Search

The defect chemistry and thermal oxidation of lanthanide (Ln) incorporated-UO2 are critical for understanding and predicting their behavior as enhanced fuels, mixed oxide (MOX) fuels, spent nuclear fuels (SNF), and particles for safeguard purposes. In this study, we independently controlled the Ln type (Ce4+, Nd3+, and Gd3+) and the preparation condition (reduced and nonreduced) to investigate their correlations to the generated non-equilibrated defects correspondingly. From early to late lanthanides: Ce and U formed close-to-ideal solid solutions in Fm-3m and oxidized to (Ce, U)4O9, Nd and U mixing under the reducing condition formed solid solutions with oxygen vacancies aggregating near Nd, and the mixing of smaller Gd with U resulted in short-range subnano-domain segregations with Ia-3 region embedded in the global Fm-3m matrix. Both trivalent Ln-incorporated UO2 oxidized to a mixture of (Ln, U)4O9 and (Ln, U)3O8. From these signature defect structures resulting from both Ln type and preparation condition, we proposed kinetic model and thermodynamic hypothesis for explaining the oxidation resistance of (Ln, U)O2. Although originated from f-block oxides, the discovery of long-range disorder short-range ordering may be not uncommon in other metal oxide systems, which can strongly influence their functionalities and properties.

This paper investigates the role of communication in improving coordination within robot swarms, focusing on a paradigm where learning and execution occur simultaneously in a decentralized manner. We highlight the role communication can play in addressing the credit assignment problem (individual contribution to the overall performance), and how it can be influenced by it. We propose a taxonomy of existing and future works on communication, focusing on information selection and physical abstraction as principal axes for classification: from low-level lossless compression with raw signal extraction and processing to high-level lossy compression with structured communication models. The paper reviews current research from evolutionary robotics, multi-agent (deep) reinforcement learning, language models, and biophysics models to outline the challenges and opportunities of communication in a collective of robots that continuously learn from one another through local message exchanges, illustrating a form of social learning., Comment: 17 pages, 3 Figures

We compare methods for signal classification applied to voltage traces from transition edge sensors (TES) which are photon-number resolving detectors fundamental for accessing quantum advantages in information processing, communication and metrology. We quantify the impact of numerical analysis on the distinction of such signals. Furthermore, we explore dimensionality reduction techniques to create interpretable and precise photon number embeddings. We demonstrate that the preservation of local data structures of some nonlinear methods is an accurate way to achieve unsupervised classification of TES traces. We do so by considering a confidence metric that quantifies the overlap of the photon number clusters inside a latent space. Furthermore, we demonstrate that for our dataset previous methods such as the signal's area and principal component analysis can resolve up to 16 photons with confidence above $90\%$ while nonlinear techniques can resolve up to 21 with the same confidence threshold. Also, we showcase implementations of neural networks to leverage information within local structures, aiming to increase confidence in assigning photon numbers. Finally, we demonstrate the advantage of some nonlinear methods to detect and remove outlier signals., Comment: Code is publicly available at https://github.com/polyquantique/Photon-Number-Classification. Data is publicly available at https://doi.org/10.5281/zenodo.14042152 and https://doi.org/10.5281/zenodo.14101974

The $t_{2g}$ band-structure of SrTiO$_3$-based two-dimensional electron gasses (2DEGs), have been found to play a role in features such as the superconducting dome, high-mobility transport, and the magnitude of spin-orbit coupling. This adds to the already very diverse range of phenomena, including magnetism and extreme magnetoresistance, exhibited by this particular material platform. Tuning and/or combining these intriguing attributes could yield significant progress within quantum and spintronics technologies. Doing so demands precise control of the parameters, which requires a better understanding of the factors that affect them. Here we present effects of the $t_{2g}$ band-order inversion, stemming from the growth of spinel-structured $\gamma$-Al$_2$O$_3$ onto perovskite SrTiO$_3$. Electronic transport measurements show that with LaAlO$_3$/SrTiO$_3$ as the reference, the carrier density and electron mobility are enhanced, and the sample displays a reshaping of the superconducting dome. Additionally, unpaired spins are evidenced by increasing Anomalous Hall Effect with decreasing temperature, entering the same temperature range as the superconducting transition, and a Kondo-like upturn in the sheet resistance. Finally, it is argued that the high-mobility $d_{xz/yz}$-band is more likely than the $d_{xy}$-band to host the supercurrent., Comment: 10 pages, 5 figures

Ferroelectric materials have promising applications in solar-energy conversion and electro-optic devices. The internal gradient fields produced by the macroscopic polarization may improve electronic and transport semiconducting properties. However, ferroelectrics tend to display relatively large band gaps and hence low solar-energy conversion efficiencies. In this work, we explore materials with an intrinsic inverse relation between band gap and polarization in a single ferroelectric phase. Ferroelectrics with an inverse band gap versus polarization relation are characterized by low density of states contribution at the conduction states and negligible orbital hybridization at the valence states. We use high-throughput and first principles methods to find 11 ferroelectric materials with an inverse band gap versus polarization relation in the Materials Project database. Our work provides a new pathway to design small-band gap large-polarization ferroelectrics, by simultaneously tailoring the band gap and polarization of ferroelectrics with an inverse relation through an external tuning parameter.

This report summarizes what has happened in the mini-workshops entitled ''My Earth in 180 minutes'' organized during the lunch break at the SF2A 2024 conference in Marseille. The project showcased an innovative serious game designed to raise awareness of greenhouse gas (GHG) emissions in astronomical research laboratories. Participants, organized into teams, simulate strategies to reduce their carbon footprints by 50\%, focusing on key astronomical activities such as space instrumentation, data analysis, and laboratory work. The sessions highlight the challenges of achieving significant emissions reductions without disrupting core research activities, such as telescope observations. While the serious game facilitates important discussions on sustainable practices, the results point to the need for broader engagement, adaptation to different cultural contexts, and institutional support. The project highlights the importance of integrating climate action into the academic environment and suggests potential future directions for expanding its impact., Comment: SF2A 2024, Jun 2024, Marseille, France

We report ring resonators on a silicon germanium on silicon platform operating in the mid-infrared wavelength range around 3.5 - 4.6 {\mu}m with quality factors reaching up to one million. Advances in fabrication technology enable us to demonstrate such high Q-factors, which put silicon germanium at the forefront of mid-infrared integrated photonic platforms. The achievement of high Q is attested by the observation of degeneracy lifting between clockwise (CW) and counter-clockwise (CCW) resonances, as well as optical bistability due to an efficient power buildup in the rings.

