Search

The Pandora Software Development Kit and algorithm libraries perform reconstruction of neutrino interactions in liquid argon time projection chamber detectors. Pandora is the primary event reconstruction software used at the Deep Underground Neutrino Experiment, which will operate four large-scale liquid argon time projection chambers at the far detector site in South Dakota, producing high-resolution images of charged particles emerging from neutrino interactions. While these high-resolution images provide excellent opportunities for physics, the complex topologies require sophisticated pattern recognition capabilities to interpret signals from the detectors as physically meaningful objects that form the inputs to physics analyses. A critical component is the identification of the neutrino interaction vertex. Subsequent reconstruction algorithms use this location to identify the individual primary particles and ensure they each result in a separate reconstructed particle. A new vertex-finding procedure described in this article integrates a U-ResNet neural network performing hit-level classification into the multi-algorithm approach used by Pandora to identify the neutrino interaction vertex. The machine learning solution is seamlessly integrated into a chain of pattern-recognition algorithms. The technique substantially outperforms the previous BDT-based solution, with a more than 20\% increase in the efficiency of sub-1\,cm vertex reconstruction across all neutrino flavours., Comment: 32 pages, 18 figures

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 2,700 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: 27 pages, 6 figures

UX Orionis stars are the most active young stars; they undergo sporadic fadings of 2 - 4 magnitudes in the V-band, due to variable circumstellar extinction caused by a nearly edge-on star-disc system. The long-lasting monitoring of a number of stars of this type with the Nordic Optical Telescope from 2019 to 2024 has given a rich collection of material of high-resolution (R ~ 25000) spectra obtained during different brightness states of the stars. In this paper, we present the results of observations for UX Ori itself. Until now only one spectrum of high resolution had been obtained for this star during brightness minimum, making it difficult to do a comprehensive analysis. Our aim is to analyse how different spectral lines change during such irregular fading events, when the star is going in and out of eclipses, obscured by dust along the line of sight. For this purpose we provide a comparative analysis of the profiles and equivalent widths of the spectral lines belonging to the different atoms and ions. In addition we compare the results for UX Ori with those made for another target in our sample: RR Tau. Common features of variability are revealed: 1) a strengthening of the H-alpha line relatively to the continuum during eclipses; 2) the appearance of additional emission on the frequencies of photospheric lines (e.g. Fe II, Ca II, Si II). The different behaviour of the spectral lines during fading found for UX Ori and RR Tau may be caused by two effects: a different contribution of the scattered light to the stellar flux during eclipses or a less intense disc wind of UX Ori.

We reconsider the footprints of radiation friction in a head on collision of a bunch of relativistic charged particles with a laser pulse by demonstrating that in a dense enough bunch forward and backward radiation and radiation friction are coherently enhanced. This should make it possible to observe radiation friction effects in laser-matter interactions at much lower energies and laser intensities than accepted ever previously. A simple estimate for the energy losses of the particles in the bunch over the collision due to radiation friction in terms of laser and bunch parameters is derived and validated by comparing with the results of three dimensional particle-in-cell simulations.

A classical and central problem in the theory of water waves is to classify parameter regimes for which nontrivial solitary waves exist. In the two-dimensional, irrotational, pure gravity case, the Froude number $Fr$ (a non-dimensional wave speed) plays the central role. So far, the best analytic result $Fr < \sqrt{2}$ was obtained by Starr (1947), while the numerical evidence of Longuet-Higgins & Fenton (1974) states $Fr \le 1.294$. On the other hand, as shown recently by Kozlov (2023), the upper bound $Fr < 1.399$ is related to the existence of subharmonic bifurcations of Stokes waves. In this paper, we develop a new strategy utilizing the flow force function and rigorously establish the improved upper bound $Fr < 1.37838$.

We consider vanishing viscosity approximations to solutions of the stochastic incompressible Euler equations in two space dimensions with additive noise. We identify sufficient and necessary conditions under which martingale solutions of the stochastic Euler equations satisfy an exact energy balance in mean. We find that the tightness of the laws of the approximating sequence of solutions of the stochastic Navier-Stokes equations in $L^2([0,T]\times D)$ is equivalent to the limiting martingale solution satisfying an energy balance in mean. Numerical simulations illustrate the theoretical findings.

A re-construction of the fundamentals of programming as a small mathematical theory (PRISM) based on elementary set theory. Highlights: $\bullet$ Zero axioms. No properties are assumed, all are proved (from standard set theory). $\bullet$ A single concept covers specifications and programs. $\bullet$ Its definition only involves one relation and one set. $\bullet$ Everything proceeds from three operations: choice, composition and restriction. $\bullet$ These techniques suffice to derive the axioms of classic papers on the "laws of programming" as consequences and prove them mechanically. $\bullet$ The ordinary subset operator suffices to define both the notion of program correctness and the concepts of specialization and refinement. $\bullet$ From this basis, the theory deduces dozens of theorems characterizing important properties of programs and programming. $\bullet$ All these theorems have been mechanically verified (using Isabelle/HOL); the proofs are available in a public repository. This paper is a considerable extension and rewrite of an earlier contribution [arXiv:1507.00723]

Minimally doubled fermions realize one pair of Dirac fermions on the lattice. Similarities to Kogut-Susskind fermions exist, namely, spin and taste degrees of freedom become intertwined, and a peculiar non-singlet chiral symmetry and ultralocality are maintained. However, charge conjugation, some space-time reflection symmetries and isotropy are broken by the cutoff. We address the most simple variant, Karsten-Wilczek fermions, in its parallel and in its perpendicular version. We derive the correct spin-taste representation from first principles. The spin-taste representation on the quark level permits construction of local or extended hadron interpolating operators for any spin-taste combination, albeit with contamination by parity partners and taste-symmetry breaking. We classify all interpolating operator for mesons and diquarks, and give examples for baryons. We also discuss the appropriate discretizations for taste-singlet or taste-isovector mass or chemical potential terms. We explain the counterterms in the spin-taste framework, and derive generic constraints on the parametric form and cutoff effects from the KW determinant and hadronic correlation functions. We derive how and why non-perturbative tuning schemes for the counterterms work, and obtain analytic, assumption-free, non-perturbative predictions for taste-symmetry breaking and other hadronic properties from first principles. In particular, we identify the origin and nature of two different types of taste-symmetry breaking cutoff effects. The few available numerical results for KW fermions validate these predictions., Comment: 51 pages, 5 figures

