Search

This paper proposes a novel k-medoids approximation algorithm to handle large-scale datasets with reasonable computational time and memory complexity. We develop a local-search algorithm that iteratively improves the medoid selection based on the estimation of the k-medoids objective. A single batch of size m << n provides the estimation, which reduces the required memory size and the number of pairwise dissimilarities computations to O(mn), instead of O(n^2) compared to most k-medoids baselines. We obtain theoretical results highlighting that a batch of size m = O(log(n)) is sufficient to guarantee, with strong probability, the same performance as the original local-search algorithm. Multiple experiments conducted on real datasets of various sizes and dimensions show that our algorithm provides similar performances as state-of-the-art methods such as FasterPAM and BanditPAM++ with a drastically reduced running time., Comment: Paper accepted by AAAI 2025

The International Semantic Web Research School (ISWS) is a week-long intensive program designed to immerse participants in the field. This document reports a collaborative effort performed by ten teams of students, each guided by a senior researcher as their mentor, attending ISWS 2023. Each team provided a different perspective to the topic of creative AI, substantiated by a set of research questions as the main subject of their investigation. The 2023 edition of ISWS focuses on the intersection of Semantic Web technologies and Creative AI. ISWS 2023 explored various intersections between Semantic Web technologies and creative AI. A key area of focus was the potential of LLMs as support tools for knowledge engineering. Participants also delved into the multifaceted applications of LLMs, including legal aspects of creative content production, humans in the loop, decentralised approaches to multimodal generative AI models, nanopublications and AI for personal scientific knowledge graphs, commonsense knowledge in automatic story and narrative completion, generative AI for art critique, prompt engineering, automatic music composition, commonsense prototyping and conceptual blending, and elicitation of tacit knowledge. As Large Language Models and semantic technologies continue to evolve, new exciting prospects are emerging: a future where the boundaries between creative expression and factual knowledge become increasingly permeable and porous, leading to a world of knowledge that is both informative and inspiring., Comment: Technical Report

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 3,000 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai., Comment: 26 pages, 6 figures

The advent of both multi-messenger and time-domain astrophysics over the last decade has seen a large interest in the development of small-scale, cheap, and robust gamma-ray detectors. This has been further encouraged by the availability of CubeSat platforms. Of particular interest are detectors capable of producing spectral and localization measurements of X and gamma-ray transients to allow for accurate follow-up measurements at different wavelengths. A vast number of the instruments developed for such purposes in the last years use a combination of scintillators and Silicon Photomultipliers (SiPMs) for photon detection. Here, we present the design, performance, and space qualification of a readout system capable of reading out 64 SiPM channels. This low-power and low-cost system was originally designed for the POLAR-2 mission, a large scale gamma-ray polarimeter. However, its flexible design makes it equally suitable for use on various CubeSat missions. The system was found to perform well when reading out both plastic and high Z scintillators using a total of 1.8~W. The space qualified design furthermore relies on commercial off-the-shelf components, thereby also removing most international export issues. In this paper, we will present the overall design, the performance of the electronics, its performance when reading out various scintillators and the successful space-qualification of this design., Comment: Submitted for publication in JINST

The graph coloring game is a famous two-player game (re)introduced by Bodlaender in $1991$. Given a graph $G$ and $k \in \mathbb{N}$, Alice and Bob alternately (starting with Alice) color an uncolored vertex with some color in $\{1,\cdots,k\}$ such that no two adjacent vertices receive a same color. If eventually all vertices are colored, then Alice wins and Bob wins otherwise. The game chromatic number $\chi_g(G)$ is the smallest integer $k$ such that Alice has a winning strategy with $k$ colors in $G$. It has been recently (2020) shown that, given a graph $G$ and $k\in \mathbb{N}$, deciding whether $\chi_g(G)\leq k$ is PSPACE-complete. Surprisingly, this parameter is not well understood even in ``simple" graph classes. Let $P_n$ denote the path with $n\geq 1$ vertices. For instance, in the case of Cartesian grids, it is easy to show that $\chi_g(P_m \times P_n) \leq 5$ since $\chi_g(G)\leq \Delta+1$ for any graph $G$ with maximum degree $\Delta$. However, the exact value is only known for small values of $m$, namely $\chi_g(P_1\times P_n)=3$, $\chi_g(P_2\times P_n)=4$ and $\chi_g(P_3\times P_n) =4$ for $n\geq 4$ [Raspaud, Wu, 2009]. Here, we prove that, for every $n\geq 18$, $\chi_g(P_4\times P_n) =4$., Comment: 22 pages, 22 figures

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.

The teacher workforce in science, technology, engineering, and math (STEM) has been a perpetual weak spot in public schools' teaching rosters. Prior reports show the pipeline of new STEM teachers into the profession is weak while demand for instruction in STEM fields continues to grow. This paper seeks to document whether and how the STEM teacher workforce in high-need settings has been impacted by these pressures. It analyzes successive waves of nationally representative teacher survey data to explore demographics and qualifications among the secondary STEM teacher workforce in high-need settings has fared over time. Results show the STEM teacher workforce in high-need schools is consistently less likely to be experienced, less likely to hold any degree in a STEM field, less likely to hold a master's degree, and less likely to be fully certified than STEM teachers in more advantaged settings. Yet, surprisingly, the observed qualifications gaps across high- versus low-need settings are either stable or slightly narrowing over time. Certain STEM fields--namely, physical sciences and computer science--rely on a less qualified workforce than those in math or biology, with low levels of teacher qualifications observed across both high- and low-need settings. Though even when considering field-specific alignment between teachers' background qualifications and their teaching assignments, the qualifications gap between high- and low-need settings has been slowly shrinking in three of four STEM fields analyzed here. In addition to high-need schools, small schools (based on enrollment size) rely more heavily on underqualified STEM teachers.

