Search

The purpose of this technical appendix is to share detailed results and more fully describe the sample and methods used to produce the research brief, "COVID's Impact on Science Achievement: Trends from 2019 through 2024. We investigated three main research questions in this brief: 1) How did science achievement in 2021 and 2024 compare to achievement in 2019, before the pandemic? 2) How much additional schooling was required in spring 2021 and spring 2024 to return to spring 2019 levels? 3) How do these patterns differ by race/ethnicity?

This brief continues ongoing research by NWEA® examining the degree to which the COVID-19 pandemic, and its associated school closures, influenced student learning. Prior research focused on math and reading and found that the pandemic's negative impact steadily accumulated during the 2020-2021 school year, with significant disparities in achievement and growth compared to prepandemic trends (Lewis, Kuhfeld, Ruzek, & McEachin, 2021). Although there were modest rebounds in reading and math during the 2021-2022 school year, progress toward pandemic recovery largely stalled in 2022-2023 and 2023-2024, with achievement and growth trends still trailing behind prepandemic levels (Kuhfeld & Lewis, 2022; Lewis & Kuhfeld, 2023; Lewis & Kuhfeld, 2024). Building on previous research series focused on math and reading achievement, this brief explores how the pandemic influenced student achievement in science. It examines science test scores across seven school years to understand how COVID-19 influenced science achievement.

The Latino and Adult Student Success (LASS) Academy is a multi-year initiative administered by CAEL (the Council for Adult and Experiential Learning) focused on supporting Hispanic-Serving Institutions (HSIs) in their efforts to improve outcomes for adult and Latino adult learners. In its first phase, from 2018 through 2021, the LASS Academy supported fifteen postsecondary initiatives across the country. In the second phase, from 2022 through 2024, the LASS Academy continued its work with four Texas institutions: Austin Community College (ACC), South Texas College (STC), Texas A&M University-Corpus Christi (TAMU-CC), and the University of Texas Rio Grande Valley (UTRGV). The LASS Academy model facilitates institutional change through four key activities: identify key priorities and action steps; provide technical assistance; collect, analyze, and report data; and facilitate peer-to-peer learning. During the second LASS Academy, each of the four participating institutions made progress toward their goals, buttressed by change management support from CAEL and collaboration with other technical assistance partners. Throughout this work, the institutions gained important insights into ways they can improve their services for adult and Latino adult learners.

This brief is a continuation of NWEA's research series examining the impact of the COVID-19 pandemic on student achievement and progress toward academic recovery. Initiated in the early phase of the pandemic, this series has leveraged NWEA's large national sample of longitudinal MAP® Growth™ data to track student performance and compare it to historical trends. The authors have tracked two critical aspects of student performance in the wake of the pandemic: achievement and growth. Achievement data reveals the extent of unfinished learning. The authors refer to the distance between current test scores and prepandemic trends as "achievement gaps." Growth data estimate how much test scores increase over time. They use these data to indicate whether students are making gains that keep pace with prepandemic trends. Understanding and tracking both achievement and growth is important. Achievement gaps quantify how much unfinished learning remains, while growth patterns help gauge the rate at which gaps will close. The system needs accelerated or above-average growth for students to catch up. The data provide insights into how effectively this is happening. The cumulative research shows that the harmful effects of the pandemic on student achievement steadily accumulated over the course of the 2020-21 school year (Lewis, Kuhfeld, Ruzek, & McEachin, 2021). Growth generally returned to, or slightly exceeded, prepandemic trends in the 2021-22 school year (Kuhfeld & Lewis, 2022). Progress stalled in 2022-23 when growth in nearly all grades fell short of prepandemic trends (Lewis & Kuhfeld, 2023). In this analysis, the authors use 2023-24 school year data to examine the current progress toward recovery. They examine test scores from approximately 7.7 million students currently in grades 3-8 in 22,400 public schools who have taken MAP Growth reading and math assessments since the onset of the pandemic. The "COVID sample" consists of six separate cohorts of students followed longitudinally across the last three school years. For instance, current fifth-graders are part of the grade 3-5 cohort; they measure this cohort's achievement across third grade in 2021-22, fourth grade in 2022-23, and fifth grade in 2023-24. The authors compared this COVID sample to a comparable group of 10 million students who tested in grades 3-8 in the pre-COVID school years of 2016-17 through 2018-19.