Blockchain is an emerging technology that enables new forms of decentralized software architectures, where distributed components can reach agreements on shared system states without trusting a central integration point. Blockchain provides a shared infrastructure to execute programs, called smart contracts, and to store data. Since blockchain technologies are at an early stage, there is a lack of a systematically organized knowledge providing a holistic view on designing software systems that use blockchain. We view blockchain as a component of a bigger software system, which requires patterns for using blockchain in the design of the software architecture. In this paper, we collect a list of patterns for blockchain-based applications. The pattern collection is categorized into five categories, including interaction with external world patterns, data management patterns, security patterns, structural patterns of contracts, and user interaction patterns. Some patterns are designed considering the nature of blockchain and how blockchains can be specifically introduced within real-world applications. Others are variants of existing design patterns applied in the context of blockchain-based applications and smart contracts., Comment: Accepted by Transactions on Pattern Languages of Programming (https://www.springer.com/series/8677?srsltid=AfmBOoomdUEOPa-9tSOF-DDhmUxUOwk1Vfj_0dkhwjrptiERHvxb-yfX)

Testing processes usually aim at high coverage, but loops severely limit coverage ambitions since the number of iterations is generally not predictable. Most testing teams address this issue by adopting the extreme solution of limiting themselves to branch coverage, which only considers loop executions that iterate the body either once or not at all. This approach misses any bug that only arises after two or more iterations. To achieve more meaningful coverage, testing strategies may unroll loops, in the sense of using executions that iterate loops up to n times for some n greater than one, chosen pragmatically in consideration of the available computational power. While loop unrolling is a standard part of compiler optimization techniques, its use in testing is far less common. Part of the reason is that the concept, while seemingly intuitive, lacks a generally accepted and precise specification. The present article provides a formal definition and a set of formal properties of unrolling. All the properties have mechanically been proved correct (through the Isabelle proof assistant). Using this definition as the conceptual basis, we have applied an unrolling strategy to an existing automated testing framework and report the results: how many more bugs get detected once we unroll loops more than once? These results provide a first assessment of whether unrolling should become a standard part of test generation and test coverage measurement.

The continuous displacive phase transition in SrTiO$_3$ near $T_c \approx 105$ K features a central elastic peak in neutron scattering investigations at temperatures above $T_c$, i.e., before the corresponding soft phonon mode is overdamped upon cooling. The origin of this central peak is still not understood. Here, we report an inelastic x-ray scattering investigation of the cubic-to-tetragonal phase transition in SrTiO$_3$. We compare quantitatively measurements of the soft phonon mode on two differently grown samples and discuss the findings regarding results from thermodynamic probes such as specific heat and thermal conductivity. Furthermore, we use inelastic x-ray scattering to perform elastic scans with both high momentum- and milli-electronvolt energy-resolution and, thus, be able to separate elastic intensities of the central peak from low-energy quasielastic phonon scattering. Our results indicate that the evolution of the soft mode is similar in both samples though the intensities of the central peak differ by a factor of four. Measurements revealing anisotropic correlation lengths on cooling towards $T_c$, indicate that local properties of the crystals to which collective lattice excitations are insensitive are likely at the origin of the central elastic line in SrTiO$_3$., Comment: Manuscript contains 8 pages, 4 figures. Supplementary information contains 26 pages, 13 figures

Efficient solid-liquid separation is crucial in industries like mining, but traditional chamber filter presses depend heavily on manual monitoring, leading to inefficiencies, downtime, and resource wastage. This paper introduces a machine learning-powered digital twin framework to improve operational flexibility and predictive control. A key challenge addressed is the degradation of the filter medium due to repeated cycles and clogging, which reduces filtration efficiency. To solve this, a neural network-based predictive model was developed to forecast operational parameters, such as pressure and flow rates, under various conditions. This predictive capability allows for optimized filtration cycles, reduced downtime, and improved process efficiency. Additionally, the model predicts the filter mediums lifespan, aiding in maintenance planning and resource sustainability. The digital twin framework enables seamless data exchange between filter press sensors and the predictive model, ensuring continuous updates to the training data and enhancing accuracy over time. Two neural network architectures, feedforward and recurrent, were evaluated. The recurrent neural network outperformed the feedforward model, demonstrating superior generalization. It achieved a relative $L^2$-norm error of $5\%$ for pressure and $9.3\%$ for flow rate prediction on partially known data. For completely unknown data, the relative errors were $18.4\%$ and $15.4\%$, respectively. Qualitative analysis showed strong alignment between predicted and measured data, with deviations within a confidence band of $8.2\%$ for pressure and $4.8\%$ for flow rate predictions. This work contributes an accurate predictive model, a new approach to predicting filter medium cycle impacts, and a real-time interface for model updates, ensuring adaptability to changing operational conditions.

Hierarchical clustering is a powerful tool for exploratory data analysis, organizing data into a tree of clusterings from which a partition can be chosen. This paper generalizes these ideas by proving that, for any reasonable hierarchy, one can optimally solve any center-based clustering objective over it (such as $k$-means). Moreover, these solutions can be found exceedingly quickly and are themselves necessarily hierarchical. Thus, given a cluster tree, we show that one can quickly access a plethora of new, equally meaningful hierarchies. Just as in standard hierarchical clustering, one can then choose any desired partition from these new hierarchies. We conclude by verifying the utility of our proposed techniques across datasets, hierarchies, and partitioning schemes.