We present a modified version of the Bravyi-Terhal bound that applies to quantum codes defined by local parity-check constraints on a $D$-dimensional lattice quotient. Specifically, we consider a quotient $\mathbb{Z}^D/\Lambda$ of $\mathbb{Z}^D$ of cardinality $n$, where $\Lambda$ is some $D$-dimensional sublattice of $\mathbb{Z}^D$: we suppose that every vertex of this quotient indexes $m$ qubits of a stabilizer code $C$, which therefore has length $nm$. We prove that if all stabilizer generators act on qubits whose indices lie within a ball of radius $\rho$, then the minimum distance $d$ of the code satisfies $d \leq m\sqrt{\gamma_D}(\sqrt{D} + 4\rho)n^\frac{D-1}{D}$ whenever $n^{1/D} \geq 8\rho\sqrt{\gamma_D}$, where $\gamma_D$ is the $D$-dimensional Hermite constant. We apply this bound to derive an upper bound on the minimum distance of Abelian Two-Block Group Algebra (2BGA) codes whose parity-check matrices have the form $[\mathbf{A} \, \vert \, \mathbf{B}]$ with each submatrix representing an element of a group algebra over a finite abelian group.

In this work, we present a preliminary study combining two approaches in the context of solving PDEs: the classical finite element method (FEM) and more recent techniques based on neural networks. Indeed, in recent years, physics-informed neural networks (PINNs) have become particularly interesting for rapidly solving such problems, especially in high dimensions. However, their lack of accuracy is a significant drawback in this context, hence the interest in combining them with FEM, for which error estimators are already known. The complete pipeline proposed here, therefore, consists of modifying classical FEM approximation spaces by taking information from a prior, chosen here as the prediction of a neural network. On the one hand, this combination improves and certifies the prediction of neural networks to obtain a fast and accurate solution. On the other hand, error estimates are proven, showing that such strategies outperform classical ones by a factor that depends only on the quality of the prior. We validate our approach with numerical results obtained for this preliminary work on parametric problems with one- and two-dimensional geometries. They demonstrate that to achieve a fixed error target, a coarser mesh can be used with our enhanced FEM compared to the standard one, leading to reduced computation time, particularly for parametric problems.

Bent monolithic active pixel sensors are the basis for the planned fully cylindrical ultra low material budget tracking detector ITS3 of the ALICE experiment. This paper presents results from testbeam campaigns using high-energy particles to verify the performance of 50 um thick bent ALPIDE chips in terms of efficiency and spatial resolution. The sensors were bent to radii of 18, 24 and 30 mm, slightly smaller than the foreseen bending radii of the future ALICE ITS3 layers. An efficiency larger than $99.9\%$ and a spatial resolution of approximately 5 um, in line with the nominal operation of flat ALPIDE sensors, is obtained at nominal operating conditions. These values are found to be independent of the bending radius and thus constitute an additional milestone in the demonstration of the feasibility of the planned ITS3 detector. In addition, a special geometry in which the beam particles graze the chip and traverse it laterally over distances of up to 3 mm is investigated.

In this paper, we present a learning-based nonlinear model predictive controller (NMPC) using an original reinforcement learning (RL) method to learn the optimal weights of the NMPC scheme. The controller is used as the current action-value function of a deep Expected Sarsa where the subsequent action-value function, usually obtained with a secondary NMPC, is approximated with a neural network (NN). With respect to existing methods, we add to the NN's input the current value of the NMPC's learned parameters so that the network is able to approximate the action-value function and stabilize the learning performance. Additionally, with the use of the NN, the real-time computational burden is approximately halved without affecting the closed-loop performance. Furthermore, we combine gradient temporal difference methods with parametrized NMPC as function approximator of the Expected Sarsa RL method to overcome the potential parameters divergence and instability issues when nonlinearities are present in the function approximation. The simulation result shows that the proposed approach converges to a locally optimal solution without instability problems.

The real type of a finite family of univariate polynomials characterizes the combined sign behavior of the polynomials over the real line. We derive an explicit formula for the number of real types subject to given degree bounds. For the special case of a single polynomial we present a closed-form expression involving Fibonacci numbers. This allows us to precisely describe the asymptotic growth of the number of real types as the degree increases, in terms of the golden ratio.

We consider positive solutions, possibly unbounded, to the semilinear equation $-\Delta u=f(u)$ on continuous epigraphs bounded from below. Under the homogeneous Dirichlet boundary condition, we prove new monotonicity results for $u$, when $f$ is a (locally or globally) Lipschitz-continuous function satisfying $ f(0) \geq 0$. As an application of our new monotonicity theorems, we prove some classification and/or non-existence results. To prove our results, we first establish some new comparison principles for semilinear problems on general unbounded open sets of $\mathbb{R}^N$, and then we use them to start and to complete a modified version of the moving plane method adapted to the geometry of the epigraph $\Omega$. As a by-product of our analysis, we also prove some new results of uniqueness and symmetry for solutions (possibly unbounded and sign-changing) to the homogeneous Dirichlet BVP for the semilinear Poisson equation in fairly general unbounded domains.

In this work we establish some rigidity results for Serrin's overdetermined problem \begin{equation*} \left\{ \begin{array}{cll} - \Delta u=f(u) & \text{in}& \Omega,\newline u > 0& \text{in} & \Omega,\newline u=0 & \text{on} & \partial \Omega,\newline \dfrac{\partial u}{\partial \eta} = \mathfrak{c} = const. & \text{on} & \partial \Omega, \end{array} \right. \end{equation*} when $\Omega \subset \mathbb{R}^N$ is an epigraph (not necessarily globally Lipschitz-continuous) and $u$ is a classical solution, possibly unbounded. In broad terms, our main results prove that $\Omega$ must be an affine half-space and $u$ must be one-dimensional, provided the epigraph is bounded from below. These results hold when $f$ is of Allen-Cahn type and $ N \geq 2$ or, alternatively, when $f$ is locally Lipschitz-continuous (with no restriction on the sign of $f(0)$) and $ N \leq 3$. These results partially answer a question raised by Berestycki, Caffarelli and Nirenberg in [1]. Finally, when $f(0) <0$, we also prove a new monotonicity result, valid in any dimension $ N \geq 2$.

