Search

The unique experimental connection to the QCD energy-momentum tensor offered by generalised parton distributions has been strongly highlighted in the past few years with attempts to extract the pressure and shear forces distributions within the nucleon. If, in principle, this can be performed in a model independent way from experimental data, in practice, the current limited precision and kinematic coverage make such an extraction very challenging. Moreover, the limitation to a leading-order description in the strong coupling of the data has provided only an indirect and weakly sensitive access to gluon degrees of freedoms, solely through their mixing to quarks via evolution. In this paper we address this issue by pro viding a next-to-leading order formalism allowing a reanalysis of global fits with genuine gluonic degrees of freedom. In addition, we provide an estimate of the reduction in uncertainty that could stem from the extended kinematic range relevant for the future Electron Ion Collider. Finally, we stress the connection between the analysis of the dispersion relation in terms of generalised parton distributions and the deconvolution problem., Comment: 21 pages, 14 figures

We address the online unconstrained submodular maximization problem (Online USM), in a setting with stochastic bandit feedback. In this framework, a decision-maker receives noisy rewards from a nonmonotone submodular function, taking values in a known bounded interval. This paper proposes Double-Greedy - Explore-then-Commit (DG-ETC), adapting the Double-Greedy approach from the offline and online full-information settings. DG-ETC satisfies a O(d log(dT)) problemdependent upper bound for the 1/2-approximate pseudo-regret, as well as a O(dT^{2/3}log(dT)^{1/3}) problem-free one at the same time, outperforming existing approaches. To that end, we introduce a notion of hardness for submodular functions, characterizing how difficult it is to maximize them with this type of strategy.

Inspired by models for synchronous spiking activity in neuroscience, we consider a scaling-limit framework for sequences of strong Markov processes. Within this framework, we establish the convergence of certain additive functionals toward L\'evy subordinators, which are of interest in synchronous input drive modeling in neuronal models. After proving an abstract theorem in full generality, we provide detailed and explicit conclusions in the case of reflected one-dimensional diffusions. Specializing even further, we provide an in-depth analysis of the limiting behavior of a sequence of integrated Wright-Fisher diffusions. In neuroscience, such diffusions serve to parametrize synchrony in doubly-stochastic models of spiking activity. Additional explicit examples involving the Feller diffusion and the Brownian motion with drift are also given.

LLM watermarks stand out as a promising way to attribute ownership of LLM-generated text. One threat to watermark credibility comes from spoofing attacks, where an unauthorized third party forges the watermark, enabling it to falsely attribute arbitrary texts to a particular LLM. While recent works have demonstrated that state-of-the-art schemes are in fact vulnerable to spoofing, they lack deeper qualitative analysis of the texts produced by spoofing methods. In this work, we for the first time reveal that there are observable differences between genuine and spoofed watermark texts. Namely, we show that regardless of their underlying approach, all current spoofing methods consistently leave observable artifacts in spoofed texts, indicative of watermark forgery. We build upon these findings to propose rigorous statistical tests that reliably reveal the presence of such artifacts, effectively discovering that a watermark was spoofed. Our experimental evaluation shows high test power across all current spoofing methods, providing insights into their fundamental limitations, and suggesting a way to mitigate this threat.

Decentralized learning (DL) is an emerging approach that enables nodes to collaboratively train a machine learning model without sharing raw data. In many application domains, such as healthcare, this approach faces challenges due to the high level of heterogeneity in the training data's feature space. Such feature heterogeneity lowers model utility and negatively impacts fairness, particularly for nodes with under-represented training data. In this paper, we introduce \textsc{Facade}, a clustering-based DL algorithm specifically designed for fair model training when the training data exhibits several distinct features. The challenge of \textsc{Facade} is to assign nodes to clusters, one for each feature, based on the similarity in the features of their local data, without requiring individual nodes to know apriori which cluster they belong to. \textsc{Facade} (1) dynamically assigns nodes to their appropriate clusters over time, and (2) enables nodes to collaboratively train a specialized model for each cluster in a fully decentralized manner. We theoretically prove the convergence of \textsc{Facade}, implement our algorithm, and compare it against three state-of-the-art baselines. Our experimental results on three datasets demonstrate the superiority of our approach in terms of model accuracy and fairness compared to all three competitors. Compared to the best-performing baseline, \textsc{Facade} on the CIFAR-10 dataset also reduces communication costs by 32.3\% to reach a target accuracy when cluster sizes are imbalanced.