Explainable AI (XAI) is concerned with how to make AI models more understandable to people. To date these explanations have predominantly been technocentric - mechanistic or productivity oriented. This paper introduces the Explainable AI for the Arts (XAIxArts) manifesto to provoke new ways of thinking about explainability and AI beyond technocentric discourses. Manifestos offer a means to communicate ideas, amplify unheard voices, and foster reflection on practice. To supports the co-creation and revision of the XAIxArts manifesto we combine a World Caf\'e style discussion format with a living manifesto to question four core themes: 1) Empowerment, Inclusion, and Fairness; 2) Valuing Artistic Practice; 3) Hacking and Glitches; and 4) Openness. Through our interactive living manifesto experience we invite participants to actively engage in shaping this XIAxArts vision within the CHI community and beyond., Comment: Author version of paper in: Extended Abstracts of the CHI Conference on Human Factors in Computing Systems, April 26-May 1, 2025, Yokohama, Japan DOI 10.1145/3706599.3716227 ISBN 979-8-4007-1395-8/25/04

The deployment of Large Language Models (LLMs) in customer support is constrained by hallucination-generating false information-and the high cost of proprietary models. To address these challenges, we propose a retrieval-augmented question-answering (QA) pipeline and explore how to balance human input and automation. Using a dataset of questions about a Samsung Smart TV user manual, we demonstrate that synthetic data generated by LLMs outperforms crowdsourced data in reducing hallucination in finetuned models. We also compare self-training (fine-tuning models on their own outputs) and knowledge distillation (fine-tuning on stronger models' outputs, e.g., GPT-4o), and find that self-training achieves comparable hallucination reduction. We conjecture that this surprising finding can be attributed to increased exposure bias issues in the knowledge distillation case and support this conjecture with post hoc analysis. We also improve robustness to unanswerable questions and retrieval failures with contextualized "I don't know" responses. These findings show that scalable, cost-efficient QA systems can be built using synthetic data and self-training with open-source models, reducing reliance on proprietary tools or costly human annotations.

We generalize the enhanced power graph by replacing elements with classes under automorphisms. We show that the connectivity and diameter of this graph is similar to that of the enhanced power graph. We consider the universal vertices of this graph and when this graph is a complete graph. Finally, we classify when this graph is the empty graph.

Under certain hypotheses, we prove a loop space decomposition for simply-connected Poincar\'e Duality complexes of dimension $n$ whose $(n-1)$-skeleton is a co-$H$-space. This unifies many known decompositions obtained in different contexts and establishes many new families of examples. As consequences, we show that such a looped Poincar\'{e} Duality complex retracts off the loops of its $(n-1)$-skeleton and describe its homology as a one-relator algebra., Comment: 33 pages

We report on Green Bank Telescope observations of the radio magnetar Swift J1818.0--1607 between 820 MHz and 35 GHz, taken from six to nine months after its 2020 March outburst. We obtained multi-hour observations at six frequencies, recording polarimetric, spectral, and single-pulse information. The spectrum peaks at a frequency of $5.4 \pm 0.6$ GHz, making Swift J1818.0--1607 one of many radio magnetars which exhibit a gigahertz-peaked spectrum (GPS). The radio flux decays steeply above the peak frequency, with in-band spectral indices $\alpha < -2.3 $ above 9 GHz. The emission is highly ($> 50\%$) linearly polarized, with a lower degree ($< 30\%$) of circular polarization which can change handedness between single pulses. Across the frequency range of our observations, the time-integrated radio profiles share a common shape: a narrow ``pulsar-like'' central component flanked by ``magnetar-like'' components comprised of bright, spiky subpulses. The outer profile components exhibit larger degrees of flux modulation and flatter spectral indices when compared to the central pulse component., Comment: 20 pages, 16 figures. Submitted to ApJ