We prove that every graph $G$ with minimum degree at least $k\geq 2$ contains a cycle $C$ that is dense in several respects. First, it has at least $k+1$ vertices each having at least $k$ neighbors in $C$ (so $C$ has at least $\frac{(k+1)(k-2)}{2}$ chords). The inequality is tight and provides a sharp upper bound for the chromatic number of graphs where all cycles have less than $\ell$ chords for all $\ell \geq 0$. Then, we show that some edges of $C$ can be contracted to obtain a graph with high minimum degree (we call such a minor of $C$ a \emph{cyclic minor}). We then investigate further cycles having cliques as cyclic minors, and show that minimum degree at least $O(k^2)$ guarantees a cyclic $K_k$-minor.

We point out that, although SMEFT is constructed to be model-independent, this holds only if complete sets of operators are used. SMEFT analyses that focus on subsets of operators are only semi-model-independent. They probe classes of ultraviolet models, opening a window to the underlying new physics. This model dependence originates from the need to choose a weak basis, usually taken to be the down or up basis. This naturally leads to the introduction of a third SMEFT weak basis, the generic basis, which interpolates between the above two bases. Due to its generality, we argue that the generic weak basis is the most natural for new physics. In the down and up weak bases, only a specific product of the two transformation matrices from the weak to the mass basis is observable -- the CKM matrix. However, in the generic weak basis, both transformation matrices are measurable in principle, possibly up to some phases., Comment: 9 pages

Preparing quantum states is a fundamental task in various quantum algorithms. In particular, state preparation in quantum harmonic oscillators (HOs) is crucial for the creation of qudits and the implementation of high-dimensional algorithms. In this work, we develop a methodology for preparing quantum states in HOs. The HO is coupled to an auxiliary qubit to ensure that any state can be prepared in the oscillator [J. Math. Phys. 59, 072101]. By applying a sequence of square pulses to both the qubit and the HO, we drive the system from an initial state to a target state. To determine the required pulses, we use a neural network that predicts the pulse parameters needed for state preparation. Specifically, we present results for preparing qubit and qutrit states in the HO, achieving average fidelities of 99.9% and 97.0%, respectively.

Despite the impressive performance of large language models (LLMs) across various benchmarks, their ability to address ambiguously specified problems--frequent in real-world interactions--remains underexplored. To address this gap, we introduce a formal definition of task ambiguity and frame the problem of task disambiguation through the lens of Bayesian Experimental Design. By posing clarifying questions, LLM agents can acquire additional task specifications, progressively narrowing the space of viable solutions and reducing the risk of generating unsatisfactory outputs. Yet, generating effective clarifying questions requires LLM agents to engage in a form of meta-cognitive reasoning, an ability LLMs may presently lack. Our proposed approach of active task disambiguation enables LLM agents to generate targeted questions maximizing the information gain. Effectively, this approach shifts the load from implicit to explicit reasoning about the space of viable solutions. Empirical results demonstrate that this form of question selection leads to more effective task disambiguation in comparison to approaches relying on reasoning solely within the space of questions.

The properties of the Wilson rational functions ${}_{10}\phi_9$ with three different normalizations are described. For one normalization, it satisfies an $R_{II}$ recurrence relation, whereas for the two other ones, they satisfy a generalized eigenvalue problem. The so-called Wilson rational algebra is introduced, which encodes algebraically the spectral properties of these special functions. Finally, different limits are considered, leading up to functions proportional to ${}_{4}\phi_3$. For one of these, the spectral algebra simplifies to yield the meta $q$-Racah algebra.

Isospin-1/2 charmed axial-vector $D^*\pi-D^*\eta-D^*_s\bar{K}$ scattering amplitudes are computed, along with interactions in several other $I=1/2$ $J^P$ channels. Using lattice QCD, we work at a light-quark mass corresponding to $m_\pi\approx 391$ MeV, where the lowest three-hadron threshold ($D\pi\pi$) lies high enough to enable a rigorous treatment of this system considering only two-hadron scattering channels. At this light-quark mass, an axial-vector $D_1$ bound state is observed just below $D^*\pi$ threshold, that is strongly coupled to $D^*\pi$ in a relative $S$-wave and influences a wide energy region up to the $D^*\eta$ threshold. An axial-vector $D_1^\prime$ resonance is observed in the elastic $D^*\pi$ energy-region, which is coupled more strongly to $D$-wave $D^*\pi$. A single narrow tensor state is seen in $J^P=2^+$ coupled to both $D\pi$ and $D^*\pi$. In the region where $D^*\eta$ and $D^*_s\bar{K}$ are kinematically open, the available energy levels indicate significant $S$-wave interactions. Upon searching this region for poles, several possibilities exist with large uncertainties. One additional state consistently arises, predominantly coupled to the $S$-wave $D^*\pi-D^*\eta-D^*_s\bar{K}$ amplitudes around the upper energy limit of this analysis., Comment: 52 pages, 14 figures

A graph class $\mathcal{C}$ has polynomial expansion if there is a polynomial function $f$ such that for every graph $G\in \mathcal{C}$, each of the depth-$r$ minors of $G$ has average degree at most $f(r)$. In this note, we study bounded-radius variants of some classical graph parameters such as bramble number, linkedness and well-linkedness, and we show that they are pairwise polynomially related. Furthermore, in a monotone graph class with polynomial expansion they are all uniformly bounded by a polynomial in $r$., Comment: 10 pages