We investigated the ultrafast structural dynamics of cyclobutanone following photoexcitation at $\lambda=200$ nm using gas-phase megaelectronvolt ultrafast electron diffraction. Our investigation complements the simulation studies of the same process within this special issue. It provides information about both electronic state population and structural dynamics through well-separable inelastic and elastic electron scattering signatures. We observe the depopulation of the photoexcited S$_2$ state of cyclobutanone with n3s Rydberg character through its inelastic electron scattering signature with a time constant of $(0.29 \pm 0.2)$ ps towards the S$_1$ state. The S$_1$ state population undergoes ring-opening via a Norrish Type-I reaction, likely while passing through a conical intersection with S$_0$. The corresponding structural changes can be tracked by elastic electron scattering signatures. These changes appear with a delay of $(0.14 \pm 0.05)$ ps with respect the initial photoexcitation, which is less than the S$_2$ depopulation time constant. This behavior provides evidence for the ballistic nature of the ring-opening once the S$_1$ state is reached. The resulting biradical species react further within $(1.2 \pm 0.2)$ ps via two rival fragmentation channels yielding ketene and ethylene, or propene and carbon monoxide. Our study showcases both the value of gas-phase ultrafast diffraction studies as an experimental benchmark for nonadiabatic dynamics simulation methods and the limits in the interpretation of such experimental data without comparison to such simulations.

Large Language Models (LLMs) are increasingly used in working environments for a wide range of tasks, excelling at solving individual problems in isolation. However, are they also able to effectively collaborate over long-term interactions? To investigate this, we introduce MemoryCode, a synthetic multi-session dataset designed to test LLMs' ability to track and execute simple coding instructions amid irrelevant information, simulating a realistic setting. While all the models we tested handle isolated instructions well, even the performance of state-of-the-art models like GPT-4o deteriorates when instructions are spread across sessions. Our analysis suggests this is due to their failure to retrieve and integrate information over long instruction chains. Our results highlight a fundamental limitation of current LLMs, restricting their ability to collaborate effectively in long interactions.

Microchips are fundamental components of modern electronic devices, yet they remain opaque to the users who rely on them daily. This opacity, compounded by the complexity of global supply chains and the concealment of proprietary information, raises significant security, trust, and accountability issues. We investigate end users' understanding of microchips, exploring their perceptions of the societal implications and information needs regarding these essential technologies. Through an online survey with 250 participants, we found that while our participants were aware of some microchip applications, they lacked awareness of the broader security, societal, and economic implications. While our participants unanimously desired more information on microchips, their specific information needs were shaped by various factors such as the microchip's application environment and one's affinity for technology interaction. Our findings underscore the necessity for improving end users' awareness and understanding of microchips, and we provide possible directions to pursue this end.

High-resource languages such as English, enables the pretraining of high-quality large language models (LLMs). The same can not be said for most other languages as LLMs still underperform for non-English languages, likely due to a gap in the quality and diversity of the available multilingual pretraining corpora. In this work, we find that machine-translated texts from a single high-quality source language can contribute significantly to the pretraining quality of multilingual LLMs. We translate FineWeb-Edu, a high-quality English web dataset, into nine languages, resulting in a 1.7-trillion-token dataset, which we call TransWebEdu and pretrain a 1.3B-parameter model, TransWebLLM, from scratch on this dataset. Across nine non-English reasoning tasks, we show that TransWebLLM matches or outperforms state-of-the-art multilingual models trained using closed data, such as Llama3.2, Qwen2.5, and Gemma, despite using an order of magnitude less data. We demonstrate that adding less than 5% of TransWebEdu as domain-specific pretraining data sets a new state-of-the-art in Arabic, Italian, Indonesian, Swahili, and Welsh understanding and commonsense reasoning tasks. To promote reproducibility, we release our corpus, models, and training pipeline under Open Source Initiative-approved licenses.

The comprehensibility and reliability of data download packages (DDPs) provided under the General Data Protection Regulation's (GDPR) right of access are vital for both individuals and researchers. These DDPs enable users to understand and control their personal data, yet issues like complexity and incomplete information often limit their utility. Also, despite their growing use in research to study emerging online phenomena, little attention has been given to systematically assessing the reliability and comprehensibility of DDPs. To bridge this research gap, in this work, we perform a comparative analysis to assess the comprehensibility and reliability of DDPs provided by three major social media platforms, namely, TikTok, Instagram, and YouTube. By recruiting 400 participants across four countries, we assess the comprehensibility of DDPs across various requirements, including conciseness, transparency, intelligibility, and clear and plain language. Also, by leveraging automated bots and user-donated DDPs, we evaluate the reliability of DDPs across the three platforms. Among other things, we find notable differences across the three platforms in the data categories included in DDPs, inconsistencies in adherence to the GDPR requirements, and gaps in the reliability of the DDPs across platforms. Finally, using large language models, we demonstrate the feasibility of easily providing more comprehensible DDPs., Comment: This is a work in progress and the draft may undergo some changes in future

We present MicroBooNE's first search for dark sector $e^+e^-$ explanations of the long-standing MiniBooNE anomaly. The MiniBooNE anomaly has garnered significant attention over the past 20 years including previous MicroBooNE investigations into both anomalous electron and photon excesses, but its origin still remains unclear. In this letter, we provide the first direct test of dark sector models in which dark neutrinos, produced through neutrino-induced scattering, decay into missing energy and visible $e^+e^-$ pairs comprising the MiniBooNE anomaly. Many such models have recently gained traction as a viable solution to the anomaly while evading past bounds. Using an exposure of $6.87 \times 10^{20}$ protons-on-target in the Booster Neutrino Beam, we implement a selection targeting forward-going, coherently produced $e^+e^-$ events. After unblinding, we observe 95 events, which we compare with the constrained background-only prediction of $69.7 \pm 17.3$. This analysis sets the world's first direct limits on these dark sector models and, at the 95\% confidence level, excludes the majority of the parameter space viable as a solution to the MiniBooNE anomaly., Comment: 7 pages, 5 figures, Supplemental Materials included in Ancillary files