We analyse high resolution slip rate data obtained during dynamic shear rupture experiments by Berman et al. (2020). We use an inverse method to extract the details of strength evolution within the cohesive zone. The overall behaviour is slip-weakening at high rupture speeds ($>0.76C_\mathrm{R}$, where $C_\mathrm{R}$ is the Rayleigh wavespeed), but non-monotonic at low rupture speeds ($<0.76C_\mathrm{R}$), with a transient increase after an initial strong weakening. The slower ruptures are associated to more weakening in the cohesive zone. The fraction of breakdown work associated to the initial weakening, immediately behind the rupture tip, matches the fracture energy estimated by independent methods, but the total breakdown work can be much larger than fracture energy. Complex stress evolution in the cohesive zone is compatible with a well-defined fracture energy that explains rupture tip propagation, but the complexity is reflected in local slip rates that will impact radiated waves.

Resonances are common in wave physics and their full and rigorous characterization is crucial to correctly tailor the response of a system in both time and frequency domains. However, they have been conventionally described by the quality factor, a real-valued number quantifying the sharpness of a single peak in the amplitude spectrum, and associated with a singularity in the complex frequency plane. But the amplitude of a physical signal does not hold all the information on the resonance and it has not been established that even the knowledge of the full distribution of singularities carries this information. Here we derive a quality function that fully characterizes resonances from the knowledge of the phase spectrum of the signal. This function is driven by the spectral derivative of the phase and does not require the amplitude of the signal. It can be obtained by numerically calculating the spectral derivative of the phase from numerical or experimental acquisitions of the phase spectrum. Alternatively, it can be retrieved from the distribution of poles and zeros in the complex frequency plane through an analytic expression, which demonstrates that singularities do not suffice to fully characterize the resonances and that both singularities and zeros must be taken into account to retrieve the quality function. This approach permits to extract all the characteristics of resonances from arbitrary spectral response functions without a priori knowledge on the physical system., Comment: Manuscript: 13 pages, 4 figures Supplementary information: 4 pages, 1 figure

Spectral clustering is a widely used method for community detection in networks. We focus on a semi-supervised community detection scenario in the Partially Labeled Stochastic Block Model (PL-SBM) with two balanced communities, where a fixed portion of labels is known. Our approach leverages random walks in which the revealed nodes in each community act as absorbing states. By analyzing the quasi-stationary distributions associated with these random walks, we construct a classifier that distinguishes the two communities by examining differences in the associated eigenvectors. We establish upper and lower bounds on the error rate for a broad class of quasi-stationary algorithms, encompassing both spectral and voting-based approaches. In particular, we prove that this class of algorithms can achieve the optimal error rate in the connected regime. We further demonstrate empirically that our quasi-stationary approach improves performance on both real-world and simulated datasets., Comment: 26 pages, 5 figures

A novel coalescence process is shown to take place in plasma fluid simulations, leading to the formation of large-scale magnetic islands that become dynamically important in the system. The parametric dependence of the process on the plasma $\beta$ and the background magnetic shear is studied, and the process is broken down at a fundamental level, allowing to clearly identify its causes and dynamics. The formation of magnetic-island-like structures at the spatial scale of the unstable modes is observed quite early in the non-linear phase of the simulation for most cases studied, as the unstable modes change their structure from interchange-like to tearing-like. This is followed by a slow coalescence process that evolves these magnetic structures towards larger and larger scales, adding to the large-scale tearing-like modes that already form by direct coupling of neighbouring unstable modes, but remain sub-dominant without the contribution from the smaller scales through coalescence. The presence of the cubic non-linearities retained in the model is essential in the dynamics of this process. The zonal fields are key actors of the overall process, acting as mediators between the competitive mechanisms from which Turbulence Driven Magnetic Islands can develop. The zonal current is found to slow down the formation of large-scale magnetic islands, acting as an inhibitor, while the zonal flow is needed to allow the system to transfer energy to the larger scales, acting as a catalyst for the island formation process.

The solar corona is highly structured by bunches of magnetic field lines forming either loops, or twisted flux ropes representing prominences/filaments, or very dynamic structures such as jets. The aim of this paper is to understand the interaction between filament channels and jets. We use high-resolution H$\alpha$ spectra obtained by the ground-based Telescope Heliographique pour lEtude du Magnetisme et des Instabilites Solaires (THEMIS) in Canary Islands, and data from Helioseismic Magnetic Imager (HMI) and Atmospheric Imaging Assembly (AIA) aboard the Solar Dynamics Observatory (SDO). In this paper we present a multi-wavelength study of the interaction of filaments and jets. They both consist of cool plasma embedded in magnetic structures. A jet is particularly well studied in all the AIA channels with a flow reaching 100-180 km s$^{-1}$. Its origin is linked to cancelling flux at the edge of the active region. Large Dopplershifts in H$\alpha$ are derived in a typical area for a short time (order of min). They correspond to flows around 140 km s$^{-1}$. In conclusion we conjecture that these flows correspond to some interchange of magnetic field lines between the filament channel and the jets leading to cool plasmoid ejections or reconnection jets perpendicularly to the jet trajectory., Comment: 13 Figures, 14 pages

We present the experimental observation of spin and density modes in a binary mixture of superfluids of light. A miscible Bose-Bose mixture with repulsive interactions is obtained by propagating, in the paraxial limit, the two circular polarization components of a laser through a non-linear hot atomic vapor. Controlling the intensity and phase for both polarizations allows us to selectively excite the fundamental modes of the mixture. Using a Bragg-like spectroscopy technique, we measure the dispersion relation and identify two distinct branches with different speeds of sound corresponding to the spin and density modes. At large photon density, we observe a crossing of these branches, which is due to an effective photon-photon interaction beyond the two-body regime and related to the saturation of the medium nonlinearity. This novel degree of freedom allows for precise control over the ratio of spin and density sound velocities and provide new insights into the control of binary mixtures collective dynamics., Comment: 10 pages, 5 figures