Perpendicular magnetic anisotropy (PMA) and Dzyaloshinskii-Moriya interactions are key interactions in modern spintronics. These interactions are thought to be dominated by the oxidation of the Co|Al interface in the archetypal Platinum-Cobalt-Aluminum oxide system. Here, we observe a double sign change in the anisotropy and about threefold variation in interfacial chiral interaction, influenced not only by the oxidation, but also by the metallic Al thickness. Contrary to previous assumptions about negligible spin-orbit effects at light metal interfaces, we not only observe strong PMA with fully oxidized Al, decreasing and turning negative (in-plane) with less oxygen at the Co|Al interface, we also observe that the magnetic anisotropy reverts to positive (out-of-plane) values at fully metallic Co|Al interface. These findings suggest modification in Co d band via Co|Al orbital hybridization, an effect supported by X-ray absorption spectroscopy and ab initio theory calculations, highlighting the key impact of strain on interfacial mechanisms at fully metallic Co|Al interface., Comment: 6 pages, 5 figures

For a reductive group over a nonarchimedean local field, we define the stack of spherical Langlands parameters, using the inertia-invariants of the Langlands dual group. This generalizes the stack of unramified Langlands parameters in case the group is unramified. We then use this stack to deduce the Eichler--Shimura congruence relations for Hodge type Shimura varieties, without restrictions on the ramification., Comment: 14 pages, comments welcome!

Background: Medication review is a structured interview of the patient, performed by the pharmacist and aimed at optimizing drug treatments. In practice, medication review is a long and cognitively-demanding task that requires specific knowledge. Clinical practice guidelines have been proposed, but their application is tedious. Methods: We designed ABiMed, a clinical decision support system for medication reviews, based on the implementation of the STOPP/START v2 guidelines and on the visual presentation of aggregated drug knowledge using tables, graphs and flower glyphs. We evaluated ABiMed with 39 community pharmacists during a randomized simulation trial, each pharmacist performing a medication review for two fictitious patients without ABiMed, and two others with ABiMed. We recorded the problems identified by the pharmacists, the interventions proposed, the response time, the perceived usability and the comments. Pharmacists' medication reviews were compared to an expert-designed gold standard. Results: With ABiMed, pharmacists found 1.6 times more relevant drug-related problems during the medication review (p=1.1e-12) and proposed better interventions (p=9.8e-9), without needing more time (p=0.56). The System Usability Scale score is 82.7, which is ranked "excellent". In their comments, pharmacists appreciated the visual aspect of ABiMed and its ability to compare the current treatment with the proposed one. A multifactor analysis showed no difference in the support offered by ABiMed according to the pharmacist's age or sex, in terms of percentage of problems identified or quality of the proposed interventions. Conclusions: The use of an intelligent and visual clinical decision support system can help pharmacists when they perform medication reviews. Our main perspective is the validation of the system in clinical conditions., Comment: 8 pages, 6 figures

Ultra-Wideband (UWB) technology re-emerges as a groundbreaking ranging technology with its precise micro-location capabilities and robustness. This paper highlights the security dimensions of UWB technology, focusing in particular on the intricacies of device fingerprinting for authentication, examined through the lens of state-of-the-art deep learning techniques. Furthermore, we explore various potential enhancements to the UWB standard that could realize a sovereign UWB data network. We argue that UWB data communication holds significant potential in healthcare and ultra-secure environments, where the use of the common unlicensed 2.4~GHz band-centric wireless technology is limited or prohibited. A sovereign UWB network could serve as an alternative, providing secure localization and short-range data communication in such environments., Comment: NetAISys Workshop at ACM Mobisys 2024

The tensor cross interpolation (TCI) algorithm is a rank-revealing algorithm for decomposing low-rank, high-dimensional tensors into tensor trains/matrix product states (MPS). TCI learns a compact MPS representation of the entire object from a tiny training data set. Once obtained, the large existing MPS toolbox provides exponentially fast algorithms for performing a large set of operations. We discuss several improvements and variants of TCI. In particular, we show that replacing the cross interpolation by the partially rank-revealing LU decomposition yields a more stable and more flexible algorithm than the original algorithm. We also present two open source libraries, xfac in Python/C++ and TensorCrossInterpolation.jl in Julia, that implement these improved algorithms, and illustrate them on several applications. These include sign-problem-free integration in large dimension, the superhigh-resolution quantics representation of functions, the solution of partial differential equations, the superfast Fourier transform, the computation of partition functions, and the construction of matrix product operators., Comment: 73 pages, 14 figures, codes at http://tensor4all.org

We present Virgil, a MIRI extremely red object (MERO) detected with the F1000W filter as part of the MIRI Deep Imaging Survey (MIDIS) observations of the Hubble Ultra Deep Field (HUDF). Virgil is a Lyman-$\alpha$ emitter (LAE) at $z_{spec} = 6.6312\pm 0.0019$ (from VLT/MUSE) with a rest-frame UV-to-optical spectral energy distribution (SED) typical of LAEs at similar redshifts. However, MIRI observations reveal an unexpected extremely red color at rest-frame near-infrared wavelengths, $\rm F444W - F1000W = 2.33 \pm 0.06$. Such steep rise in the near-infrared, completely missed without MIRI imaging, is poorly reproduced by models including only stellar populations and hints towards the presence of an Active Galactic Nucleus (AGN). Interestingly, the overall SED shape of Virgil resembles that of the recently discovered population of Little Red Dots (LRDs) but does not meet their compactness criterion: at rest-frame UV-optical wavelengths Virgil's morphology follows a 2D-S\'ersic profile with average index $n = 0.93^{+0.85}_{-0.31}$ and $r_e = 0.43$~pkpc. Only at MIRI wavelengths Virgil is unresolved due to the coarser PSF. We also estimate a bolometric luminosity $L_{\rm bol} = (8.4-11.1)\times 10^{44}\rm~erg~s^{-1}$ and a supermassive black hole mass $M_{\rm BH} = (4-7)\times 10^7\rm ~ M_\odot$ in agreement with recently reported values for LRDs. This discovery demonstrates the crucial importance of deep MIRI surveys to find AGN amongst high-$z$ galaxies that otherwise would be completely missed and raises the question of how common Virgil-like objects could be in the early Universe., Comment: 17 pages, 10 figures, 3 tables. Submitted to ApJ