Higher-order information is crucial for relational learning in many domains where relationships extend beyond pairwise interactions. Hypergraphs provide a natural framework for modeling such relationships, which has motivated recent extensions of graph neural network architectures to hypergraphs. However, comparisons between hypergraph architectures and standard graph-level models remain limited. In this work, we systematically evaluate a selection of hypergraph-level and graph-level architectures, to determine their effectiveness in leveraging higher-order information in relational learning. Our results show that graph-level architectures applied to hypergraph expansions often outperform hypergraph-level ones, even on inputs that are naturally parametrized as hypergraphs. As an alternative approach for leveraging higher-order information, we propose hypergraph-level encodings based on classical hypergraph characteristics. While these encodings do not significantly improve hypergraph architectures, they yield substantial performance gains when combined with graph-level models. Our theoretical analysis shows that hypergraph-level encodings provably increase the representational power of message-passing graph neural networks beyond that of their graph-level counterparts.

We quantify the effect of starspots for the orbital elements of the spotted RS CVn binary $\lambda$ Andromedae ($\lambda$ And) and present an empirical correction. The aim is to obtain a more precise orbital solution that can be used to better study the system's severe orbital-rotational asynchronism. Phase-resolved high-resolution optical spectra were recorded over the course of 522 days in 2021-2022. We employed two facilities with medium and high resolution spectroscopy for the multiple activity analyses. Doppler imaging is used to reconstruct $\lambda$ And's starspots with a high resolution (R= 250 000) and high signal-to-noise ratio spectra. Optimized cross-correlation functions were used to measure precise radial velocities at a level of a few ten's of m/s. The spot-corrected radial velocities enable, on average, a threefold increase in precision of the individual orbital elements. The residual velocity jitter with a full range of 500 m/s is modulated by the rotation period of $\lambda$ And of 54.4$\pm$0.3 days. Our logarithmic gravity from spectrum synthesis of 2.8$\pm$0.2 together with the interferometrically determined stellar radius suggest a most-likely mass of the primary of $\approx$1.4 Msun. The small orbital mass function then implies a secondary mass of just $\approx$0.1 Msun, which is appropriate for an L-class brown dwarf. The Doppler image reconstructs a dominating cool spot with an umbral temperature difference of $\approx$1000 K with respect to the photosphere of 4660 K and is likely surrounded by a moat-like velocity field. Three more weaker spots add to the total surface spottedness, which is up to 25% of the visible surface. Seven optical chromospheric tracers show rotational modulation of their emission line fluxes in phase with the cool spots. This surface configuration appears to have been stable for the 522 days of our observations.

We report on a scale determination, scale setting, and determination of the strong coupling in the gradient flow scheme using the $N_f=2+1$ highly improved staggered quark (HISQ) ensembles generated by the HotQCD Collaboration for bare gauge couplings ranging from $\beta = 7.030$ to $8.400$. The gradient flow scales we obtain in this work are $\sqrt{t_0} = 0.14229(98)$~fm and $w_0 = 0.17190(140)$~fm. Using the decay constants of the kaon and $\eta_s$, as well as the bottomonium mass splitting from the literature, we also calculate the potential scale $r_1$, obtaining $r_1 = 0.3072(22)$~fm. We fit the flow scales to an Allton-type ansatz as a function of $\beta$, providing a polynomial expression that allows for the prediction of lattice spacings at new $beta$ values. As a secondary result, we make an attempt to determine $\Lambda_{\overline{\mathrm{MS}}}$ and use it to estimate the strong coupling in the $\overline{\mathrm{MS}}$ scheme., Comment: 20 pages, 11 figures

We present Pippo, a generative model capable of producing 1K resolution dense turnaround videos of a person from a single casually clicked photo. Pippo is a multi-view diffusion transformer and does not require any additional inputs - e.g., a fitted parametric model or camera parameters of the input image. We pre-train Pippo on 3B human images without captions, and conduct multi-view mid-training and post-training on studio captured humans. During mid-training, to quickly absorb the studio dataset, we denoise several (up to 48) views at low-resolution, and encode target cameras coarsely using a shallow MLP. During post-training, we denoise fewer views at high-resolution and use pixel-aligned controls (e.g., Spatial anchor and Plucker rays) to enable 3D consistent generations. At inference, we propose an attention biasing technique that allows Pippo to simultaneously generate greater than 5 times as many views as seen during training. Finally, we also introduce an improved metric to evaluate 3D consistency of multi-view generations, and show that Pippo outperforms existing works on multi-view human generation from a single image., Comment: Project Page - http://yashkant.github.io/pippo

We present Cryoscope -- a new 50 sq. deg field-of-view, 1.2 m aperture, K-dark survey telescope to be located at Dome C, Antarctica. Cryoscope has an innovative optical-thermal design wherein the entire telescope is cryogenically cooled. Cryoscope also explores new detector technology to cost-effectively tile the full focal plane. Leveraging the dark Antarctic sky and minimizing telescope thermal emission, Cryoscope achieves unprecedented deep, wide, fast and red observations, matching and exceeding volumetric survey speeds from the Ultraviolet Explorer, Vera Rubin Observatory, and Nancy Grace Roman Space Telescope. By providing coverage beyond wavelengths of 2 $\mu$m, we aim to create the most comprehensive dynamic movie of the most obscured reaches of the Universe. Cryoscope will be a dedicated discovery engine for electromagnetic emission from coalescing compact binaries, Earth-like exoplanets orbiting cold stars, and multiple facets of time-domain, stellar and solar system science. In this paper, we describe the scientific drivers and technical innovations for this new discovery engine operating in the K-dark passband, why we choose to deploy it in Antarctica, and the status of a fifth-scale prototype designed as a Pathfinder to retire technological risks prior to full-scale implementation., Comment: 36 pages, 20 figures, 4 tables; submitted to PASP on 2025-02-09