We discuss an extension of the standard model with a real scalar triplet, $T$, including non-renormalizable operators (NROs) up to $d=6$. If $T$ is odd under a $Z_2$ symmetry, the neutral component of $T$ is a good candidate for the dark matter (DM) of the universe. We calculate the relic density and constraints from direct and indirect detection on such a setup, concentrating on the differences with respect to the simple model for a DM $T$ with only renormalizable interactions. Bosonic operators can change the relic density of the triplet drastically, opening up new parameter space for the model. Indirect detection constraints, on the other hand, rule out an interesting part of the allowed parameter space already today and future CTA data will, very likely, provide a decisive test for this setup., Comment: 22 pages, 9 figures

We present TapeAgents, an agent framework built around a granular, structured log tape of the agent session that also plays the role of the session's resumable state. In TapeAgents we leverage tapes to facilitate all stages of the LLM Agent development lifecycle. The agent reasons by processing the tape and the LLM output to produce new thought and action steps and append them to the tape. The environment then reacts to the agent's actions by likewise appending observation steps to the tape. By virtue of this tape-centred design, TapeAgents can provide AI practitioners with holistic end-to-end support. At the development stage, tapes facilitate session persistence, agent auditing, and step-by-step debugging. Post-deployment, one can reuse tapes for evaluation, fine-tuning, and prompt-tuning; crucially, one can adapt tapes from other agents or use revised historical tapes. In this report, we explain the TapeAgents design in detail. We demonstrate possible applications of TapeAgents with several concrete examples of building monolithic agents and multi-agent teams, of optimizing agent prompts and finetuning the agent's LLM. We present tooling prototypes and report a case study where we use TapeAgents to finetune a Llama-3.1-8B form-filling assistant to perform as well as GPT-4o while being orders of magnitude cheaper. Lastly, our comparative analysis shows that TapeAgents's advantages over prior frameworks stem from our novel design of the LLM agent as a resumable, modular state machine with a structured configuration, that generates granular, structured logs and that can transform these logs into training text -- a unique combination of features absent in previous work.

Achieving global optimality in nonlinear model predictive control (NMPC) is challenging due to the non-convex nature of the underlying optimization problem. Since commonly employed local optimization techniques depend on carefully chosen initial guesses, this non-convexity often leads to suboptimal performance resulting from local optima. To overcome this limitation, we propose a novel diffusion model-based approach for near-globally optimal NMPC consisting of an offline and an online phase. The offline phase employs a local optimizer to sample from the distribution of optimal NMPC control sequences along generated system trajectories through random initial guesses. Subsequently, the generated diverse data set is used to train a diffusion model to reflect the multi-modal distribution of optima. In the online phase, the trained model is leveraged to efficiently perform a variant of random shooting optimization to obtain near-globally optimal control sequences without relying on any initial guesses or online NMPC solving. The effectiveness of our approach is illustrated in a numerical simulation indicating high performance benefits compared to direct neural network approximations of NMPC and significantly lower computation times than online solving NMPC using global optimizers.

Plug-and-Play methods for image restoration are iterative algorithms that solve a variational problem to restore an image. These algorithms are known to be flexible to changes of degradation and to perform state-of-the-art restoration. Recently, a lot of efforts have been made to explore new algorithms to solve this variational problem based on the Plug-and-Play or REgularization by Denoising (RED) frameworks, such as SNORE that is a converging stochastic gradient descent algorithm. A variant of this algorithm, named SNORE Prox, reaches state-of-the-art performances, especially for inpainting tasks. However, the convergence of SNORE Prox, that can be seen as a stochastic proximal gradient descent, has not been analyzed so far. In this paper, we prove the convergence of SNORE Prox under reasonable non-convex assumptions.

Modelling information from complex systems such as humans social interaction or words co-occurrences in our languages can help to understand how these systems are organized and function. Such systems can be modelled by networks, and network theory provides a useful set of methods to analyze them. Among these methods, graph embedding is a powerful tool to summarize the interactions and topology of a network in a vectorized feature space. When used in input of machine learning algorithms, embedding vectors help with common graph problems such as link prediction, graph matching, etc. Word embedding has the goal of representing the sense of words, extracting it from large text corpora. Despite differences in the structure of information in input of embedding algorithms, many graph embedding approaches are adapted and inspired from methods in NLP. Limits of these methods are observed in both domains. Most of these methods require long and resource greedy training. Another downside to most methods is that they are black-box, from which understanding how the information is structured is rather complex. Interpretability of a model allows understanding how the vector space is structured without the need for external information, and thus can be audited more easily. With both these limitations in mind, we propose a novel framework to efficiently embed network vertices in an interpretable vector space. Our Lower Dimension Bipartite Framework (LDBGF) leverages the bipartite projection of a network using cliques to reduce dimensionality. Along with LDBGF, we introduce two implementations of this framework that rely on communities instead of cliques: SINr-NR and SINr-MF. We show that SINr-MF can perform well on classical graphs and SINr-NR can produce high-quality graph and word embeddings that are interpretable and stable across runs.

The Hilda group is a set of asteroids whose mean motion is in a 3:2 orbital resonance with Jupiter. In this paper we use the planar Circular Restricted Three-Body Problem (CRTBP) as a dynamical model and we show that there exists a family of stable periodic orbits that are surrounded by islands of quasi-periodic motions. We have computed the frequencies of these quasi-periodic motions and we have shown how the Hilda family fits inside these islands. We have compared these results with the ones obtained using the Elliptic Restricted Three-Body Problem and they are similar, showing the suitability of the CRTBP model. It turns out that, to decide if a given asteroid belongs to the Hilda class, it is much better to look at its frequencies in the planar CRTBP rather than to use two-body orbital elements as it is commonly done today.