Watermarking has emerged as a promising way to detect LLM-generated text. To apply a watermark an LLM provider, given a secret key, augments generations with a signal that is later detectable by any party with the same key. Recent work has proposed three main families of watermarking schemes, two of which focus on the property of preserving the LLM distribution. This is motivated by it being a tractable proxy for maintaining LLM capabilities, but also by the idea that concealing a watermark deployment makes it harder for malicious actors to hide misuse by avoiding a certain LLM or attacking its watermark. Yet, despite much discourse around detectability, no prior work has investigated if any of these scheme families are detectable in a realistic black-box setting. We tackle this for the first time, developing rigorous statistical tests to detect the presence of all three most popular watermarking scheme families using only a limited number of black-box queries. We experimentally confirm the effectiveness of our methods on a range of schemes and a diverse set of open-source models. Our findings indicate that current watermarking schemes are more detectable than previously believed, and that obscuring the fact that a watermark was deployed may not be a viable way for providers to protect against adversaries. We further apply our methods to test for watermark presence behind the most popular public APIs: GPT4, Claude 3, Gemini 1.0 Pro, finding no strong evidence of a watermark at this point in time.

Evaluating and updating the obstacle avoidance velocity for an autonomous robot in real-time ensures robustness against noise and disturbances. A passive damping controller can obtain the desired motion with a torque-controlled robot, which remains compliant and ensures a safe response to external perturbations. Here, we propose a novel approach for designing the passive control policy. Our algorithm complies with obstacle-free zones while transitioning to increased damping near obstacles to ensure collision avoidance. This approach ensures stability across diverse scenarios, effectively mitigating disturbances. Validation on a 7DoF robot arm demonstrates superior collision rejection capabilities compared to the baseline, underlining its practicality for real-world applications. Our obstacle-aware damping controller represents a substantial advancement in secure robot control within complex and uncertain environments.

Patients diagnosed with metastatic breast cancer (mBC) typically undergo several radiographic assessments during their treatment. mBC often involves multiple metastatic lesions in different organs, it is imperative to accurately track and assess these lesions to gain a comprehensive understanding of the disease's response to treatment. Computerized analysis methods that rely on lesion-level tracking have often used manual matching of corresponding lesions, a time-consuming process that is prone to errors. This paper introduces an automated lesion correspondence algorithm designed to precisely track both targets' lesions and non-targets' lesions in longitudinal data. Here we demonstrate the applicability of our algorithm on the anonymized data from two Phase III trials. The dataset contains imaging data of patients for different follow-up timepoints and the radiologist annotations for the patients enrolled in the trials. Target and non-target lesions are annotated by either one or two groups of radiologists. To facilitate accurate tracking, we have developed a registration-assisted lesion correspondence algorithm. The algorithm employs a sequential two-step pipeline: (a) Firstly, an adaptive Hungarian algorithm is used to establish correspondence among lesions within a single volumetric image series which have been annotated by multiple radiologists at a specific timepoint. (b) Secondly, after establishing correspondence and assigning unique names to the lesions, three-dimensional rigid registration is applied to various image series at the same timepoint. Registration is followed by ongoing lesion correspondence based on the adaptive Hungarian algorithm and updating lesion names for accurate tracking. Validation of our automated lesion correspondence algorithm is performed through triaxial plots based on axial, sagittal, and coronal views, confirming its efficacy in matching lesions.

We construct the geometric Satake equivalence for quasi-split reductive groups over nonarchimedean local fields, using \'etale Artin-Tate motives with $\mathbb{Z}[\frac{1}{p}]$-coefficients. We consider local fields of both equal and mixed characteristic. Along the way, we extend the work of Gaussent--Littelmann on the connection between LS galleries and MV cycles to the case of residually split reductive groups. As an application, we generalize Zhu's integral Satake isomorphism for spherical Hecke algebras to ramified groups. Moreover, for residually split groups, we define generic spherical Hecke algebras, and construct generic Satake and Bernstein isomorphisms., Comment: Comments welcome!