This chapter is about the fundamentals of fabrication, control, and human-robot interaction of a new type of collaborative robotic manipulators, called malleable robots, which are based on adjustable architectures of varying stiffness for achieving high dexterity with lower mobility arms. Collaborative robots, or cobots, commonly integrate six or more degrees of freedom (DOF) in a serial arm in order to allow positioning in constrained spaces and adaptability across tasks. Increasing the dexterity of robotic arms has been indeed traditionally accomplished by increasing the number of degrees of freedom of the system; however, once a robotic task has been established (e.g., a pick-and-place operation), the motion of the end-effector can be normally achieved using less than 6-DOF (i.e., lower mobility). The aim of malleable robots is to close the technological gap that separates current cobots from achieving flexible, accessible manufacturing automation with a reduced number of actuators., Comment: 37 pages, 22 figures, chapter 7 of "Handbook on Soft Robotics"

Thin stellar streams, such as those resulting from the tidal disruption of globular clusters, have long been known and used as probes of the gravitational potential of our Galaxy, both its visible and dark contents. In particular, the presence of under-density regions, or gaps, along these streams is commonly interpreted as being due to the close passage of dark matter sub-halos. In this work, we investigate the perturbations induced on streams by the passage of dense stellar systems, such as globular clusters themselves, to test the possibility that they may cause the formation of gaps as well. In particular, we focus on the study of the stream of Palomar 5, a well-known globular cluster in the Galactic halo, which has particularly long tidal tails. For this purpose, we used a particle-test code to simulate Palomar 5's tidal tails when subjected to the Galaxy's gravitational field plus its whole system of globular clusters. Our study shows that the tails of Palomar 5 can be strongly perturbed by the close passage of other clusters, in particular of NGC 2808, NGC 7078, NGC 104, and that these perturbations induce the formation of gaps in the tails. These results show that globular clusters are capable of inducing gaps in streams--as other baryonic components such as giant molecular clouds and the galactic bar have been shown to do in other works. Therefore, when searching to construct the distribution function of dark matter sub halos within the Milky Way, the gap contribution from globular clusters must be included., Comment: 13 pages, 8 figures, 17 appendix pages

In the present paper, we consider the interacting partially-directed self-avoiding walk (IPDSAW) attracted by a vertical wall. The IPDSAW was introduced by Zwanzig and Lauritzen (J. Chem. Phys., 1968) as a manner of investigating the collapse transition of a homopolymer dipped in a repulsive solvent. We prove in particular that a surface transition occurs inside the collapsed phase between (i) a regime where the attractive vertical wall does not influence the geometry of the polymer and (ii) a regime where the polymer is partially attached at the wall on a length that is comparable to its horizontal extension, modifying its asymptotic Wulff shape. The latter rigorously confirms the conjecture exposed by physicists in (Physica A: Stat. Mech. \\& App., 2002). We push the analysis even further by providing sharp asymptotics of the partition function inside the collapsed phase.

A junction is a particular network given by the collection of $N\ge 1$ half lines $[0,+\infty)$ glued together at the origin. On such a junction, we consider evolutive Hamilton-Jacobi equations with $N$ coercive Hamiltonians. Furthermore,we consider a general desired junction condition at the origin, given by some monotone function $F_0:\R^N\to \R$.There is existence and uniqueness of solutions which only satisfy weakly the junction condition (at the origin, they satisfy either the desired junction condition or the PDE).We show that those solutions satisfy strongly a relaxed junction condition $\frak R F_0$ (that we can recognize as an effective junction condition). It is remarkable that this relaxed condition can be computed in three different but equivalent ways: 1) using viscosity inequalities, 2) using Godunov fluxes, 3) using Riemann problems.Our result goes beyond uniqueness theory, in the following sense: solutions to two different desired junction conditions $F_0$ and $F_1$ do coincide if $\frak R F_0=\frak R F_1$.

The widespread use of chest X-rays (CXRs), coupled with a shortage of radiologists, has driven growing interest in automated CXR analysis and AI-assisted reporting. While existing vision-language models (VLMs) show promise in specific tasks such as report generation or abnormality detection, they often lack support for interactive diagnostic capabilities. In this work we present RadVLM, a compact, multitask conversational foundation model designed for CXR interpretation. To this end, we curate a large-scale instruction dataset comprising over 1 million image-instruction pairs containing both single-turn tasks -- such as report generation, abnormality classification, and visual grounding -- and multi-turn, multi-task conversational interactions. After fine-tuning RadVLM on this instruction dataset, we evaluate it across different tasks along with re-implemented baseline VLMs. Our results show that RadVLM achieves state-of-the-art performance in conversational capabilities and visual grounding while remaining competitive in other radiology tasks. Ablation studies further highlight the benefit of joint training across multiple tasks, particularly for scenarios with limited annotated data. Together, these findings highlight the potential of RadVLM as a clinically relevant AI assistant, providing structured CXR interpretation and conversational capabilities to support more effective and accessible diagnostic workflows., Comment: 21 pages, 15 figures

This paper introduces IRIS, an immersive Robot Interaction System leveraging Extended Reality (XR), designed for robot data collection and interaction across multiple simulators, benchmarks, and real-world scenarios. While existing XR-based data collection systems provide efficient and intuitive solutions for large-scale data collection, they are often challenging to reproduce and reuse. This limitation arises because current systems are highly tailored to simulator-specific use cases and environments. IRIS is a novel, easily extendable framework that already supports multiple simulators, benchmarks, and even headsets. Furthermore, IRIS is able to include additional information from real-world sensors, such as point clouds captured through depth cameras. A unified scene specification is generated directly from simulators or real-world sensors and transmitted to XR headsets, creating identical scenes in XR. This specification allows IRIS to support any of the objects, assets, and robots provided by the simulators. In addition, IRIS introduces shared spatial anchors and a robust communication protocol that links simulations between multiple XR headsets. This feature enables multiple XR headsets to share a synchronized scene, facilitating collaborative and multi-user data collection. IRIS can be deployed on any device that supports the Unity Framework, encompassing the vast majority of commercially available headsets. In this work, IRIS was deployed and tested on the Meta Quest 3 and the HoloLens 2. IRIS showcased its versatility across a wide range of real-world and simulated scenarios, using current popular robot simulators such as MuJoCo, IsaacSim, CoppeliaSim, and Genesis. In addition, a user study evaluates IRIS on a data collection task for the LIBERO benchmark. The study shows that IRIS significantly outperforms the baseline in both objective and subjective metrics.