HI-rich absorbers seen within quasar spectra contain the bulk of neutral gas in the Universe. However, the spatial extent of these reservoirs are not extensively studied due to the pencil beam nature of quasar sightlines. Using two giant gravitational arc fields (at redshifts 1.17 and 2.06) as 2D background sources with known strong MgII absorption observed with the MUSE integral field spectrograph (IFS), we investigated whether spatially mapped MgII absorption can predict the presence of strong HI systems, and determine both the physical extent and HI mass of the two absorbing systems. We created a simple model of an ensemble of gas clouds in order to simultaneously predict the HI column density and gas covering fraction of HI-rich absorbers based on observations of the MgII rest-frame equivalent width in IFS spaxels. We first test the model on the field with HI observations already available from the literature, finding that we can recover HI column densities consistent with the previous estimates (although with large uncertainties). We then use our framework to simultaneously predict the gas covering fraction, HI column density and total HI mass ($M_{\rm{HI}}$) for both fields. We find that both of the observed strong systems have a covering fraction of $\approx70$% and are likely damped Lyman $\alpha$ systems (DLAs) with $M_{\rm{HI}}>10^9\ M_{\odot}$. Our model shows that the typical MgII metrics used in the literature to identify the presence of DLAs are sensitive to the gas covering fraction. However, these MgII metrics are still sensitive to strong HI, and can be still applied to absorbers towards gravitational arcs or other spatially extended background sources. Based on our results, we speculate that the two strong absorbers are likely representative of a neutral inner circumgalactic medium and are a significant reservoir of fuel for star formation within the host galaxies., Comment: 15 pages, 6 Figures, 3 appendices. Accepted for publication on Astronomy & Astrophysics

It is shown that in the spacetime dominated by a cosmological constant, in the far region of a Schwarzschild-de Sitter black hole, a seed magnetic field can be generated in an ambient plasma (in a state of no magnetic field) by a general-relativistic battery, which depends on the interaction of spacetime curvature with inhomogeneous plasma thermodynamics. Thus, at large distances, dark energy becomes the only gravitational source for magnetic field generation. This allows a mechanism that make dark energy manifest through its conversion to cosmic magnetic fields.

We prove that there is no relaxation gap between a quasi-dissipative nonlinear evolution equation in a Hilbert space and its linear Liouville equation reformulation on probability measures. In other words, strong and generalized solutions of such equations are unique in the class of measure-valued solutions. As a major consequence, non-convex numerical optimization over these non-linear partial differential equations can be carried out with the infinite-dimensional moment-SOS hierarchy with global convergence guarantees. This covers in particular all reaction-diffusion equations with polynomial nonlinearity.

Global visual geolocation predicts where an image was captured on Earth. Since images vary in how precisely they can be localized, this task inherently involves a significant degree of ambiguity. However, existing approaches are deterministic and overlook this aspect. In this paper, we aim to close the gap between traditional geolocalization and modern generative methods. We propose the first generative geolocation approach based on diffusion and Riemannian flow matching, where the denoising process operates directly on the Earth's surface. Our model achieves state-of-the-art performance on three visual geolocation benchmarks: OpenStreetView-5M, YFCC-100M, and iNat21. In addition, we introduce the task of probabilistic visual geolocation, where the model predicts a probability distribution over all possible locations instead of a single point. We introduce new metrics and baselines for this task, demonstrating the advantages of our diffusion-based approach. Codes and models will be made available., Comment: Project page: https://nicolas-dufour.github.io/plonk

We establish the existence of equilibrium states for geometric potentials in a family of piecewise weakly convex interval maps. This family includes systems with indifferent fixed points and non-Markov partitions. Under specific conditions, we also prove the uniqueness of equilibrium states.

We articulate fundamental mismatches between technical methods for machine unlearning in Generative AI, and documented aspirations for broader impact that these methods could have for law and policy. These aspirations are both numerous and varied, motivated by issues that pertain to privacy, copyright, safety, and more. For example, unlearning is often invoked as a solution for removing the effects of targeted information from a generative-AI model's parameters, e.g., a particular individual's personal data or in-copyright expression of Spiderman that was included in the model's training data. Unlearning is also proposed as a way to prevent a model from generating targeted types of information in its outputs, e.g., generations that closely resemble a particular individual's data or reflect the concept of "Spiderman." Both of these goals--the targeted removal of information from a model and the targeted suppression of information from a model's outputs--present various technical and substantive challenges. We provide a framework for thinking rigorously about these challenges, which enables us to be clear about why unlearning is not a general-purpose solution for circumscribing generative-AI model behavior in service of broader positive impact. We aim for conceptual clarity and to encourage more thoughtful communication among machine learning (ML), law, and policy experts who seek to develop and apply technical methods for compliance with policy objectives., Comment: Presented at the 2nd Workshop on Generative AI and Law at ICML (July 2024)

Object detection in art is a valuable tool for the digital humanities, as it allows for faster identification of objects in artistic and historical images compared to humans. However, annotating such images poses significant challenges due to the need for specialized domain expertise. We present NADA (no annotations for detection in art), a pipeline that leverages diffusion models' art-related knowledge for object detection in paintings without the need for full bounding box supervision. Our method, which supports both weakly-supervised and zero-shot scenarios and does not require any fine-tuning of its pretrained components, consists of a class proposer based on large vision-language models and a class-conditioned detector based on Stable Diffusion. NADA is evaluated on two artwork datasets, ArtDL 2.0 and IconArt, outperforming prior work in weakly-supervised detection, while being the first work for zero-shot object detection in art. Code is available at https://github.com/patrick-john-ramos/nada, Comment: 8 pages, 6 figures, to be published in WACV 2025

While the subclass of interacting supernovae with narrow hydrogen emission lines (SNe IIn) consists of some of the longest-lasting and brightest SNe ever discovered, their progenitors are still not well understood. Investigating SNe IIn as they emit across the electromagnetic spectrum is the most robust way to understand the progenitor evolution before the explosion. This work presents X-Ray, optical, infrared, and radio observations of the strongly interacting Type IIn SN 2020ywx covering a period $>1200$ days after discovery. Through multiwavelength modeling, we find that the progenitor of 2020ywx was losing mass at $\sim10^{-2}$--$10^{-3} \mathrm{\,M_{\odot}\,yr^{-1}}$ for at least 100 yr pre-explosion using the circumstellar medium (CSM) speed of $120$ km/s measured from our optical and NIR spectra. Despite the similar magnitude of mass-loss measured in different wavelength ranges, we find discrepancies between the X-ray and optical/radio-derived mass-loss evolution, which suggest asymmetries in the CSM. Furthermore, we find evidence for dust formation due to the combination of a growing blueshift in optical emission lines and near-infrared continuum emission which we fit with blackbodies at $\sim$ 1000 K. Based on the observed elevated mass loss over more than 100 years and the configuration of the CSM inferred from the multiwavelength observations, we invoke binary interaction as the most plausible mechanism to explain the overall mass-loss evolution. SN 2020ywx is thus a case that may support the growing observational consensus that SNe IIn mass loss is explained by binary interaction., Comment: Submitted to ApJ, 31 pages, 19 figures