In time-series analysis, nonlinear temporal misalignment remains a pivotal challenge that forestalls even simple averaging. Since its introduction, the Diffeomorphic Temporal Alignment Net (DTAN), which we first introduced (Weber et al., 2019) and further developed in (Weber & Freifeld, 2023), has proven itself as an effective solution for this problem (these conference papers are earlier partial versions of the current manuscript). DTAN predicts and applies diffeomorphic transformations in an input-dependent manner, thus facilitating the joint alignment (JA) and averaging of time-series ensembles in an unsupervised or a weakly-supervised manner. The inherent challenges of the weakly/unsupervised setting, particularly the risk of trivial solutions through excessive signal distortion, are mitigated using either one of two distinct strategies: 1) a regularization term for warps; 2) using the Inverse Consistency Averaging Error (ICAE). The latter is a novel, regularization-free approach which also facilitates the JA of variable-length signals. We also further extend our framework to incorporate multi-task learning (MT-DTAN), enabling simultaneous time-series alignment and classification. Additionally, we conduct a comprehensive evaluation of different backbone architectures, demonstrating their efficacy in time-series alignment tasks. Finally, we showcase the utility of our approach in enabling Principal Component Analysis (PCA) for misaligned time-series data. Extensive experiments across 128 UCR datasets validate the superiority of our approach over contemporary averaging methods, including both traditional and learning-based approaches, marking a significant advancement in the field of time-series analysis., Comment: This manuscript covers and extends the papers: Diffeomorphic Temporal Alignment Nets (DTAN; NeruIPS 2019) and Regularization-free Diffeomorphic Temporal Alignment Nets (ICML 2023). Additional contributions: Multi-tasking DTAN, PCA-DTAN and more

This article presents the first search for neutrino-induced neutral current coherent single-photon production (NC coherent 1$\gamma$). The search makes use of data from the MicroBooNE 85-tonne active volume liquid argon time projection chamber detector, situated in the Fermilab Booster Neutrino Beam (BNB), with an average neutrino energy of $\langle E_{\nu}\rangle \sim 0.8$ GeV. A targeted selection of candidate neutrino interactions with a single photon-like electromagnetic shower in the final state and no visible vertex activity was developed to search for the NC coherent 1$\gamma$ process, along with two auxiliary selections used to constrain the dominant background from NC$\pi^0$ production. With an integrated exposure of $6.87 \times 10^{20}$ protons on target delivered by the BNB, we set the world's first limit for this rare process, corresponding to an upper limit on the flux-averaged cross section of $\sigma<1.49 \times 10^{-41}\text{cm}^2$ at 90\% C.L., Comment: 20 pages, 17 figures

We present an inclusive search for anomalous production of single-photon events from neutrino interactions in the MicroBooNE experiment. The search and its signal definition are motivated by the previous observation of a low-energy excess of electromagnetic shower events from the MiniBooNE experiment. We use the Wire-Cell reconstruction framework to select a sample of inclusive single-photon final-state interactions with a final efficiency and purity of 7.0% and 40.2%, respectively. We leverage simultaneous measurements of sidebands of charged current $\nu_{\mu}$ interactions and neutral current interactions producing $\pi^{0}$ mesons to constrain signal and background predictions and reduce uncertainties. We perform a blind analysis using a dataset collected from February 2016 to July 2018, corresponding to an exposure of $6.34\times10^{20}$ protons on target from the Booster Neutrino Beam (BNB) at Fermilab. In the full signal region, we observe agreement between the data and the prediction, with a goodness-of-fit $p$-value of 0.11. We then isolate a sub-sample of these events containing no visible protons, and observe $93\pm22\text{(stat.)}\pm35\text{(syst.)}$ data events above prediction, corresponding to just above $2\sigma$ local significance, concentrated at shower energies below 600 MeV., Comment: 9 pages, 6 figures

We report Growth--Guided Local Ordering, a novel mechanism of symmetry reduction in disordered crystals. This mechanism operates through the directional ordering of point defects during crystal growth, where defect correlations develop preferentially along the growth direction, resulting in reduced symmetry that persists in the final structure through the spatial distribution of defects. We demonstrate this phenomenon in Mn[Co]-Prussian Blue Analogues, disordered cyanide crystals containing numerous Co(CN)$_6$ vacancies. Single crystal diffuse scattering reveals pronounced anisotropy in vacancy distribution: strong correlations along [001] growth direction contrast with weak correlations perpendicular to it. This local ordering reduces the Laue symmetry to tetragonal $4/mmm$, evident in properties such as birefringence, while the average structure retains cubic $m\bar 3m$ symmetry. When growth proceeds along [111] direction, the same mechanism produces domains with trigonal symmetry. Because this mechanism relies on fundamental aspects of crystal growth rather than specific material properties, it offers a general strategy for symmetry control in disordered crystals. Crucially, it transforms the complex task of altering crystal symmetry into the more manageable challenge of controlling growth direction, achievable through various established techniques such as the use of surfactants during crystallization., Comment: 19 pages, 11 figures

We report results from an updated search for neutral current (NC) resonant $\Delta$(1232) baryon production and subsequent $\Delta$ radiative decay (NC $\Delta\rightarrow N \gamma$). We consider events with and without final state protons; events with a proton can be compared with the kinematics of a $\Delta(1232)$ baryon decay, while events without a visible proton represent a more generic phase space. In order to maximize sensitivity to each topology, we simultaneously make use of two different reconstruction paradigms, Pandora and Wire-Cell, which have complementary strengths, and select mostly orthogonal sets of events. Considering an overall scaling of the NC $\Delta\rightarrow N \gamma$ rate as an explanation of the MiniBooNE anomaly, our data exclude this hypothesis at 94.4% CL. When we decouple the expected correlations between NC $\Delta\rightarrow N \gamma$ events with and without final state protons, and allow independent scaling of both types of events, our data exclude explanations in which excess events have associated protons, and do not exclude explanations in which excess events have no associated protons.