TorchSurv is a Python package that serves as a companion tool to perform deep survival modeling within the PyTorch environment. Unlike existing libraries that impose specific parametric forms, TorchSurv enables the use of custom PyTorch-based deep survival models. With its lightweight design, minimal input requirements, full PyTorch backend, and freedom from restrictive survival model parameterizations, TorchSurv facilitates efficient deep survival model implementation and is particularly beneficial for high-dimensional and complex input data scenarios., Comment: https://opensource.nibr.com/torchsurv/

This summary reviews contributions to the CKM 2023 workshop in Working Group 4: mixing and mixing-related $CP$ violation in B system. The theoretical and experimental progress is discussed., Comment: Presented at the 12th Workshop on the CKM Unitarity Triangle, 18-22 September 2023, Santiago de Compostela

Structural glasses display at low temperature a set of anomalies in thermodynamic observables. A prominent example is the linear-in-temperature scaling of the specific heat, at odds with the Debye cubic scaling found in crystals, due to acoustic phonons. Such an excess of specific heat in amorphous solids is thought of arising from phenomenological soft excitations dubbed tunneling two-level systems (TTLS). Their nature as well as their statistical properties remain elusive from a first-principle viewpoint. In this work we investigate the canonically quantized version of the KHGPS model, a mean-field glass model of coupled anharmonic oscillators, across its phase diagram, with an emphasis on the specific heat. The thermodynamics is solved in a semiclassical expansion. We show that in the replica-symmetric region of the model, up to the marginal glass transition line where replica symmetry gets continuously broken, a disordered version of the Debye approximation holds: the specific heat is dominated by harmonic vibrational excitations inducing a power-law scaling at the transition, ruled by random matrix theory. This mechanism generalizes a previous semiclassical argument in the literature. We then study the marginal glass phase where the semiclassical expansion becomes non-perturbative due to the emergence of instantons that overcome disordered Debye behavior. Inside the glass phase, a variational solution to the instanton approach provides the prevailing excitations as TTLS, which generate a linear specific heat. This phase thus hosts a mix of TTLS and harmonic excitations generated by interactions. We finally suggest to go beyond the variational approximation through an analogy with the spin-boson model., Comment: 40 pages, 16 figures

We construct a refinement of Gaitsgory's central functor for integral motivic sheaves, and show it preserves stratified Tate motives. Towards this end, we develop a reformulation of unipotent motivic nearby cycles, which also works over higher-dimensional bases. We moreover introduce Wakimoto motives and use them to show that our motivic central functor is t-exact. A decategorification of these functors yields a new approach to generic Hecke algebras for general parahorics., Comment: 56 pages, comments welcome

The exchange bias phenomenon, inherent in exchange-coupled ferromagnetic and antiferromagnetic systems, has intrigued researchers for decades. Van der Waals materials, with their layered structure, provide an optimal platform for probing such physical phenomena. However, achieving a facile and effective means to manipulate exchange bias in pristine van der Waals heterostructures remains challenging. In this study, we investigate the origin of exchange bias in MnPS3/Fe3GeTe2 van der Waals heterostructures. Our work demonstrates a method to modulate unidirectional exchange anisotropy, achieving an unprecedented nearly 1000% variation through simple thermal cycling. Despite the compensated interfacial spin configuration of MnPS3, magneto-transport measurements reveal a huge 170 mT exchange bias at 5 K, the largest observed in pristine van der Waals antiferromagnet-ferromagnet interfaces. This substantial magnitude of the exchange bias is linked to an anomalous weak ferromagnetic ordering in MnPS3 below 40 K. On the other hand, the tunability of exchange bias during thermal cycling is ascribed to the modified arrangement of interfacial atoms and changes in the vdW gap during field cooling. Our findings highlight a robust and easily adjustable exchange bias in van der Waals antiferromagnetic/ferromagnetic heterostructures, presenting a straightforward approach to enhance other interface related spintronic phenomena for practical applications.

The goal of this paper is to perform object detection in satellite imagery with only a few examples, thus enabling users to specify any object class with minimal annotation. To this end, we explore recent methods and ideas from open-vocabulary detection for the remote sensing domain. We develop a few-shot object detector based on a traditional two-stage architecture, where the classification block is replaced by a prototype-based classifier. A large-scale pre-trained model is used to build class-reference embeddings or prototypes, which are compared to region proposal contents for label prediction. In addition, we propose to fine-tune prototypes on available training images to boost performance and learn differences between similar classes, such as aircraft types. We perform extensive evaluations on two remote sensing datasets containing challenging and rare objects. Moreover, we study the performance of both visual and image-text features, namely DINOv2 and CLIP, including two CLIP models specifically tailored for remote sensing applications. Results indicate that visual features are largely superior to vision-language models, as the latter lack the necessary domain-specific vocabulary. Lastly, the developed detector outperforms fully supervised and few-shot methods evaluated on the SIMD and DIOR datasets, despite minimal training parameters.