The Sagittarius dwarf spheroidal galaxy (Sgr dSph) is a satellite orbiting the Milky Way that has experienced multiple stripping events due to tidal interactions with our Galaxy. Its accretion history has led to a distinct stellar overdensity, which is the remnant of the core of the progenitor. We present a complete chemical analysis of 111 giant stars in the core of Sgr dSph to investigate the chemical evolution and enrichment history of this satellite. Employing the metallicity-sensitive Ca H&K photometry from the Pristine Inner Galaxy Survey, we selected stars spanning a wide metallicity range and obtained high-resolution spectra with the ESO FLAMES/GIRAFFE multi-object spectrograph. For the stellar sample covering $-2.13 < \rm{[Fe/H] < -0.35}$, we derived abundances for up to 14 chemical elements with average uncertainties of $\sim 0.09$ dex and a set of stellar ages which allowed us to build an age-metallicity relation (AMR) for the entire sample. With the most comprehensive set of chemical species measured for the core of Sgr, we studied several [X/Fe] ratios. Most trends align closely with Galactic chemical trends, but notable differences emerge in the heavy $n$-capture elements, which offer independent insights into the star formation history of a stellar population. The deficiency in the $\alpha$-elements with respect the Milky Way suggests a slower, less efficient early star formation history, similar to other massive satellites. $S$-process element patterns indicate significant enrichment from AGB stars over time. The AMR and chemical ratios point to an extended star formation history, with a rapid early phase in the first Gyr, followed by declining activity and later star-forming episodes. These findings are consistent with Sgr hosting multiple stellar populations, from young ($\sim 4$ Gyr) to old, metal-poor stars ($\sim 10$ Gyr), Comment: Submitted to A&A. 23 pages and 13 figures. Comments welcome. The abstract has been modified for the ArXiv preview

Fifteen marble samples were subjected to semi-brittle deformation through triaxial compression experiments, reaching axial strains of 0.5%, 1.0%, 2.0%, 4.0%, or 7.5% at temperatures of 20C, 200C, or 350C, under a confining pressure of 400 MPa. Deformation twins, lattice curvature, and intragranular microfractures in the samples were quantitatively characterised using forescattered electron images and electron backscatter diffraction. Microstructural analyses revealed that twins accommodate most of the shortening during the first 2% strain, whereas lattice curvature associated with geometrically necessary dislocations predominantly develops in the later stages. Intragranular fracture intensity exhibits an almost linear correlation with strain during the first 2% strain but increases more slowly thereafter. The mechanical data indicate a strong temperature dependence of yield stress, consistent with the temperature dependence of the critical resolved shear stress for dislocation glide. The subsequent strain hardening is likely caused by progressively increasing intensity of interactions among dislocations and between dislocations and twin boundaries. Based on the microstructural data and interpreted hardening mechanisms, we propose a phenomenological model, with microstructural state variables, for semi-brittle deformation at our experimental conditions as a step towards development of a microphysical constitutive model of semi-brittle deformation.

This work is about estimating when a conditional generative model (CGM) can solve an in-context learning (ICL) problem. An in-context learning (ICL) problem comprises a CGM, a dataset, and a prediction task. The CGM could be a multi-modal foundation model; the dataset, a collection of patient histories, test results, and recorded diagnoses; and the prediction task to communicate a diagnosis to a new patient. A Bayesian interpretation of ICL assumes that the CGM computes a posterior predictive distribution over an unknown Bayesian model defining a joint distribution over latent explanations and observable data. From this perspective, Bayesian model criticism is a reasonable approach to assess the suitability of a given CGM for an ICL problem. However, such approaches -- like posterior predictive checks (PPCs) -- often assume that we can sample from the likelihood and posterior defined by the Bayesian model, which are not explicitly given for contemporary CGMs. To address this, we show when ancestral sampling from the predictive distribution of a CGM is equivalent to sampling datasets from the posterior predictive of the assumed Bayesian model. Then we develop the generative predictive $p$-value, which enables PPCs and their cousins for contemporary CGMs. The generative predictive $p$-value can then be used in a statistical decision procedure to determine when the model is appropriate for an ICL problem. Our method only requires generating queries and responses from a CGM and evaluating its response log probability. We empirically evaluate our method on synthetic tabular, imaging, and natural language ICL tasks using large language models.

This study examines the shift in the scientific community from X (formerly Twitter) to Bluesky, its impact on scientific communication, and consequently on social metrics (altmetrics). Analyzing 10,174 publications from multidisciplinary and library and information science (LIS) journals in 2024, the results reveal a notable increase in Bluesky activity for multidisciplinary journals in November 2024, likely influenced by political and platform changes, with mentions doubling or quadrupling for journals like Nature and Science. In LIS, the adoption of Bluesky is more limited and shows significant variations across journals, suggesting discipline-specific adoption patterns. However, overall engagement on Bluesky remains significantly lower than on X. While X currently dominates altmetric mentions, the observed growth on Bluesky suggests a potential shift in the future, underscoring its emerging role in academic dissemination and the challenges of adapting scholarly communication metrics across evolving platforms.

We extend the buckling and clamped-plate problems to the context of differential forms on compact Riemannian manifolds with smooth boundary. We characterize their smallest eigenvalues and prove that, in the case of Euclidean domains, their spectra on forms coincide with the spectra of the corresponding problems. We obtain various estimates involving the first eigenvalues of the mentioned problems and the ones of the Hodge Laplacian with respect to Dirichlet and absolute boundary conditions on forms. These estimates generalize previous ones in the case of functions.