We propose a new method to extract discriminant and explainable features from a particular machine learning model, i.e., a combination of the scattering transform and the multiclass logistic regression. Although this model is well-known for its ability to learn various signal classes with high classification rate, it remains elusive to understand why it can generate such successful classification, mainly due to the nonlinearity of the scattering transform. In order to uncover the meaning of the scattering transform coefficients selected by the multiclass logistic regression (with the Lasso penalty), we adopt zeroth-order optimization algorithms to search an input pattern that maximizes the class probability of a class of interest given the learned model. In order to do so, it turns out that imposing sparsity and smoothness of input patterns is important. We demonstrate the effectiveness of our proposed method using a couple of synthetic time-series classification problems., Comment: 5 pages; 6 figures; submitted to 2025 IEEE Statistical Signal Processing Workshop

Accurate tumor classification is essential for selecting effective treatments, but current methods have limitations. Standard tumor grading, which categorizes tumors based on cell differentiation, is not recommended as a stand-alone procedure, as some well-differentiated tumors can be malignant. Tumor heterogeneity assessment via single-cell sequencing offers profound insights but can be costly and may still require significant manual intervention. Many existing statistical machine learning methods for tumor data still require complex pre-processing of MRI and histopathological data. In this paper, we propose to build on a mathematical model that simulates tumor evolution (O\.{z}a\'{n}ski (2017)) and generate artificial datasets for tumor classification. Tumor heterogeneity is estimated using normalized entropy, with a threshold to classify tumors as having high or low heterogeneity. Our contributions are threefold: (1) the cut and graph generation processes from the artificial data, (2) the design of tumor features, and (3) the construction of Block Graph Neural Networks (BGNN), a Graph Neural Network-based approach to predict tumor heterogeneity. The experimental results reveal that the combination of the proposed features and models yields excellent results on artificially generated data ($89.67\%$ accuracy on the test data). In particular, in alignment with the emerging trends in AI-assisted grading and spatial transcriptomics, our results suggest that enriching traditional grading methods with birth (e.g., Ki-67 proliferation index) and death markers can improve heterogeneity prediction and enhance tumor classification., Comment: 27 pages, 8 figures

We present Fillerbuster, a method that completes unknown regions of a 3D scene by utilizing a novel large-scale multi-view latent diffusion transformer. Casual captures are often sparse and miss surrounding content behind objects or above the scene. Existing methods are not suitable for handling this challenge as they focus on making the known pixels look good with sparse-view priors, or on creating the missing sides of objects from just one or two photos. In reality, we often have hundreds of input frames and want to complete areas that are missing and unobserved from the input frames. Additionally, the images often do not have known camera parameters. Our solution is to train a generative model that can consume a large context of input frames while generating unknown target views and recovering image poses when desired. We show results where we complete partial captures on two existing datasets. We also present an uncalibrated scene completion task where our unified model predicts both poses and creates new content. Our model is the first to predict many images and poses together for scene completion., Comment: Project page at https://ethanweber.me/fillerbuster/

Quantum Key Distribution (QKD) is currently being discussed as a technology to safeguard communication in a future where quantum computers compromise traditional public-key cryptosystems. In this paper, we conduct a comprehensive security evaluation of QKD-based solutions, focusing on real-world use cases sourced from academic literature and industry reports. We analyze these use cases, assess their security and identify the possible advantages of deploying QKD-based solutions. We further compare QKD-based solutions with Post-Quantum Cryptography (PQC), the alternative approach to achieving security when quantum computers compromise traditional public-key cryptosystems, evaluating their respective suitability for each scenario. Based on this comparative analysis, we critically discuss and comment on which use cases QKD is suited for, considering factors such as implementation complexity, scalability, and long-term security. Our findings contribute to a better understanding of the role QKD could play in future cryptographic infrastructures and offer guidance to decision-makers considering the deployment of QKD., Comment: Submitted to EPJ Quantum Technology

To model the structure of neutron stars (NSs) theoretically,it is common to consider layers with different density regimes. Matching the equation of state (EoS) for the crust and core and obtaining a suitable description of these extreme conditions are crucial for understanding the properties of these compact objects. In this work, we construct ten different NS EoSs incorporating three distinct crust models, which are connected to the core using a thermodynamically and causally consistent formalism. For cold NSs, we propose a linear relationship between pressure and energy density in a narrow region between the crust and core, effectively establishing an interpolation function in the pressure-baryonic chemical potential plane. We then compare this EoS matching method with the classical approach, which neglects causal and thermodynamic consistency. We solve the Tolman-Oppenheimer-Volkoff equation to obtain the mass-radius relationship and compare our results with observational constraints on NSs. Furthermore, we investigate the influence of the new matching formalism on non-radial oscillation frequencies and damping times. Our findings suggest that the method used to glue the crust and core EoS impacts NS observables, such as the radius, oscillation frequencies, and damping times of non-radial modes, which may be crucial for interpreting future gravitational wave observations from neutron star mergers or isolated pulsars. The effects are particularly noticeable for low-mass NSs, regardless of the specific EoS model chosen. In particular, we find that the $p_1$ oscillation mode exhibits significant differences in frequencies among alternative matching methods, whereas the fundamental $f$-mode remains unaffected by changes in crust models or interpolation schemes., Comment: 9 pages, 5 figures, IWARA2024 proceedings published in Astronomische Nachrichten

There is strong evidence for the conjecture that the $\lambda \phi^4$ QFT- model on 4-dimensional non-commutative Moyal space can be non-perturbatively constructed. As preparation, in this paper we construct the 2-dimensional case with the method of stochastic quantization. We show the local well-posedness and global well-posedness of the stochastic quantization equation, leading to a construction of the Moyal $\lambda \phi^4_2$ measure for any non-negative coupling constant $\lambda$., Comment: 68 pages, 45 pictures