The rapid development of advanced AI agents and the imminent deployment of many instances of these agents will give rise to multi-agent systems of unprecedented complexity. These systems pose novel and under-explored risks. In this report, we provide a structured taxonomy of these risks by identifying three key failure modes (miscoordination, conflict, and collusion) based on agents' incentives, as well as seven key risk factors (information asymmetries, network effects, selection pressures, destabilising dynamics, commitment problems, emergent agency, and multi-agent security) that can underpin them. We highlight several important instances of each risk, as well as promising directions to help mitigate them. By anchoring our analysis in a range of real-world examples and experimental evidence, we illustrate the distinct challenges posed by multi-agent systems and their implications for the safety, governance, and ethics of advanced AI., Comment: Cooperative AI Foundation, Technical Report #1

The shape of the atomic nucleus is a property which underpins our understanding of nuclear systems, impacts the limits of nuclear existence, and enables probes of physics beyond the Standard Model. Nuclei can adopt a variety of shapes, including spheres, axially deformed spheroids, and pear shapes. In some regions of the nuclear chart where a spherical nucleus would naively be expected, deformed nuclear states can result from collective action of constituent protons and neutrons. In a small subset of nuclei both spherical and deformed nuclear states have been experimentally observed, a phenomenon termed shape coexistence. We present spectroscopic evidence for the coexistence of $J^{\pi}=1+$ spherical and deformed states in $^{70}$Co, separated by less than 275~keV. This close degeneracy of levels with the same $J^{\pi}$ and different shapes demonstrates an extreme example of shape coexistence resulting from the interplay of independent particle motion and collective behavior in highly unstable nuclear systems and identifies the Co isotopes as a transition point between deformed ground states observed in the Cr isotopes and spherical configurations observed in the closed-shell Ni isotopes.

Nonlocal entanglement between pair-correlated particles is a highly counter-intuitive aspect of quantum mechanics, where measurement on one particle can instantly affect the other, regardless of distance. While the rigorous Bell's inequality framework has enabled the demonstration of such entanglement in photons and atomic internal states, no experiment has yet involved motional states of massive particles. Here we report the experimental observation of Bell correlations in motional states of momentum-entangled ultracold helium atoms. Momentum-entangled pairs are generated via $s$-wave collisions. Using a Rarity-Tapster interferometer and a Bell-test framework, we observe atom-atom correlations required for violation of a Bell inequality. This result shows the potential of ultracold atoms for fundamental tests of quantum mechanics and opens new avenues to studying gravitational effects in quantum states.

How we should design and interact with social artificial intelligence depends on the socio-relational role the AI is meant to emulate or occupy. In human society, relationships such as teacher-student, parent-child, neighbors, siblings, or employer-employee are governed by specific norms that prescribe or proscribe cooperative functions including hierarchy, care, transaction, and mating. These norms shape our judgments of what is appropriate for each partner. For example, workplace norms may allow a boss to give orders to an employee, but not vice versa, reflecting hierarchical and transactional expectations. As AI agents and chatbots powered by large language models are increasingly designed to serve roles analogous to human positions - such as assistant, mental health provider, tutor, or romantic partner - it is imperative to examine whether and how human relational norms should extend to human-AI interactions. Our analysis explores how differences between AI systems and humans, such as the absence of conscious experience and immunity to fatigue, may affect an AI's capacity to fulfill relationship-specific functions and adhere to corresponding norms. This analysis, which is a collaborative effort by philosophers, psychologists, relationship scientists, ethicists, legal experts, and AI researchers, carries important implications for AI systems design, user behavior, and regulation. While we accept that AI systems can offer significant benefits such as increased availability and consistency in certain socio-relational roles, they also risk fostering unhealthy dependencies or unrealistic expectations that could spill over into human-human relationships. We propose that understanding and thoughtfully shaping (or implementing) suitable human-AI relational norms will be crucial for ensuring that human-AI interactions are ethical, trustworthy, and favorable to human well-being., Comment: 76 pages, 2 figures