Identifying flood affected areas in remote sensing data is a critical problem in earth observation to analyze flood impact and drive responses. While a number of methods have been proposed in the literature, there are two main limitations in available flood detection datasets: (1) a lack of region variability is commonly observed and/or (2) they require to distinguish permanent water bodies from flooded areas from a single image, which becomes an ill-posed setup. Consequently, we extend the globally diverse MMFlood dataset to multi-date by providing one year of Sentinel-1 observations around each flood event. To our surprise, we notice that the definition of flooded pixels in MMFlood is inconsistent when observing the entire image sequence. Hence, we re-frame the flood detection task as a temporal anomaly detection problem, where anomalous water bodies are segmented from a Sentinel-1 temporal sequence. From this definition, we provide a simple method inspired by the popular video change detector ViBe, results of which quantitatively align with the SAR image time series, providing a reasonable baseline for future works.

Large-size hybrid and pixelized GEM-Micromegas gaseous detectors (40x40 cm$^2$ active area) were developed and installed in 2014 and 2015 for the COMPASS2 physics program which started at the same time. That program involved in particular two full years of Drell-Yan studies using a high-intensity pion beam on a thick polarized target. Although the detectors were placed behind a thick absorber, they were exposed to an important flux of low energy neutrons and photons. The detectors were designed to drastically reduce the discharge rate, a major issue for non-resistive Micromegas in high hadron flux, by a factor of more than 100 compared to the former ones. A hybrid solution was chosen where a pre-amplifying GEM foil is placed 2 mm above the micromesh electrode. A pixelized readout was also added in the center of the detector, where the beam is going through, in order to track particles scattered at very low angles. The combination of the hybrid structure and the pixelized central readout allowed the detector to be operated in an environment with particle flux above 10 MHz/cm$^2$ with very good detection efficiencies and spatial resolution. The performance has remained stable since 2015 in terms of gain and resolution, showing the interest of hybrid structures associating a GEM foil to a Micromegas board to protect gaseous detectors against discharges and aging effects, Comment: 9 pages, 9 figures, proceedings to the 3rd International Conference on Detector Stability and Aging Phenomena in Gaseous Detectors, CERN, November 6-10 2023, will be published in the NIMA journal Proceedings Section

The Cosmological Advanced Survey Telescope for Optical and UV Research (CASTOR) is a planned flagship space telescope, covering the blue-optical and UV part of the spectrum. Here we introduce the CASTOR image simulator, a Python GalSim package-based script capable of generating mock CASTOR images from an input catalogue. We generate example images from the CASTOR Wide, Deep, and Ultra-Deep surveys using simulated light-cones from the Santa Cruz Semi-Analytic Model. We make predictions for the performance of these surveys by comparing galaxies that are extracted from each image using Source Extractor to the input catalogue. We find that the Wide, Deep, and Ultra-Deep surveys will be complete to ~27, 29 and 30 mag, respectively, in the UV, u, and g filters, with the UV-split and u-split filters reaching a shallower depth. With a large area of ~2200 deg$^2$, the Wide survey will detect hundreds of millions of galaxies out to z~4, mostly with $M_\ast \gtrsim 10^9 M_\odot$. The Ultra-Deep survey will probe to z~5, detecting a large fraction of $M_\ast \simeq 10^8 M_\odot$ galaxies. These powerful samples will enable precision measurements of the distribution of star formation in the cosmic web, connecting the growth of stellar mass to the assembly of dark matter halos over two thirds of the history of the Universe, and other core goals of CASTOR's legacy surveys. These image simulations and the tools developed to generate them will be a vital planning tool to estimate CASTOR's performance and iterate the telescope and survey designs prior to launch., Comment: Submitted to AJ

We report the discovery of Cerberus, an extremely red object detected with the MIRI Deep Imaging Survey (MIDIS) observations in the F1000W filter of the Hubble Ultra Deep Field. The object is detected at $S/N\sim6$, with $\mathrm{F1000W}\sim27$ mag, and undetected in the NIRCam data gathered by the JWST Advanced Deep Extragalactic Survey, JADES, fainter than the 30.0-30.5 mag $5\sigma$ detection limits in individual bands, as well as in the MIDIS F560W ultra-deep data ($\sim$29 mag, $5\sigma$). Analyzing the spectral energy distribution built with low-$S/N$ ($<5$) measurements in individual optical-to-mid-infrared filters and higher $S/N$ ($\gtrsim5$) in stacked NIRCam data, we discuss the possible nature of this red NIRCam-dark source using a battery of codes. We discard the possibility of Cerberus being a Solar System body based on the $<0.016$" proper motion in the 1-year apart JADES and MIDIS observations. A sub-stellar Galactic nature is deemed unlikely, given that the Cerberus' relatively flat NIRCam-to-NIRCam and very red NIRCam-to-MIRI flux ratios are not consistent with any brown dwarf model. The extragalactic nature of Cerberus offers 3 possibilities: (1) A $z\sim0.4$ galaxy with strong emission from polycyclic aromatic hydrocarbons; the very low inferred stellar mass, $\mathrm{M}_\star=10^{5-6}$ M$_\odot$, makes this possibility highly improbable. (2) A dusty galaxy at $z\sim4$ with an inferred stellar mass $\mathrm{M}_\star\sim10^{8}$ M$_\odot$. (3) A galaxy with observational properties similar to those of the reddest little red dots discovered around $z\sim7$, but Cerberus lying at $z\sim15$, with the rest-frame optical dominated by emission from a dusty torus or a dusty starburst., Comment: Accepted for publication in ApJL, results unchanged, manuscript shortened to publish as a letter