Sparse regularization techniques are well-established in machine learning, yet their application in neural networks remains challenging due to the non-differentiability of penalties like the $L_1$ norm, which is incompatible with stochastic gradient descent. A promising alternative is shallow weight factorization, where weights are decomposed into two factors, allowing for smooth optimization of $L_1$-penalized neural networks by adding differentiable $L_2$ regularization to the factors. In this work, we introduce deep weight factorization, extending previous shallow approaches to more than two factors. We theoretically establish equivalence of our deep factorization with non-convex sparse regularization and analyze its impact on training dynamics and optimization. Due to the limitations posed by standard training practices, we propose a tailored initialization scheme and identify important learning rate requirements necessary for training factorized networks. We demonstrate the effectiveness of our deep weight factorization through experiments on various architectures and datasets, consistently outperforming its shallow counterpart and widely used pruning methods., Comment: accepted at ICLR 2025

The increasing technological advancements towards miniaturized physiological measuring devices have enabled continuous monitoring of epileptic patients outside of specialized environments. The large amounts of data that can be recorded with such devices holds significant potential for developing automated seizure detection frameworks. In this work, we present SeizeIT2, the first open dataset of wearable data recorded in patients with focal epilepsy. The dataset comprises more than 11,000 hours of multimodal data, including behind-the-ear electroencephalography, electrocardiography, electromyography and movement (accelerometer and gyroscope) data. The dataset contains 886 focal seizures recorded from 125 patients across five different European Epileptic Monitoring Centers. We present a suggestive training/validation split to propel the development of AI methodologies for seizure detection, as well as two benchmark approaches and evaluation metrics. The dataset can be accessed on OpenNeuro and is stored in Brain Imaging Data Structure (BIDS) format., Comment: 9 pages, 5 figures, 2 tables, paper in submission on Scientific Data Journal from Nature

The strong coupling $\alpha_\mathrm{s}$ can be obtained from the static energy as shown in previous lattices studies. For short distances, the static energy can be calculated both on the lattice with the use of Wilson line correlators, and with the perturbation theory up to three loop accuracy with leading ultrasoft log resummation. Comparing the perturbative expression and lattice data allows for precise determination of $\alpha_\mathrm{s}(m_Z)$. We will present preliminary results for the determination of $\alpha_\mathrm{s} {(M_Z)}$ in (2+1+1)-flavor QCD using the configurations made availableby the MILC-collaboration with smallest lattice spacing reaching 0.0321fm., Comment: 10 pages, 4 figures, Contribution to the 41st International Symposium on Lattice Field Theory (LATTICE2024), 28 July - 3 August 2024, Liverpool, UK

Monoclonal antibodies (mAbs) represent one of the most prevalent FDA-approved modalities for treating autoimmune diseases, infectious diseases, and cancers. However, discovery and development of therapeutic antibodies remains a time-consuming and expensive process. Recent advancements in machine learning (ML) and artificial intelligence (AI) have shown significant promise in revolutionizing antibody discovery and optimization. In particular, models that predict antibody biological activity enable in-silico evaluation of binding and functional properties; such models can prioritize antibodies with the highest likelihoods of success in costly and time-intensive laboratory testing procedures. We here explore an AI model for predicting the binding and receptor blocking activity of antibodies against influenza A hemagglutinin (HA) antigens. Our present model is developed with the MAMMAL framework for biologics discovery to predict antibody-antigen interactions using only sequence information. To evaluate the model's performance, we tested it under various data split conditions to mimic real-world scenarios. Our models achieved an AUROC $\geq$ 0.91 for predicting the activity of existing antibodies against seen HAs and an AUROC of 0.9 for unseen HAs. For novel antibody activity prediction, the AUROC was 0.73, which further declined to 0.63-0.66 under stringent constraints on similarity to existing antibodies. These results demonstrate the potential of AI foundation models to transform antibody design by reducing dependence on extensive laboratory testing and enabling more efficient prioritization of antibody candidates. Moreover, our findings emphasize the critical importance of diverse and comprehensive antibody datasets to improve the generalization of prediction models, particularly for novel antibody development.

In this article we introduce the notion of Floer function which has the property that the Hessian is a Fredholm operator of index zero in a scale of Hilbert spaces. Since the Hessian has a complicated transformation under chart transition, in general this is not an intrinsic condition. Therefore we introduce the concept of Floerfolds for which we show that the notion of Floer function is intrinsic., Comment: 19 pages

Polytopes are one of the most primitive concepts underlying geometry. Discovery and study of polytopes with complex structures provides a means of advancing scientific knowledge. Construction of polytopes with specific extremal structure is very difficult and time-consuming. Having an automated tool for the generation of such extremal examples is therefore of great value. We present an Artificial Intelligence system capable of generating novel polytopes with very high complexity, whose abilities we demonstrate in three different and challenging scenarios: the Hirsch Conjecture, the k-neighbourly problem and the longest monotone paths problem. For each of these three problems the system was able to generate novel examples, which match or surpass the best previously known bounds. Our main focus was the Hirsch Conjecture, which had remained an open problem for over 50 years. The highly parallel A.I. system presented in this paper was able to generate millions of examples, with many of them surpassing best known previous results and possessing properties not present in the earlier human-constructed examples. For comparison, it took leading human experts over 50 years to handcraft the first example of a polytope exceeding the bound conjectured by Hirsch, and in the decade since humans were able to construct only a scarce few families of such counterexample polytopes. With the adoption of computer-aided methods, the creation of new examples of mathematical objects stops being a domain reserved only for human expertise. Advances in A.I. provide mathematicians with yet another powerful tool in advancing mathematical knowledge. The results presented demonstrate that A.I. is capable of addressing problems in geometry recognized as extremely hard, and also to produce extremal examples different in nature from the ones constructed by humans., Comment: 18 pages, 5 figures