Compositionally graded alloys (CGAs) are often proposed for use in structural components where the combination of two or more alloys within a single part can yield substantial enhancement in performance and functionality. For these applications, numerous design methodologies have been developed, one of the most sophisticated being the application of path planning algorithms originally designed for robotics to solve CGA design problems. In addition to the traditional application to structural components, this work proposes and demonstrates the application of this CGA design framework to rapid alloy design, synthesis, and characterization. A composition gradient in the CoCrFeNi alloy space was planned between the maximum and minimum stacking fault energy (SFE) as predicted by a previously developed model in a face-centered cubic (FCC) high entropy alloy (HEA) space. The path was designed to be monotonic in SFE and avoid regions that did not meet FCC phase fraction and solidification range constraints predicted by CALculation of PHAse Diagrams (CALPHAD). Compositions from the path were selected to produce a linear gradient in SFE, and the CGA was built using laser directed energy deposition (L-DED). The resulting gradient was characterized for microstructure and mechanical properties, including hardness, elastic modulus, and strain rate sensitivity. Despite being predicted to contain a single FCC phase throughout the gradient, part of the CGA underwent a martensitic transformation, thereby demonstrating a limitation of using equilibrium CALPHAD calculations for phase stability predictions. More broadly, this demonstrates the ability of the methods employed to bring attention to blind spots in alloy models.

Using data samples of 102 million Upsilon(1S) events and 158 million Upsilon(2S) events collected by the Belle detector at the KEKB asymmetric-energy $e^+e^-$ collider, we search for [udsccbar] pentaquark states decaying to Jpsi Lambda. Using the first observations of Upsilon(1S, 2S) inclusive decays to Jpsi Lambda, we find evidence of the P_ccbars(4459)0 state with a significance of 3.3 standard deviations, including statistical and systematic uncertainties. We measure the mass and width of the Pccbars(4459)0 to be (4471.7 +- 4.8 +- 0.6) MeV/c2 and (21.9 +- 13.1 +- 2.7) MeV, respectively. The branching fractions for P_ccbars(4459)0 production are measured to be B[Upsilon(1S) -> P_ccbars(4459)0/ Pbar_ccbars(4459)0 + anything] = (3.5 +- 2.0 +- 0.2)*10-6 and B[Upsilin(2S) -> P_ccbars(4459)0/ Pbar_ccbars(4459)0 +anything] = (2.9 +- 1.7 +- 0.4)*10-6. The inclusive branching fractions of Upsilon(1S, 2S) -> Jpsi Lambda/Lambdabar are measured to be B[Upsilin(1S) -> Jpsi Lambda/Lambdabar + anything] = (36.9 +- 5.3 +- 2.4)*10-6 and B[Upsilon(2S) -> Jpsi Lambda/Lambdabar + anything] = (22.3 +- 5.7 +- 3.1)*10-6. We measure the visible cross section $\sigma(e^+e^- \to J/psi \Lambda/\bar\Lambda$ + anything) = (90 +- 14 +- 6) fb for the continuum production at $\sqrt{s} = 10.52$ GeV. In all cases, the first uncertainties are statistical and the second are systematic., Comment: 8 pages, 3 figures

Flood risk managers seek to optimise Blue-Green Infrastructure (BGI) designs to maximise return on investment. Current systems often use optimisation algorithms and detailed flood models to maximise benefit-cost ratios for single rainstorm return periods. However, these schemes may lack robustness in mitigating flood risks across different storm magnitudes. For example, a BGI scheme optimised for a 100-year return period may differ from one optimised for a 10-year return period. This study introduces a novel methodology incorporating five return periods (T = 10, 20, 30, 50, and 100 years) into a multi-objective BGI optimisation framework. The framework combines a Non-dominated Sorting Genetic Algorithm II (NSGA-II) with a fully distributed hydrodynamic model to optimise the spatial placement and combined size of BGI features. For the first time, direct damage cost (DDC) and expected annual damage (EAD), calculated for various building types, are used as risk objective functions, transforming a many-objective problem into a multi-objective one. Performance metrics such as Median Risk Difference (MedRD), Maximum Risk Difference (MaxRD), and Area Under Pareto Front (AUPF) reveal that a 100-year optimised BGI design performs poorly when evaluated for other return periods, particularly shorter ones. In contrast, a BGI design optimised using composite return periods enhances performance metrics across all return periods, with the greatest improvements observed in MedRD (22%) and AUPF (73%) for the 20-year return period, and MaxRD (23%) for the 50-year return period. Furthermore, climate uplift stress testing confirms the robustness of the proposed design to future rainfall extremes. This study advocates a paradigm shift in flood risk management, moving from single maximum to multiple rainstorm return period-based designs to enhance resilience and adaptability to future climate extremes., Comment: Preprint submitted to Wiley in February 2025. It has 38 pages and 15 figures