We address the problem of stochastic combinatorial semi-bandits, where a player selects among $P$ actions from the power set of a set containing $d$ base items. Adaptivity to the problem's structure is essential in order to obtain optimal regret upper bounds. As estimating the coefficients of a covariance matrix can be manageable in practice, leveraging them should improve the regret. We design ``optimistic'' covariance-adaptive algorithms relying on online estimations of the covariance structure, called OLSUCBC and COSV (only the variances for the latter). They both yields improved gap-free regret. Although COSV can be slightly suboptimal, it improves on computational complexity by taking inspiration from Thompson Sampling approaches. It is the first sampling-based algorithm satisfying a $\sqrt{T}$ gap-free regret (up to poly-logs). We also show that in some cases, our approach efficiently leverages the semi-bandit feedback and outperforms bandit feedback approaches, not only in exponential regimes where $P\gg d$ but also when $P\leq d$, which is not covered by existing analyses.

We reveal an interplay between temperature and radical concentration necessary to establish thermal mixing (TM) as an efficient dynamic nuclear polarization (DNP) mechanism. We conducted DNP experiments by hyperpolarizing widely used DNP samples, i.e., sodium pyruvate-1-$^{13}$C in water/glycerol mixtures at varying nitroxide radical (TEMPOL) concentrations and microwave irradiation frequencies, measuring proton and carbon-13 spin temperatures. Using a cryogen consumption-free prototype-DNP apparatus, we could probe cryogenic temperatures between 1.5 and 6.5 K, i.e., below and above the boiling point of liquid helium. We identify two mechanisms for the breakdown of TM: (i) Anderson type of quantum localization for low radical concentration, or (ii) quantum Zeno localization occurring at high temperature. This observation allowed us to reconcile the recent diverging observations regarding the relevance of TM as a DNP mechanism by proposing a unifying picture and, consequently, to find a trade-off between radical concentration and electron relaxation times, which offers a pathway to improve experimental DNP performance based on TM., Comment: Main article 15 pages, 5 figures. Appendix 6 pages, 7 figures

In recent years, large Transformer-based Pre-trained Language Models (PLM) have changed the Natural Language Processing (NLP) landscape, by pushing the performance boundaries of the state-of-the-art on a wide variety of tasks. However, this performance gain goes along with an increase in complexity, and as a result, the size of such models (up to billions of parameters) represents a constraint for their deployment on embedded devices or short-inference time tasks. To cope with this situation, compressed models emerged (e.g. DistilBERT), democratizing their usage in a growing number of applications that impact our daily lives. A crucial issue is the fairness of the predictions made by both PLMs and their distilled counterparts. In this paper, we propose an empirical exploration of this problem by formalizing two questions: (1) Can we identify the neural mechanism(s) responsible for gender bias in BERT (and by extension DistilBERT)? (2) Does distillation tend to accentuate or mitigate gender bias (e.g. is DistilBERT more prone to gender bias than its uncompressed version, BERT)? Our findings are the following: (I) one cannot identify a specific layer that produces bias; (II) every attention head uniformly encodes bias; except in the context of underrepresented classes with a high imbalance of the sensitive attribute; (III) this subset of heads is different as we re-fine tune the network; (IV) bias is more homogeneously produced by the heads in the distilled model.

The effectiveness of advertising in e-commerce largely depends on the ability of merchants to bid on and win impressions for their targeted users. The bidding procedure is highly complex due to various factors such as market competition, user behavior, and the diverse objectives of advertisers. In this paper we consider the problem at the level of user timelines instead of individual bid requests, manipulating full policies (i.e. pre-defined bidding strategies) and not bid values. In order to optimally allocate policies to users, typical multiple treatments allocation methods solve knapsack-like problems which aim at maximizing an expected value under constraints. In the industrial contexts such as online advertising, we argue that optimizing for the probability of success is a more suited objective than expected value maximization, and we introduce the SuccessProbaMax algorithm that aims at finding the policy allocation which is the most likely to outperform a fixed reference policy. Finally, we conduct comprehensive experiments both on synthetic and real-world data to evaluate its performance. The results demonstrate that our proposed algorithm outperforms conventional expected-value maximization algorithms in terms of success rate., Comment: accepted to AAAI 2024, 7 pages main text, 9 pages references and appendix, 22 figures

Radiance fields have been a major breakthrough in the field of inverse rendering, novel view synthesis and 3D modeling of complex scenes from multi-view image collections. Since their introduction, it was shown that they could be extended to other modalities such as LiDAR, radio frequencies, X-ray or ultrasound. In this paper, we show that, despite the important difference between optical and synthetic aperture radar (SAR) image formation models, it is possible to extend radiance fields to radar images thus presenting the first "radar fields". This allows us to learn surface models using only collections of radar images, similar to how regular radiance fields are learned and with the same computational complexity on average. Thanks to similarities in how both fields are defined, this work also shows a potential for hybrid methods combining both optical and SAR images.