Machine learning (ML) is increasingly vital for smart-grid research, yet restricted access to realistic, diverse data - often due to privacy concerns - slows progress and fuels doubts within the energy sector about adopting ML-based strategies. We propose integrating Large Language Models (LLMs) in energy modeling to generate realistic, culturally sensitive, and behavior-specific data for household energy usage across diverse geographies. In this study, we employ and compare five different LLMs to systematically produce family structures, weather patterns, and daily consumption profiles for households in six distinct countries. A four-stage methodology synthesizes contextual daily data, including culturally nuanced activities, realistic weather ranges, HVAC operations, and distinct `energy signatures' that capture unique consumption footprints. Additionally, we explore an alternative strategy where external weather datasets can be directly integrated, bypassing intermediate weather modeling stages while ensuring physically consistent data inputs. The resulting dataset provides insights into how cultural, climatic, and behavioral factors converge to shape carbon emissions, offering a cost-effective avenue for scenario-based energy optimization. This approach underscores how prompt engineering, combined with knowledge distillation, can advance sustainable energy research and climate mitigation efforts. Source code is available at https://github.com/Singularity-AI-Lab/LLM-Energy-Knowledge-Distillation ., Comment: Source code is available at https://github.com/Singularity-AI-Lab/LLM-Energy-Knowledge-Distillation

Superconducting traveling-wave parametric amplifiers (TWPAs) are increasingly used in various applications, including quantum computing, quantum sensing, and dark matter detection. However, one important characteristic of these amplifiers, gain compression, has not received much attention. As a result, there is a lack of comprehensive experimental exploration of this phenomenon in the existing literature. In this study, we present an experimental investigation of gain compression in a Josephson traveling-wave parametric amplifier based on a four-wave mixing process. We have implemented a novel setup to monitor the complex transmission of both the pump and signal tones, which allows us to simultaneously track pump depletion and signal amplification as functions of signal power and frequency across the entire bandwidth of the device. Our findings indicate that, while pump depletion occurs during gain compression, it is not the only mechanism involved in the saturation of a TWPA. Power-induced phase-matching processes also take place within the device. This study provides valuable insights for optimizing TWPAs for applications that require high total input power, such as multiplexed qubit readout or broadband photon emission.

In this study, the crystal structure and magnetic properties of La$_{2-x} A'_{x}$Ni$_{7}$ compounds with magnetic rare earth elements ($A$' = Sm, Gd) have been investigated combining X-ray powder diffraction and magnetic measurements. These intergrowth compounds crystallize in a mixture of 2$H$ hexagonal (Ce$_{2}$Ni$_{7}$-type) and 3$R$ rhombohedral (Gd$_{2}$Co$_{7}$-type) polymorphic structures which are related to the stacking of [$AB_{5}$] and [$A_{2}B_{4}$] subunits along the $c$-axis.The average cell volume decreases linearly versus $A$' content, whereas the $c/a$ ratio reaches a minimum at $x$ = 1, due to geometric constraints upon $A$' for La substitution between the two different subunits. The magnetic properties strongly depend on the structure type and the $A$' content. Hexagonal La$_{2}$Ni$_{7}$ is a weak antiferromagnet (wAFM) at low field and temperature and undergoes metamagnetic transitions towards weak ferromagnetic state (wFM) under applied field. Under an applied field of 0.1 T, La$_{2-x}A'_{x}$Ni$_{7}$ intermetallic compounds display two different transition temperatures $T_{1}$ and $T_{2}$ that both increase with $x$. $T_{1}$ is associated with a wFM-wAFM transition in the 2$H$ phase for $A$'= Sm, whereas $T_{2}$ is related to the Curie temperature of both 2$H$ and 3$R$ phases. A metamagnetic behaviour is observed between $T_{1}$ and $T_{2}$ with transition field $\mu_{0}H_{Trans}$ between 2 and 3.5 T for compounds with $A$' = Sm. The La$_{2-x}Sm_{x}$Ni$_{7}$ compounds ($x$ > 0} behave as hard magnets with a large coercive field $\mu_{0}H_{C}$ at low temperature ($\mu_{0}H_{C}$ > 9 T at 5 K for $x$ = 2), whereas the La$_{2-x} Gd_{x}$Ni$_{7}$ compounds ($x$ > 0) are soft ferrimagnets with a linear increase of the saturation magnetization versus Gd content., Comment: 29 pages (20 main article, 9 in supplementary Material, 1 table, 19 figures (10 in main article, 9 in supplementary Material)

Recent advances in machine learning have opened new avenues for optimizing detector designs in high-energy physics, where the complex interplay of geometry, materials, and physics processes has traditionally posed a significant challenge. In this work, we introduce the $\textit{end-to-end}$ optimization framework AIDO that leverages a diffusion model as a surrogate for the full simulation and reconstruction chain, enabling gradient-based design exploration in both continuous and discrete parameter spaces. Although this framework is applicable to a broad range of detectors, we illustrate its power using the specific example of a sampling calorimeter, focusing on charged pions and photons as representative incident particles. Our results demonstrate that the diffusion model effectively captures critical performance metrics for calorimeter design, guiding the automatic search for layer arrangement and material composition that aligns with known calorimeter principles. The success of this proof-of-concept study provides a foundation for future applications of end-to-end optimization to more complex detector systems, offering a promising path toward systematically exploring the vast design space in next-generation experiments., Comment: 14 pages, 9 figures, submitted to MDPI particles