The Pandora SmallSat is a NASA flight project aimed at studying the atmospheres of exoplanets -- planets orbiting stars outside our Solar System. Pandora will provide the first dataset of simultaneous, multiband (visible and NIR), long-baseline observations of exoplanets and their host stars. Pandora is an ambitious project that will fly a 0.44 m telescope in a small form factor. To achieve the scientific goals, the mission requires a departure from the traditional cost-schedule paradigm of half-meter-class observatories. Pandora achieves this by leveraging existing capabilities that necessitate minimal engineering development, disruptive and agile management, trusted partnerships with vendors, and strong support from the lead institutions. The Pandora team has developed a suite of high-fidelity parameterized simulation and modeling tools to estimate the performance of both imaging channels. This has enabled a unique bottom-up approach to deriving trades and system requirements. Pandora is a partnership between NASA and Lawrence Livermore National Laboratory. The project completed its Critical Design Review in October 2023 and is slated for launch into Sun-synchronous, low-Earth orbit in Fall 2025., Comment: Paper accepted to the IEEE Aerospace Conference 2025

Recent research in consensus has often focussed on protocols for State-Machine-Replication (SMR) that can handle high throughputs. Such state-of-the-art protocols (generally DAG-based) induce undue overhead when the needed throughput is low, or else exhibit unnecessarily-poor latency and communication complexity during periods of low throughput. Here we present Morpheus Consensus, which naturally morphs from a quiescent low-throughput leaderless blockchain protocol to a high-throughput leader-based DAG protocol and back, excelling in latency and complexity in both settings. During high-throughout, Morpheus pars with state-of-the-art DAG-based protocols, including Autobahn. During low-throughput, Morpheus exhibits competitive complexity and lower latency than standard protocols such as PBFT and Tendermint, which in turn do not perform well during high-throughput. The key idea of Morpheus is that as long as blocks do not conflict (due to Byzantine behaviour, network delays, or high-throughput simultaneous production) it produces a forkless blockchain, promptly finalizing each block upon arrival. It assigns a leader only if one is needed to resolve conflicts, in a manner and with performance not unlike Autobahn.

We investigate the properties of momentum-dark excitons and trions formed in two-dimensional (2D) materials that exhibit an inverted Mexican-hat-shaped dispersion relation, taking monolayer InSe as an example. We employ variational techniques to obtain the momentum-dark and bright ground-states (non-zero and zero quasiparticle momenta, respectively). These states are of particular interest due to their peaks in the quasiparticle density of states, the largest contribution comes from the momentum-dark ground state due to the presence of a van Hove singularity (VHS). These momentum-dark systems require a physical process to provide the necessary momentum to become bright. We study the brightening of this state due to coupling with phonons and compute the resulting photoluminescence spectrum. This work opens new avenues of research, such as exploiting dark excitons in solar cells and other semiconductor-based optoelectronic devices., Comment: 28 pages, 9 figures

We present 18 pulsar discoveries from the AO327 pulsar survey, along with their timing solutions and those for an additional 31 AO327-discovered pulsars. Timing solutions were constructed using observations from a follow-up timing campaign taken between the periods of 2013 -- 2019 using the Arecibo Observatory's 327-MHz receiver. Aside from PSR J0916+0658, an isolated pulsar that shows evidence for partial recycling, the remaining discoveries are non-recycled pulsars. We present a brief census of emission features for all pulsars with the following standouts. PSR~J1942+0142 is found to exhibit the very rare phenomenon of subpulse bi-drifting and PSR~J0225+1727 has an interpulse. We also report distance estimates using the NE2001 and YMW16 Galactic electron density models, and identify at least 10 sources where either one or both models underestimate the maximum Galactic line of sight dispersion measure., Comment: 30 pages, 19 figures, submitted to ApJ