The breaking of translational symmetry in the inner crust of a neutron star leads to the depletion of the neutron superfluid reservoir similarly to cold atomic condensates in optical lattices and in supersolids. The suppression of the superfluid fraction is studied in the BCS theory for superfluid velocities much smaller than Landau's critical velocity treating the crust as a body-centered cubic polycrystal. To this end, fully three-dimensional band structure calculations have been carried out. Although the formation of Cooper pairs is essential for the occurrence of superfluidity, the superfluid fraction is found to be insensitive to the pairing gap. In the intermediate region of the inner crust at the average baryon number density 0.03~fm$^{-3}$, only 8\% of the free neutrons are found to participate to the superflow. This very low superfluid fraction challenges the classical interpretation of pulsar frequency glitches., Comment: 27 pages, 12 figures

Subtracting the changing sky contribution from the near-infrared (NIR) spectra of faint astronomical objects is challenging and crucial to a wide range of science cases such as estimating the velocity dispersions of dwarf galaxies, studying the gas dynamics in faint galaxies, measuring accurate redshifts, and any spectroscopic studies of faint targets. Since the sky background varies with time and location, NIR spectral observations, especially those employing fiber spectrometers and targeting extended sources, require frequent sky-only observations for calibration. However, sky subtraction can be optimized with sufficient a priori knowledge of the sky's variability. In this work, we explore how to optimize sky subtraction by analyzing 1075 high-resolution NIR spectra from the CFHT's SPIRou on Maunakea, and we estimate the variability of 481 hydroxyl (OH) lines. These spectra were collected during two sets of three nights dedicated to obtaining sky observations every five and a half minutes. During the first set, we observed how the Moon affects the NIR, which has not been accurately measured at these wavelengths. We suggest accounting for the Moon contribution at separation distances less than 10 degrees when 1) reconstructing the sky using principal component analysis 2) observing targets at Y JHK mags fainter than ~15 and 3) attempting a sky subtraction better than 1%. We also identified 126 spectral doublets, or OH lines that split into at least two components, at SPIRou's resolution. In addition, we used Lomb-Scargle Periodograms and Gaussian process regression to estimate that most OH lines vary on similar timescales, which provides a valuable input for IR spectroscopic survey strategies. The data and code developed for this study are publicly available., Comment: 23 pages, 12 figures

The BrowserGym ecosystem addresses the growing need for efficient evaluation and benchmarking of web agents, particularly those leveraging automation and Large Language Models (LLMs) for web interaction tasks. Many existing benchmarks suffer from fragmentation and inconsistent evaluation methodologies, making it challenging to achieve reliable comparisons and reproducible results. BrowserGym aims to solve this by providing a unified, gym-like environment with well-defined observation and action spaces, facilitating standardized evaluation across diverse benchmarks. Combined with AgentLab, a complementary framework that aids in agent creation, testing, and analysis, BrowserGym offers flexibility for integrating new benchmarks while ensuring consistent evaluation and comprehensive experiment management. This standardized approach seeks to reduce the time and complexity of developing web agents, supporting more reliable comparisons and facilitating in-depth analysis of agent behaviors, and could result in more adaptable, capable agents, ultimately accelerating innovation in LLM-driven automation. As a supporting evidence, we conduct the first large-scale, multi-benchmark web agent experiment and compare the performance of 6 state-of-the-art LLMs across all benchmarks currently available in BrowserGym. Among other findings, our results highlight a large discrepancy between OpenAI and Anthropic's latests models, with Claude-3.5-Sonnet leading the way on almost all benchmarks, except on vision-related tasks where GPT-4o is superior. Despite these advancements, our results emphasize that building robust and efficient web agents remains a significant challenge, due to the inherent complexity of real-world web environments and the limitations of current models.

One key ingredient of image restoration is to define a realistic prior on clean images to complete the missing information in the observation. State-of-the-art restoration methods rely on a neural network to encode this prior. Moreover, typical image distributions are invariant to some set of transformations, such as rotations or flips. However, most deep architectures are not designed to represent an invariant image distribution. Recent works have proposed to overcome this difficulty by including equivariance properties within a Plug-and-Play paradigm. In this work, we propose a unified framework named Equivariant Regularization by Denoising (ERED) based on equivariant denoisers and stochastic optimization. We analyze the convergence of this algorithm and discuss its practical benefit.

Access to large-scale high-quality healthcare databases is key to accelerate medical research and make insightful discoveries about diseases. However, access to such data is often limited by patient privacy concerns, data sharing restrictions and high costs. To overcome these limitations, synthetic patient data has emerged as an alternative. However, synthetic data generation (SDG) methods typically rely on machine learning (ML) models trained on original data, leading back to the data scarcity problem. We propose an approach to generate synthetic tabular patient data that does not require access to the original data, but only a description of the desired database. We leverage prior medical knowledge and in-context learning capabilities of large language models (LLMs) to generate realistic patient data, even in a low-resource setting. We quantitatively evaluate our approach against state-of-the-art SDG models, using fidelity, privacy, and utility metrics. Our results show that while LLMs may not match the performance of state-of-the-art models trained on the original data, they effectively generate realistic patient data with well-preserved clinical correlations. An ablation study highlights key elements of our prompt contributing to high-quality synthetic patient data generation. This approach, which is easy to use and does not require original data or advanced ML skills, is particularly valuable for quickly generating custom-designed patient data, supporting project implementation and providing educational resources., Comment: 12 pages, 2 figures, 3 tables

In this work, we introduce a novel approach for predicting thermodynamic properties of binary mixtures, which we call the similarity-based method (SBM). The method is based on quantifying the pairwise similarity of components, which we achieve by comparing quantum-chemical descriptors of the components, namely $\sigma$-profiles. The basic idea behind the approach is that mixtures with similar pairs of components will have similar thermodynamic properties. The SBM is trained on a matrix that contains some data for a given property for different binary mixtures; the missing entries are then predicted by the SBM. As an example, we consider the prediction of isothermal activity coefficients at infinite dilution ($\gamma^\infty_{ij}$) and show that the SBM outperforms the well-established physical methods modified UNIFAC (Dortmund) and COSMO-SAC-dsp. In this case, the matrix is only sparsely occupied, and it is shown that the SBM works also if only a limited number of data for similar mixtures is available. The SBM idea can be transferred to any mixture property and is a powerful tool for generating essential data for many applications.