Context. The localized formation of planetesimals can be triggered with the help of streaming instability when the local pebble density is high. This can happen at various locations in the disk leading to the formation of local planetesimal rings. The planetesimals in these rings subsequently grow from mutual collisions and by pebble accretion. Aims. We investigate the early growth of protoplanetary embryos from a ring of planetesimals created from streaming instability to see if they reach sizes where they accrete pebbles efficiently. Methods. We simulate the early stages of planet formation for rings of planetesimals that we assume were created by streaming instability at various separations from the star and for various stellar masses using a semi-analytic model. Results. The rings in the inner disk are able to produce protoplanetary embryos in a short time whereas at large separations there is little to no growth. The growth of the largest bodies is significantly slower around lower-mass stars. Conclusions. The formation of planetary embryos from filaments during the disk lifetime is possible but strongly dependent on the separation from the star and the mass of the host star. It remains difficult to form the seeds of pebble accretion early in the outer disk \sim 50AU, especially for low-mass stars., Comment: 16 pages, 19 figures

Dielectric structures can support low-absorption optical modes, which are attractive for engineering light-matter interactions with excitonic resonances in two-dimensional (2D) materials. However, the coupling strength is often limited by the electromagnetic field being confined inside the dielectric, reducing spatial overlap with the active excitonic material. Here, we demonstrate a scheme for enhanced light-matter coupling by embedding excitonic tungsten disulfide (WS$_2$) within dielectric hexagonal boron nitride (hBN), forming a van der Waals (vdW) heterostructure that optimizes the field overlap and alignment between excitons and optical waveguide modes. To tailor diffractive coupling between free-space light and the waveguide modes in the vdW heterostructure, we fabricate Fourier surfaces in the top hBN layer using thermal scanning-probe lithography and etching, producing sinusoidal topographic landscapes with nanometer precision. We observe the formation of exciton-polaritons with a Rabi splitting indicating that the system is at the onset of strong coupling. These results demonstrate the potential of Fourier-tailored vdW heterostructures for exploring advanced optoelectronic and quantum devices.

We implement a paraxial azimuthally-radially polarized beam (ARPB), a novel class of structured light beams that can be optimal chiral (OC), leading to maximum chirality density at a given energy density. By using vectorial light shaping techniques, we successfully generated a paraxial ARPB with precise control over its features, validating theoretical predictions. Our findings demonstrate the ability to finely adjust the chirality density of the ARPB across its entire range by manipulating a single beam parameter. Although our experimental investigations are primarily focused on the transverse plane, we show that fields whose transverse components satisfy the optimal chirality condition are optimally chiral in all directions, and our results highlight the promising potential of OC structured light for applications in the sensing and manipulation of chiral particles. We show that optical chirality is more general than the concept of handedness. This work represents a significant advancement toward practical optical enantioseparation and enantiomer detection at the nanoscale.

In two recent works, Hoffman, Johnson and Junge proved the density conjecture, the hockey stick conjecture and the ball conjecture for Activated Random Walks in dimension 1, showing an equality between several different definitions of the critical density of the model. This establishes a kind of self-organized criticality, that was originally predicted for the Abelian Sandpile Model. The proof of Hoffman, Johnson and Junge uses a comparison with a percolation process, which exhibits a superadditivity property. In the present note, we revisit their argument by providing a new proof of superadditivity directly for Activated Random Walks, without relying on a percolation process. The proof relies on a simple comparison between the stabilization of two neighbouring segments and that of their union. We then explain how this superaddivity property implies the three mentioned conjectures. Yet, so far it does not seem that this approach yields as much information as does the percolation technology developed by Hoffman, Johnson and Junge, which yields an exponential concentration bound on the stationary density, whereas the superadditivity property alone only ensures an exponential bound on the lower tail., Comment: 16 pages. Preliminary version, comments most welcome!

Medical multimodal large language models (MLLMs) are becoming an instrumental part of healthcare systems, assisting medical personnel with decision making and results analysis. Models for radiology report generation are able to interpret medical imagery, thus reducing the workload of radiologists. As medical data is scarce and protected by privacy regulations, medical MLLMs represent valuable intellectual property. However, these assets are potentially vulnerable to model stealing, where attackers aim to replicate their functionality via black-box access. So far, model stealing for the medical domain has focused on classification; however, existing attacks are not effective against MLLMs. In this paper, we introduce Adversarial Domain Alignment (ADA-STEAL), the first stealing attack against medical MLLMs. ADA-STEAL relies on natural images, which are public and widely available, as opposed to their medical counterparts. We show that data augmentation with adversarial noise is sufficient to overcome the data distribution gap between natural images and the domain-specific distribution of the victim MLLM. Experiments on the IU X-RAY and MIMIC-CXR radiology datasets demonstrate that Adversarial Domain Alignment enables attackers to steal the medical MLLM without any access to medical data., Comment: Accepted at AAAI 2025

Discrete diffusion models have recently gained significant attention due to their ability to process complex discrete structures for language modeling. However, fine-tuning these models with policy gradient methods, as is commonly done in Reinforcement Learning from Human Feedback (RLHF), remains a challenging task. We propose an efficient, broadly applicable, and theoretically justified policy gradient algorithm, called Score Entropy Policy Optimization (SEPO), for fine-tuning discrete diffusion models over non-differentiable rewards. Our numerical experiments across several discrete generative tasks demonstrate the scalability and efficiency of our method. Our code is available at https://github.com/ozekri/SEPO, Comment: 23 pages, 4 figures, 5 tables