Background: Polypharmacy, i.e. taking five drugs or more, is both a public health and an economic issue. Medication reviews are structured interviews of the patient by the community pharmacist, aiming at optimizing the drug treatment and deprescribing useless, redundant or dangerous drugs. However, they remain difficult to perform and time-consuming. Several clinical decision support systems were developed for helping clinicians to manage polypharmacy. However, most were limited to the implementation of clinical practice guidelines. In this work, our objective is to design an innovative clinical decision support system for medication reviews and polypharmacy management, named ABiMed. Methods: ABiMed associates several approaches: guidelines implementation, but the automatic extraction of patient data from the GP's electronic health record and its transfer to the pharmacist, and the visual presentation of contextualized drug knowledge using visual analytics. We performed an ergonomic assessment and qualitative evaluations involving pharmacists and GPs during focus groups and workshops. Results: We describe the proposed architecture, which allows a collaborative multi-user usage. We present the various screens of ABiMed for entering or verifying patient data, for accessing drug knowledge (posology, adverse effects, interactions), for viewing STOPP/START rules and for suggesting modification to the treatment. Qualitative evaluations showed that health professionals were highly interested by our approach, associating the automatic guidelines execution with the visual presentation of drug knowledge. Conclusions: The association of guidelines implementation with visual presentation of knowledge is a promising approach for managing polypharmacy. Future works will focus on the improvement and the evaluation of ABiMed., Comment: 10 pages, 7 figures

We investigate the dynamic phase transition in two-dimensional Ising models whose equilibrium characteristics are influenced by either anisotropic interactions or quenched defects. The presence of anisotropy reduces the dynamical critical temperature, leading to the expected result that the critical temperature approaches zero in the full-anisotropy limit. We show that a comprehensive understanding of the dynamic behavior of systems with quenched defects requires a generalized definition of the dynamic order parameter. By doing so, we demonstrate that the inclusion of quenched defects lowers the dynamic critical temperature as well, with a linear trend across the range of defect fractions considered. We also explore if and how it is possible to predict the dynamic behavior of specific magnetic systems with quenched randomness. Various geometric quantities, such as a defect potential index, the defect dipole moment, and the properties of the defect Delaunay triangulation, prove useful for this purpose., Comment: 15 pages, 11 figures

Group fairness is a central research topic in text classification, where reaching fair treatment between sensitive groups (e.g. women vs. men) remains an open challenge. This paper presents a novel method for mitigating biases in neural text classification, agnostic to the model architecture. Considering the difficulty to distinguish fair from unfair information in a text encoder, we take inspiration from adversarial training to induce Wasserstein independence between representations learned to predict our target label and the ones learned to predict some sensitive attribute. Our approach provides two significant advantages. Firstly, it does not require annotations of sensitive attributes in both testing and training data. This is more suitable for real-life scenarios compared to existing methods that require annotations of sensitive attributes at train time. Second, our approach exhibits a comparable or better fairness-accuracy trade-off compared to existing methods.

Phosphatidylserine (PS) is a negatively-charged glycerophospholipid found mainly in the plasma membrane (PM) and in the late secretory/endocytic compartments, where it regulates cellular activity and can mediate apoptosis. Export of PS from the endoplasmic reticulum, its site of synthesis, to other compartments, and its transbilayer asymmetry must therefore be precisely regulated. We review recent findings on non-vesicular transport of PS by lipid transfer proteins (LTPs) at membrane contact sites, on PS flip-flop between membrane leaflets by flippases and scramblases, and on PS nano-clustering at the PM. We also discuss emerging data on cooperation between scramblases and LTPs, how perturbation of PS distribution can lead to disease, and the specific role of PS in viral infection.