In the last few decades, ultrafast demagnetization elicited by ultrashort laser pulses has been the subject of a large body of work that aims to better understand and control this phenomenon. Although specific magnetic materials' properties play a key role in defining ultrafast demagnetization dynamics, features of the driving laser pulse such as its duration and photon energy might also contribute. Here, we report ultrafast demagnetization of a cobalt/platinum multilayer in a broad spectral range spanning from the near-infrared (near-IR) to the mid-infrared (mid-IR), with wavelengths between 0.8 and 8.7 $\mu$m. The ultrafast dynamics of the macroscopic magnetization is tracked via the time-resolved magneto-optical Kerr effect. We show that the ultrafast demagnetization of the sample can be efficiently induced over that entire excitation spectrum with minimal dependence on the excitation wavelength. Instead, we confirm that the temporal profile of the pump excitation pulse is an important factor influencing ultrafast demagnetization dynamics.

Tongue contour extraction from real-time magnetic resonance images is a nontrivial task due to the presence of artifacts manifesting in form of blurring or ghostly contours. In this work, we present results of automatic tongue delineation achieved by means of U-Net auto-encoder convolutional neural network. We present both intra- and inter-subject validation. We used real-time magnetic resonance images and manually annotated 1-pixel wide contours as inputs. Predicted probability maps were post-processed in order to obtain 1-pixel wide tongue contours. The results are very good and slightly outperform published results on automatic tongue segmentation.

Precision metrology underpins scientific and technological advancements. Quantum metrology offers a pathway to surpass classical sensing limits by leveraging quantum states and measurement strategies. However, measuring multiple incompatible observables suffers from quantum backaction, where measurement of one observable pollutes a subsequent measurement of the other. This is a manifestation of Heisenberg's uncertainty principle for two non-commuting observables, such as position and momentum. Here, we demonstrate measurements of small changes in position and momentum where the uncertainties are simultaneously reduced below the standard quantum limit (SQL). We measure $\textit{modular observables}$ using tailored, highly non-classical states that ideally evade measurement backactions. The states are deterministically prepared in the single mode of the mechanical motion of a trapped ion using an optimal quantum control protocol. Our experiment uses grid states to measure small changes in position and momentum and shows a metrological gain of up to 5.1(5)~dB over the simultaneous SQL. Using an adaptive-phase estimation algorithm with Bayesian inference, we estimate these displacements with a combined variance of 2.6(1.1)~dB below the SQL. Furthermore, we examine simultaneously estimating $\textit{number}$ and $\textit{phase}$, which are the polar counterparts of position and momentum. This is performed by preparing a novel quantum resource -- number-phase states -- and we demonstrate a metrological gain over their SQL. The combination of quantum control and multi-parameter quantum metrology marks a significant step towards unprecedented precision with applications ranging from fundamental physics to advanced quantum technologies., Comment: Main text: 8 pages, 4 figures; Supplemental material: 37 pages, 16 figures

The circumgalactic medium (CGM) is a key component needed to understand the physical processes governing the flows of gas around galaxies. Quantifying its evolution and its dependence on galaxy properties is particularly important for our understanding of accretion and feedback mechanisms. We select a volume-selected sample of 66 {\it isolated} star-forming galaxies (SFGs) at $0.4< z <1.5$ with $\log(M_\star/M_{\odot})> 9$ from the MusE GAs FLOw and Wind (MEGAFLOW) survey. Using MgII 2796,2803 absorptions in background quasars, we measure the covering fraction $f_c$ and quantify how the cool gas profile depends on galaxy properties (such as star-formation rate (SFR), stellar mass ($M_\star$) or azimuthal angle relative to the line of sight) and how these dependencies evolve with redshift. The MgII covering fraction of isolated galaxies is a strong function of impact parameter, and is steeper than previously reported. The impact parameter $b_{50}$ at which $f_c = $50\% is $b_{50}=50\pm7$kpc ($65\pm7$ kpc) for $W_r^{2796}>$0.5 \AA ($W_r^{2796}>0.1$ \AA), respectively. It is weakly correlated with SFR ($\propto$ SFR$^{0.08\pm0.09}$) and decreases with cosmic time ($\propto (1+z)^{0.8 \pm 0.7}$), contrary to the expectation of increasingly larger halos with time. The covering fraction is also higher along the minor axis than along the major axis at the $\approx 2 \sigma$ level. The CGM traced by \MgII{} is similar across the isolated galaxy population. Indeed, among the isolated galaxies with an impact parameter below 55 kpc, all have associated MgII absorption with $W_r^{2796}>$0.3\AA, resulting in a steep covering fraction $f_c(b)$., Comment: 23 pages, 7 figures, accepted for publication in A&A

Multimodal AI has the potential to significantly enhance document-understanding tasks, such as processing receipts, understanding workflows, extracting data from documents, and summarizing reports. Code generation tasks that require long-structured outputs can also be enhanced by multimodality. Despite this, their use in commercial applications is often limited due to limited access to training data and restrictive licensing, which hinders open access. To address these limitations, we introduce BigDocs-7.5M, a high-quality, open-access dataset comprising 7.5 million multimodal documents across 30 tasks. We use an efficient data curation process to ensure our data is high-quality and license-permissive. Our process emphasizes accountability, responsibility, and transparency through filtering rules, traceable metadata, and careful content analysis. Additionally, we introduce BigDocs-Bench, a benchmark suite with 10 novel tasks where we create datasets that reflect real-world use cases involving reasoning over Graphical User Interfaces (GUI) and code generation from images. Our experiments show that training with BigDocs-Bench improves average performance up to 25.8% over closed-source GPT-4o in document reasoning and structured output tasks such as Screenshot2HTML or Image2Latex generation. Finally, human evaluations showed a preference for outputs from models trained on BigDocs over GPT-4o. This suggests that BigDocs can help both academics and the open-source community utilize and improve AI tools to enhance multimodal capabilities and document reasoning. The project is hosted at https://bigdocs.github.io ., Comment: The project is hosted at https://bigdocs.github.io