Aligning visual features with language embeddings is a key challenge in vision-language models (VLMs). The performance of such models hinges on having a good connector that maps visual features generated by a vision encoder to a shared embedding space with the LLM while preserving semantic similarity. Existing connectors, such as multilayer perceptrons (MLPs), often produce out-of-distribution or noisy inputs, leading to misalignment between the modalities. In this work, we propose a novel vision-text alignment method, AlignVLM, that maps visual features to a weighted average of LLM text embeddings. Our approach leverages the linguistic priors encoded by the LLM to ensure that visual features are mapped to regions of the space that the LLM can effectively interpret. AlignVLM is particularly effective for document understanding tasks, where scanned document images must be accurately mapped to their textual content. Our extensive experiments show that AlignVLM achieves state-of-the-art performance compared to prior alignment methods. We provide further analysis demonstrating improved vision-text feature alignment and robustness to noise.

In Distributed Manipulator Systems (DMS), decentralization is a highly desirable property as it promotes robustness and facilitates scalability by distributing computational burden and eliminating singular points of failure. However, current DMS typically utilize a centralized approach to sensing, such as single-camera computer vision systems. This centralization poses a risk to system reliability and offers a significant limiting factor to system size. In this work, we introduce a decentralized approach for sensing and in a Distributed Manipulator Systems using Neural Cellular Automata (NCA). Demonstrating a decentralized sensing in a hardware implementation, we present a novel inductive sensor board designed for distributed sensing and evaluate its ability to estimate global object properties, such as the geometric center, through local interactions and computations. Experiments demonstrate that NCA-based sensing networks accurately estimate object position at 0.24 times the inter sensor distance. They maintain resilience under sensor faults and noise, and scale seamlessly across varying network sizes. These findings underscore the potential of local, decentralized computations to enable scalable, fault-tolerant, and noise-resilient object property estimation in DMS, Comment: 8 pages, 8 figures

The analyses of ancient coins, and especially the identification of those struck with the same die, provides invaluable information for archaeologists and historians. Nowadays, these die links are identified manually, which makes the process laborious, if not impossible when big treasures are discovered as the number of comparisons is too large. This study introduces advances that promise to streamline and enhance archaeological coin analysis. Our contributions include: 1) First publicly accessible labeled dataset of coin pictures (329 images) for die link detection, facilitating method benchmarking; 2) Novel SSIM-based scoring method for rapid and accurate discrimination of coin pairs, outperforming current techniques used in this research field; 3) Evaluation of clustering techniques using our score, demonstrating near-perfect die link identification. We provide datasets, to foster future research and the development of even more powerful tools for archaeology, and more particularly for numismatics.

This paper presents a methodology combining multimodal semantic analysis with an eye-tracking experimental protocol to investigate the cognitive effort involved in understanding the communication of future scenarios. To demonstrate the methodology, we conduct a pilot study examining how visual fixation patterns vary during the evaluation of valence and counterfactuality in fictional ad pieces describing futuristic scenarios, using a portable eye tracker. Participants eye movements are recorded while evaluating the stimuli and describing them to a conversation partner. Gaze patterns are analyzed alongside semantic representations of the stimuli and participants descriptions, constructed from a frame semantic annotation of both linguistic and visual modalities. Preliminary results show that far-future and pessimistic scenarios are associated with longer fixations and more erratic saccades, supporting the hypothesis that fractures in the base spaces underlying the interpretation of future scenarios increase cognitive load for comprehenders.

Bioprocesses are often characterised by nonlinear and uncertain dynamics, posing particular challenges for model predictive control (MPC) algorithms due to their computational demands when applied to nonlinear systems. Recent advances in optimal control theory have demonstrated that concepts from convex optimisation, tube MPC, and differences of convex functions (DC) enable efficient, robust online process control. Our approach is based on DC decompositions of nonlinear dynamics and successive linearisations around predicted trajectories. By convexity, the linearisation errors have tight bounds and can be treated as bounded disturbances within a robust tube MPC framework. We describe a systematic, data-driven method for computing DC model representations using deep learning neural networks with a special convex structure, and explain how the resulting MPC optimisation can be solved using convex programming. For the problem of maximising product formation in a cultivation with uncertain model parameters, we design a controller that ensures robust constraint satisfaction and allows online estimation of unknown model parameters. Our results indicate that this method is a promising solution for computationally tractable, robust MPC of bioprocesses.

The Hong-Ou-Mandel effect is a paradigmatic quantum phenomenon demonstrating the interference of two indistinguishable photons that are linearly coupled at a 50:50 beam splitter. Here, we transpose such a two-particle quantum interference effect to the nonlinear regime, when two single photons are impinging on a parametric down-conversion crystal. Formally, this transposition amounts to exchanging space and time variables, giving rise to an unknown form of timelike quantum interference. The two-photon component of the output state is a superposition of the incident photons being either transmitted or reborn, that is, replaced by indistinguishable substitutes due to their interaction with the nonlinear crystal. We experimentally demonstrate the suppression of the probability of detecting precisely one photon pair when the amplification gain is tuned to 2, which arises from the destructive interference between the transmitted and reborn photon pairs. This heretofore unobserved quantum manifestation of indistinguishability in time pushes nonlinear quantum interference towards a new regime with multiple photons. Hence, composing this effect with larger linear optical circuits should provide a tool to generate multimode quantum non-Gaussian states, which are essential resources for photonic quantum computers.

