solely from the data}, \emph{without} actually solving the RS equations. We emphasize that this approach can be applied when the estimator is computed via any method and is not restricted to COX-AMP. Once the RS order parameters are estimated, we have access to the amount of signal and noise in the RMPLE, but also its generalization error, directly from the data. Although we focus on the Partial Likelihood objective, we envisage broader application of the methodology proposed here, for instance to GLMs with nuisance parameters, which include some non-proportional hazards models, e.g. Accelerated Failure Time models.

The Small Magellanic Cloud (SMC) shows a large variation in ultraviolet (UV) dust extinction curves, ranging from Milky Way-like (MW) to significantly steeper curves with no detectable 2175 A bump. This result is based on a sample of only nine sightlines. From HST/STIS and IUE spectra of OB stars, we have measured UV extinction curves along 32 SMC sightlines where eight of these curves were published previously. We find 16 sightlines with steep extinction with no detectable 2175 A bump, four sightlines with MW-like extinction with a detectable 2175 A bump, two sightlines with fairly flat UV extinction and weak/absent 2175 A bumps, and 10 sightlines with unreliable curves due to low SMC dust columns. Our expanded sample shows that the sightlines with and without the 2175 A bump are located throughout the SMC and not limited to specific regions. The average extinction curve of the 16 bumpless sightlines is very similar to the previous average based on four sightlines. We find no correlation between dust column and the strength of the 2175 A bump. We test the hypothesis that the 2175 A bump is due to the same dust grains that are responsible for the mid-infrared carbonaceous (PAH) emission features and find they are correlated, confirming recent work in the MW. Overall, the slope of the UV extinction increases as the amplitudes of the 2175 A bump and far-UV curvature decrease. Finally, the UV slope is correlated with $N(HI)/A(V)$ and the 2175 A bump and nonlinear far-UV rise amplitudes are anti-correlated with $N(HI)/A(V)$., Comment: 24 pages, 12 figures, ApJ, accepted

In cancer research, overall survival and progression free survival are often analyzed with the Cox model. To estimate accurately the parameters in the model, sufficient data and, more importantly, sufficient events need to be observed. In practice, this is often a problem. Merging data sets from different medical centers may help, but this is not always possible due to strict privacy legislation and logistic difficulties. Recently, the Bayesian Federated Inference (BFI) strategy for generalized linear models was proposed. With this strategy the statistical analyses are performed in the local centers where the data were collected (or stored) and only the inference results are combined to a single estimated model; merging data is not necessary. The BFI methodology aims to compute from the separate inference results in the local centers what would have been obtained if the analysis had been based on the merged data sets. In this paper we generalize the BFI methodology as initially developed for generalized linear models to survival models. Simulation studies and real data analyses show excellent performance; i.e., the results obtained with the BFI methodology are very similar to the results obtained by analyzing the merged data. An R package for doing the analyses is available., Comment: 22 pages, 5 figures, 2 tables

This position paper introduces Urgent Edge Computing (UEC) as a paradigm shift addressing the evolving demands of time-sensitive applications in distributed edge environments, in time-critical scenarios. With a focus on ultra-low latency, availability, resource management, decentralization, self-organization, and robust security, UEC aims to facilitate operations in critical scenarios such as disaster response, environmental monitoring, and smart city management. This paper outlines and discusses the key requirements, challenges, and enablers along with a conceptual architecture. The paper also outlines the potential applications of Urgent Edge Computing

In this work we study the stability, convergence, and pressure-robustness of discretization methods for incompressible flows with hybrid velocity and pressure. Specifically, focusing on the Stokes problem, we identify a set of assumptions that yield inf-sup stability as well as error estimates which distinguish the velocity- and pressure-related contributions to the error. We additionally identify the key properties under which the pressure-related contributions vanish in the estimate of the velocity, thus leading to pressure-robustness. Several examples of existing and new schemes that fit into the framework are provided, and extensive numerical validation of the theoretical properties is provided.

ALICE (A Large Ion Collider Experiment) studies the Quark-Gluon Plasma (QGP): a deconfined state of matter obtained in ultra-relativistic heavy-ion collisions. One of the probes for QGP study are quarkonia and open heavy flavour, of which ALICE exploits the muonic decay. A set of Resistive Plate Chambers (RPCs), placed in the forward rapidity region of the ALICE detector, is used for muon identification purposes. The correct operation of these detectors is ensured by the choice of the proper gas mixture. Currently they are operated with a mixture of C$_{2}$H$_{2}$F$_{4}$, i-C$_{4}$H$_{10}$ and SF$_{6}$ but, starting from 2017, new EU regulations have enforced a progressive phase-out of C$_{2}$H$_{2}$F$_{4}$ because of its large Global Warming Potential (GWP), making it difficult and costly to purchase. CERN asked LHC experiments to reduce greenhouse gases emissions, to which RPC operation contributes significantly. A possible candidate for C$_{2}$H$_{2}$F$_{4}$ replacement is the C$_{3}$H$_{2}$F$_{4}$ (diluted with other gases, such as CO$_{2}$), which has been extensively tested using cosmic rays. Promising gas mixtures have been devised; the next crucial steps are the detailed in-beam characterization of such mixtures as well as the study of their performance under increasing irradiation levels. This contribution will describe the methodology and results of beam tests carried out at the CERN GIF++ (equipped with a high activity $^{137}$Cs source and muon beam) with an ALICE-like RPC prototype, operated with several mixtures with varying proportions of CO$_{2}$, C$_{3}$H$_{2}$F$_{4}$, i-C$_{4}$H$_{10}$ and SF$_{6}$ . Absorbed currents, efficiencies, prompt charges, cluster sizes, time resolutions and rate capabilities will be presented, both from digitized (for detailed shape and charge analysis) and discriminated (using the same front-end electronics as employed in ALICE) signals.