The parameter space of freeze-in dark matter (DM) with mass $m_\chi$ through light dark photon (``minimal freeze-in DM'') is currently being probed by direct detection experiments through electron and nuclear recoil. Exploring the DM production in the mass range $10^{-2}~{\rm MeV} < m_\chi < 10^3$ TeV, we quantify the impact of quantum statistics and the reheating dynamics (beyond the instantaneous reheating approximation) on the DM production in the early universe, in particular, the dependence on the cosmic equation of state and the scaling of the temperature of the Standard Model bath during reheating. Special cases corresponding to matter-domination and kination are carefully studied. To fit the entire observed DM relic abundance, low-temperature reheating scenarios require an increase in the coupling between dark and visible sectors which, in turn, enhances the regions of the parameter space that are already tested and will be probed by next-generation direct detection experiments for diverse reheating scenarios., Comment: 10+8 pages, 4 figures

The coldest metal-poor population made of T and Y dwarfs are archaeological tracers of our Galaxy because they are very old and have kept the pristine material. The optical properties of these objects are important to characterise their atmospheric properties. We aim at characterising further the optical properties of ultracool metal-poor population with deep far-red optical images and parallax determinations. We solve trigonometric parallaxes of five metal-poor T dwarf candidates using 2-year monitoring with Calar-Alto 3.5-m telescope. We obtain $z'$-band photometry for the other 12 metal-poor T dwarf candidates using the 10.4-m GTC, the 8.2-m VLT, and the DES, increasing the sample of T subdwarfs with optical photometry from 12 to 24. We report a 3-$\sigma$ limit for the Accident in five optical bands using the 10.4-m GTC. We confirm four T subdwarfs and the Accident as a Y subdwarf, and propose two more Y subdwarf candidates. We emphasise that the $z_{PS1}-W1$ colour combining with the $W1-W2$ colour could break the metallicity-temperature degeneracy for T and possibly for Y dwarfs. The $z_{PS1}-W1$ colour shifts redward when metallicity decreases for a certain temperature, which is not predicted by state-of-the-art ultracool models. The Accident has the reddest $z_{PS1}-W1$ colour among our sample. The $z_{PS1}-W1$ colour will be useful to search for other examples of this cold and old population in upcoming and existing deep optical and infrared large-area surveys., Comment: 17 pages, 7 figures, 2 appendices, submitted to A&A, comments are welcome

We propose to learn legged robot locomotion skills by watching thousands of wild animal videos from the internet, such as those featured in nature documentaries. Indeed, such videos offer a rich and diverse collection of plausible motion examples, which could inform how robots should move. To achieve this, we introduce Reinforcement Learning from Wild Animal Videos (RLWAV), a method to ground these motions into physical robots. We first train a video classifier on a large-scale animal video dataset to recognize actions from RGB clips of animals in their natural habitats. We then train a multi-skill policy to control a robot in a physics simulator, using the classification score of a third-person camera capturing videos of the robot's movements as a reward for reinforcement learning. Finally, we directly transfer the learned policy to a real quadruped Solo. Remarkably, despite the extreme gap in both domain and embodiment between animals in the wild and robots, our approach enables the policy to learn diverse skills such as walking, jumping, and keeping still, without relying on reference trajectories nor skill-specific rewards., Comment: Project website: https://elliotchanesane31.github.io/RLWAV/

We study the multi-height distribution of rational points of smooth, projective and split toric varieties over $\mathbf{Q}$ using the lift of the number of points to universal torsors.

Geospatial Foundation Models (GFMs) have emerged as powerful tools for extracting representations from Earth observation data, but their evaluation remains inconsistent and narrow. Existing works often evaluate on suboptimal downstream datasets and tasks, that are often too easy or too narrow, limiting the usefulness of the evaluations to assess the real-world applicability of GFMs. Additionally, there is a distinct lack of diversity in current evaluation protocols, which fail to account for the multiplicity of image resolutions, sensor types, and temporalities, which further complicates the assessment of GFM performance. In particular, most existing benchmarks are geographically biased towards North America and Europe, questioning the global applicability of GFMs. To overcome these challenges, we introduce PANGAEA, a standardized evaluation protocol that covers a diverse set of datasets, tasks, resolutions, sensor modalities, and temporalities. It establishes a robust and widely applicable benchmark for GFMs. We evaluate the most popular GFMs openly available on this benchmark and analyze their performance across several domains. In particular, we compare these models to supervised baselines (e.g. UNet and vanilla ViT), and assess their effectiveness when faced with limited labeled data. Our findings highlight the limitations of GFMs, under different scenarios, showing that they do not consistently outperform supervised models. PANGAEA is designed to be highly extensible, allowing for the seamless inclusion of new datasets, models, and tasks in future research. By releasing the evaluation code and benchmark, we aim to enable other researchers to replicate our experiments and build upon our work, fostering a more principled evaluation protocol for large pre-trained geospatial models. The code is available at https://github.com/VMarsocci/pangaea-bench.

This paper presents a dense depth estimation approach from light-field (LF) images that is able to compensate for strong rolling shutter (RS) effects. Our method estimates RS compensated views and dense RS compensated disparity maps. We present a two-stage method based on a 2D Gaussians Splatting that allows for a ``render and compare" strategy with a point cloud formulation. In the first stage, a subset of sub-aperture images is used to estimate an RS agnostic 3D shape that is related to the scene target shape ``up to a motion". In the second stage, the deformation of the 3D shape is computed by estimating an admissible camera motion. We demonstrate the effectiveness and advantages of this approach through several experiments conducted for different scenes and types of motions. Due to lack of suitable datasets for evaluation, we also present a new carefully designed synthetic dataset of RS LF images. The source code, trained models and dataset will be made publicly available at: https://github.com/ICB-Vision-AI/DenseRSLF

We present a quantum optics-based detection method for determining the position and current of an electron beam. As electrons pass through a dilute vapor of rubidium atoms, their magnetic field perturb the atomic spin's quantum state and causes polarization rotation of a laser resonant with an optical transition of the atoms. By measuring the polarization rotation angle across the laser beam, we recreate a 2D projection of the magnetic field and use it to determine the e-beam position, size and total current. We tested this method for an e-beam with currents ranging from 30 to 110 {\mu}A. Our approach is insensitive to electron kinetic energy, and we confirmed that experimentally between 10 to 20 keV. This technique offers a unique platform for non-invasive characterization of charged particle beams used in accelerators for particle and nuclear physics research.