The circumgalactic medium (CGM) plays a critical role in galaxy evolution, influencing gas flows, feedback processes, and galactic dynamics. Observations show a substantial cold gas reservoir in the CGM, but the mechanisms driving its formation and evolution remain unclear. Cosmic rays (CRs), as a source of non-thermal pressure, are increasingly recognized as key regulators of cold gas dynamics. This study explores how CRs affect cold clouds that condense from the hot CGM via thermal instability (TI). Using 3D CR-magnetohydrodynamic (CRMHD) simulations with AREPO, we assess the impact of various CR transport models on cold gas evolution. Under purely advective CR transport, CR pressure significantly suppresses the collapse of thermally unstable regions, altering the CGM's structure. In contrast, realistic CR transport models reveal that CRs escape collapsing regions via streaming and diffusion along magnetic fields, diminishing their influence on the thermal and dynamic structure of the cold CGM. The ratio of the CR escape time to the cloud collapse time emerges as a critical factor in determining the impact of CRs on TI. CRs remain confined within cold clouds when effective CR diffusion is slow which maximizes their pressure support and inhibits collapse. Fast effective CR diffusion, as realized in our 2-moment CRMHD model, facilitates rapid CR escape, reducing their stabilizing effect. This realistic CR transport model shows a wide dynamic range of the effective CR diffusion coefficient, ranging from $10^{29}$ to $10^{30}\,\mathrm{cm^{2}\,s^{-1}}$ for thermally- to CR-dominated atmospheres, respectively. In addition to these CR transport-related effects, we demonstrate that high numerical resolution is crucial to avoid spuriously large clouds formed in low-resolution simulations, which would result in overly long CR escape times and artificially amplified CR pressure support., Comment: 20 pages and 17 figures in main text; submitted to A&A

The contributions of this short technical note are two-fold. Firstly, we introduce a modified version of a generalized Bellman-Ford algorithm calculating the value function of optimal control problems defined on hyper-graphs. Those Bellman-Ford algorithms can be used in particular for the synthesis of near-optimal controllers by the principle of symbolic control. Our modification causes less nodes of the hyper-graph being iterated during the execution compared to our initial version of the algorithm published in 2020. Our second contribution lies in the field of Plan recognition applied to drone missions driven by symbolic controllers. We address and resolve the Plan and Goal Recognition monitor's dependence on a pre-defined initial guess for a drone's task allocation and mission execution. To validate the enhanced implementation, we use a more challenging scenario for UAV-based aerial firefighting, demonstrating the practical applicability and robustness of the system architecture.

We present an extensive study of 4H-SiC nanomechanical resonators electrochemically etched out of a monocrystalline wafer. Combining piezo-driven interferometric determination of the mechanical spectra with scanning-laser-Doppler vibrometry, an unambiguous assignment of resonance peaks to flexural and torsional modes is achieved. The investigation of multiple harmonic eigenmodes of singly and doubly clamped resonators with varying geometry allows for a comprehensive characterization. Excellent intrinsic mechanical quality factors up to $2\times10^5$ are found at room temperature, approaching the thermoelastic limit at eigenfrquencies exceeding 10 MHz. Mechanical stress is essentially absent. Young's modulus in agreement with literature. These findings are robust under post-processing treatments, in particular atomic layer etching and high-temperature thermal annealing. The resulting on-chip high-quality mechanical resonators represent a valuable technological element for a broad range of applications. In particular, the monolithic architecture meets the requirements of spin-based photonic quantum technologies on the upcoming SiC platform.

This article provides a systematic up-to-date survey of automatic summarization techniques, datasets, models, and evaluation methods in the legal domain. Through specific source selection criteria, we thoroughly review over 120 papers spanning the modern `transformer' era of natural language processing (NLP), thus filling a gap in existing systematic surveys on the matter. We present existing research along several axes and discuss trends, challenges, and opportunities for future research.

This study documents the relationship between computer skills/digital literacy and influenza vaccination take-up among older adults in Europe during and after the COVID-19 pandemic. Using data from the Survey of Health, Aging and Retirement in Europe, we find a positive partial association between influenza vaccination take-up and two indicators of computer skills/digital literacy, self-assessed pre-pandemic computer skills and having used a computer at work in any pre-pandemic job. We do not estimate significant behavioural changes for individuals with better computer skills that may have been driven by spillover effects from the pandemic experience.

The fundamental understanding of the root causes of failure requires information on atomic-scale processes. Molecular dynamics (MD) simulations are widely used to provide these insights while maintaining chemical specificity. However, the boundary conditions that can be applied in MD fracture simulations are limited to the linear-elastic fracture mechanics (LEFM) far-field solution, posing restrictions in simulating bending or shearing deformations commonly applied in classical experimental setups. This has to date prevented the accurate and precise prediction of key fracture quantities such as critical stress intensity factors. To overcome these limitations, we here apply the domain-decomposition Capriccio method to couple atomistic MD samples representing silica glass synthesized at various quenching rates with the finite element (FE) method and perform both mode I (tension, three- and four-point bending) and mode III simulations at different loading rates. We investigate multiple criteria to best identify the onset of crack propagation based on the virial stress, the number of pair interactions, the kinetic energy/temperature, the crack velocity, and the crack opening displacement. Based on this, we propose a novel protocol that allows us to determine the fracture toughness of silica glass under mode I and III conditions from atomistic data with increased fidelity, with our results being in good agreement with experimental data and predictions from LEFM. Overall, our contribution demonstrates how coupled FE-MD simulations enable chemically specific quantitative predictions of the fracture behavior of amorphous materials under arbitrary mechanical loading conditions.

We present a dynamic theory for the droplet size distribution in chemically active emulsions, considering a simple ternary mixture with a conserved density. We show that the collective behavior of many droplets, such as monodispersity, emerges through a coupling of the conserved density in the far field. Using our theory, we determine the stationary state of such emulsions and characterize the relaxation of the droplet size distributions at late times. A key finding is a universal scaling behavior, leading to the collapse of the rescaled size distributions at late times. Our results suggest that the key features of our dynamic theory are generic and apply to the broader class of multi-component systems with conservation laws and active chemical reactions.