In this paper, we define a family of dimensions for Borel measures that lie between the Hausdorff and Minkowski dimensions for measures, analogous to the intermediate dimensions of sets. Previously, Hare et. al. in [10] defined intermediate dimensions that interpolate between the Minkowski and Assouad dimensions for measures. Additionally, Fraser, in [7] introduced intermediate dimensions that interpolate between the Fourier and Hausdorff dimensions of measures. Our results address a "gap" in the study of dimension interpolation for measures, almost completing the spectrum of intermediate dimensions for measures: from Fourier to Assouad dimensions. Furthermore, Theorem 3.11 can be interpreted as a "reverse Frostman" lemma for intermediate dimensions. We also obtain a capacity-theoretic definition that enables us to estimate the intermediate dimensions of pushforward measures by projections.

Multi-Agent Deep Reinforcement Learning (MADRL) was proven efficient in solving complex problems in robotics or games, yet most of the trained models are hard to interpret. While learning intrinsically interpretable models remains a prominent approach, its scalability and flexibility are limited in handling complex tasks or multi-agent dynamics. This paper advocates for direct interpretability, generating post hoc explanations directly from trained models, as a versatile and scalable alternative, offering insights into agents' behaviour, emergent phenomena, and biases without altering models' architectures. We explore modern methods, including relevance backpropagation, knowledge edition, model steering, activation patching, sparse autoencoders and circuit discovery, to highlight their applicability to single-agent, multi-agent, and training process challenges. By addressing MADRL interpretability, we propose directions aiming to advance active topics such as team identification, swarm coordination and sample efficiency.

There has been tremendous progress in the physical realization of quantum computing hardware in recent times, bringing us closer than ever before to realizing the promise of quantum computing. However, noise continues to pose a crucial challenge when it comes to scaling up present day quantum processors. While decoherence limits the qubits ability to store information for long periods in the presence of uncontrollable noise sources, the erroneous implementation of control methods for state preparation and measurements leads to faulty implementations of quantum circuits. Conventional noise spectroscopy protocols can characterize and model environmental noise but are usually resource intensive and lengthy. Moreover, the underlying noise can vary in nature over time, making noise profile extraction futile as this new information cannot be harnessed to improve quantum error correction or dynamical decoupling protocols. In this work, we address this challenge using a machine learning-based methodology to quickly extract noise spectra of multiple qubits and demonstrate a possible noise mitigation strategy. The procedure involves implementing undemanding dynamical decoupling sequences to record coherence decays of the investigated qubits and then predict the underlying noise spectra with the help of a convolution neural network pre-trained on a synthetic dataset. While our protocol is virtually hardware-agnostic, we validate its effectiveness using superconducting qubits available on the IBM Quantum platform. We further use these rapidly obtained, yet accurate, noise spectra to design bespoke dynamic decoupling sequences and perform time-dependent noise spectroscopy., Comment: First two authors have equal contribution to this work

We derive asymptotically precise expressions for test and train errors of denoising score matching (DSM) in generative diffusion models. The score function is parameterized by random features neural networks, with the target distribution being $d$-dimensional standard Gaussian. We operate in a regime where the dimension $d$, number of data samples $n$, and number of features $p$ tend to infinity while keeping the ratios $\psi_n=\frac{n}{d}$ and $\psi_p=\frac{p}{d}$ fixed. By characterizing the test and train errors, we identify regimes of generalization and memorization in diffusion models. Furthermore, our work sheds light on the conditions enhancing either generalization or memorization. Consistent with prior empirical observations, our findings indicate that the model complexity ($p$) and the number of noise samples per data sample ($m$) used during DSM significantly influence generalization and memorization behaviors., Comment: 8 pages

We present a class of diffusion-based algorithms to draw samples from high-dimensional probability distributions given their unnormalized densities. Ideally, our methods can transport samples from a Gaussian distribution to a specified target distribution in finite time. Our approach relies on the stochastic interpolants framework to define a time-indexed collection of probability densities that bridge a Gaussian distribution to the target distribution. Subsequently, we derive a diffusion process that obeys the aforementioned probability density at each time instant. Obtaining such a diffusion process involves solving certain Hamilton-Jacobi-Bellman PDEs. We solve these PDEs using the theory of forward-backward stochastic differential equations (FBSDE) together with machine learning-based methods. Through numerical experiments, we demonstrate that our algorithm can effectively draw samples from distributions that conventional methods struggle to handle., Comment: 8 pages

We analyze the time reversed dynamics of generative diffusion models. If the exact empirical score function is used in a regime of large dimension and exponentially large number of samples, these models are known to undergo transitions between distinct dynamical regimes. We extend this analysis and compute the transitions for an analytically tractable manifold model where the statistical model for the data is a mixture of lower dimensional Gaussians embedded in higher dimensional space. We compute the so-called speciation and collapse transition times, as a function of the ratio of manifold-to-ambient space dimensions, and other characteristics of the data model. An important tool used in our analysis is the exact formula for the mutual information (or free energy) of Generalized Linear Models.

The interplay between stellar multiplicity and protoplanetary discs represents a cornerstone of modern astrophysics, offering key insights into the processes of planet formation. Protoplanetary discs act as cradles for planetary systems, yet their evolution and capacity to form planets are profoundly affected by gravitational forces within multiple stellar systems. This review synthesises recent advancements in observational and theoretical studies to explore the rich diversity of circumstellar and circumbinary discs within multiple stellar systems. We examine how stellar companions shape disc morphology through truncation, spirals, and misalignment. We also outline how dust dynamics and planetesimal formation are impacted by stellar multiplicity. On top of this, observations at high angular resolution reveal detailed disc structures, while simulations offer key insights into their evolution. Last, we consider the implications of stellar multiplicity for planetary system architectures, emphasising the diversity of planetary outcomes in such environments. Looking ahead, coordinated efforts combining high-resolution observations with advanced numerical models will be critical for unravelling the role of multiple stellar systems in shaping planetary formation and evolution., Comment: 37 pages, 6 figures, 2 tables, 315 references. Invited review article accepted in Symmetry