Indolent cancers are characterized by long overall survival (OS) times. Therefore, powering a clinical trial to provide definitive assessment of the effects of an experimental intervention on OS in a reasonable timeframe is generally infeasible. Instead, the primary outcome in many pivotal trials is an intermediate clinical response such as progression-free survival (PFS). In several recently reported pivotal trials of interventions for indolent cancers that yielded promising results on an intermediate outcome, however, more mature data or post-approval trials showed concerning OS trends. These problematic results have prompted a keen interest in quantitative approaches for monitoring OS that can support regulatory decision-making related to the risk of an unacceptably large detrimental effect on OS. For example, the US Food and Drug Administration, the American Association for Cancer Research, and the American Statistical Association recently organized a one-day multi-stakeholder workshop entitled 'Overall Survival in Oncology Clinical Trials'. In this paper, we propose OS monitoring guidelines tailored for the setting of indolent cancers. Our pragmatic approach is modeled, in part, on the monitoring guidelines the FDA has used in cardiovascular safety trials conducted in Type 2 Diabetes Mellitus. We illustrate proposals through application to several examples informed by actual case studies., Comment: 28 pages, 5 tables

We study the stellar population properties of 182 spectroscopically-confirmed (MUSE/VLT) Lyman-$\alpha$ emitters (LAEs) and 450 photometrically-selected Lyman-Break galaxies (LBGs) at z = 2.8 - 6.7 in the Hubble eXtreme Deep Field (XDF). Leveraging the combined power of HST and JWST NIRCam and MIRI observations, we analyse their rest-frame UV-through-near-IR spectral energy distributions (SEDs) with MIRI playing a crucial role in robustly assessing the LAE's stellar mass and ages. Our LAEs are low-mass objects (log$_{10}$(M$_\star$[M$_\odot$]) ~ 7.5), with little or no dust extinction (E(B - V) ~ 0.1) and a blue UV continuum slope ($\beta$ ~ -2.2). While 75% of our LAEs are young (< 100 Myr), the remaining 25% have significantly older stellar populations (> 100 Myr). These old LAEs are statistically more massive, less extinct and have lower specific star formation rate (sSFR) compared to young LAEs. Besides, they populate the M$_\star$ - SFR plane along the main-sequence (MS) of star-forming galaxies, while young LAEs populate the starburst region. The comparison between the LAEs properties to those of a stellar-mass matched sample of LBGs shows no statistical difference between these objects, except for the LBGs redder UV continuum slope and marginally larger E(B - V) values. Interestingly, 48% of the LBGs have ages < 10 Myr and are classified as starbursts, but lack detectable Ly$\alpha$ emission. This is likely due to HI resonant scattering and/or selective dust extinction. Overall, we find that JWST observations are crucial in determining the properties of LAEs and shedding light on the properties and similarities between LAEs and LBGs., Comment: 22 pages, 15 figures, 5 tables. Accepted for publication in APJ

We present deep James Webb Space Telescope (JWST)/MIRI F560W observations of a flux-limited, ALMA-selected sample of 28 galaxies at z=0.5-3.6 in the Hubble Ultra Deep Field (HUDF). The data from the MIRI Deep Imaging Survey (MIDIS) reveal the stellar structure of the HUDF galaxies at rest-wavelengths of >1 micron for the first time. We revise the stellar mass estimates using new JWST photometry and find good agreement with pre-JWST analysis; the few discrepancies can be explained by blending issues in the earlier lower-resolution Spitzer data. At z~2.5, the resolved rest-frame near-infrared (1.6 micron) structure of the galaxies is significantly more smooth and centrally concentrated than seen by HST at rest-frame 450 nm (F160W), with effective radii of Re(F560W)=1-5 kpc and S\'ersic indices mostly close to an exponential (disk-like) profile (n~1), up to n~5 (excluding AGN). We find an average size ratio of Re(F560W)/Re(F160W)~0.7 that decreases with stellar mass. The stellar structure of the ALMA-selected galaxies is indistinguishable from a HUDF reference sample of galaxies with comparable MIRI flux density. We supplement our analysis with custom-made, position-dependent, empirical PSF models for the F560W observations. The results imply that a smoother stellar structure is in place in massive gas-rich, star-forming galaxies at Cosmic Noon, despite a more clumpy rest-frame optical appearance, placing additional constraints on galaxy formation simulations. As a next step, matched-resolution, resolved ALMA observations will be crucial to further link the mass- and light-weighted galaxy structures to the dusty interstellar medium., Comment: 19 pages, 12 figures, 1 table, Accepted for publication in ApJ

Manually grading structural changes with the modified Stoke Ankylosing Spondylitis Spinal Score (mSASSS) on spinal X-ray imaging is costly and time-consuming due to bone shape complexity and image quality variations. In this study, we address this challenge by prototyping a 2-step auto-grading pipeline, called VertXGradeNet, to automatically predict mSASSS scores for the cervical and lumbar vertebral units (VUs) in X-ray spinal imaging. The VertXGradeNet utilizes VUs generated by our previously developed VU extraction pipeline (VertXNet) as input and predicts mSASSS based on those VUs. VertXGradeNet was evaluated on an in-house dataset of lateral cervical and lumbar X-ray images for axial spondylarthritis patients. Our results show that VertXGradeNet can predict the mSASSS score for each VU when the data is limited in quantity and imbalanced. Overall, it can achieve a balanced accuracy of 0.56 and 0.51 for 4 different mSASSS scores (i.e., a score of 0, 1, 2, 3) on two test datasets. The accuracy of the presented method shows the potential to streamline the spinal radiograph readings and therefore reduce the cost of future clinical trials.