Linear stationary reaction-convection-diffusion equations with Dirichlet boundary conditions are approximated using a simple finite difference method corresponding to central differences and the addition of a high-order stabilization term called a numerical moment. The focus is on convection-dominated equations, and the formulation for the method is motivated by various results for fully nonlinear problems. The method features higher-order local truncation errors than monotone methods consistent with the use of the central difference approximation for the gradient. Stability and rates of convergence are derived in the $\ell^2$ norm for the constant-coefficient case. Numerical tests are provided to compare the new methods to monotone methods. The methods are also tested for stationary Hamilton-Jacobi equations where they demonstrate higher rates of convergence than the Lax-Friedrich's method when the underlying viscosity solution is smooth and comparable performance when the underlying viscosity solution is not smooth.

The estimation of uncertainties associated with predictions from quantitative structure-activity relationship (QSAR) models can accelerate the drug discovery process by identifying promising experiments and allowing an efficient allocation of resources. Several computational tools exist that estimate the predictive uncertainty in machine learning models. However, deviations from the i.i.d. setting have been shown to impair the performance of these uncertainty quantification methods. We use a real-world pharmaceutical dataset to address the pressing need for a comprehensive, large-scale evaluation of uncertainty estimation methods in the context of realistic distribution shifts over time. We investigate the performance of several uncertainty estimation methods, including ensemble-based and Bayesian approaches. Furthermore, we use this real-world setting to systematically assess the distribution shifts in label and descriptor space and their impact on the capability of the uncertainty estimation methods. Our study reveals significant shifts over time in both label and descriptor space and a clear connection between the magnitude of the shift and the nature of the assay. Moreover, we show that pronounced distribution shifts impair the performance of popular uncertainty estimation methods used in QSAR models. This work highlights the challenges of identifying uncertainty quantification methods that remain reliable under distribution shifts introduced by real-world data.

Let $D: \mathbb{Z}_m^n \to \mathbb{Z}_m^n$ be defined so that \[D(x_1, x_2, ..., x_n)=(x_1+x_2 \; \text{mod} \; m, x_2+x_3 \; \text{mod} \; m, ..., x_n+x_1 \; \text{mod} \; m).\] We call $D$ the Ducci function and the sequence $\{D^{\alpha}(\mathbf{u})\}_{\alpha=0}^{\infty}$ the Ducci sequence of $\mathbf{u}$ for $\mathbf{u} \in \mathbb{Z}_m^n$. Every Ducci sequence enters a cycle, so we can let $\text{Per}(\mathbf{u})$ be the number of tuples in the Ducci cycle of $\mathbf{u}$, or the period of $\mathbf{u}$. In this paper, we will look at what different possible values of $\text{Per}(\mathbf{u})$ we can have and some conditions that if $\mathbf{u}$ meets at least one of them, $\mathbf{u}$ will generate a period smaller than the maximum period.

Let $D$ be an endomorphism on $\mathbb{Z}_m^n$ so that \[D(x_1, x_2, ..., x_n)=(x_1+x_2 \; \text{mod} \; m, x_2+x_3 \; \text{mod} \; m, ..., x_n+x_1 \; \text{mod} \; m).\] We call the sequence $\{D^{\alpha}(\mathbf{u})\}_{\alpha=0}^{\infty}$ the Ducci sequence of $\mathbf{u} \in \mathbb{Z}_m^n$, which always enters a cycle. Now let $H$ be an endomorphism on $\mathbb{Z}_m^n$ such that \[H(x_1, x_2, ..., x_n)=(x_2, x_3, ..., x_n, x_1).\] In this paper, we will talk about a few cases when $\mathbf{u}$ and $H^{\beta}(\mathbf{u})$ have the same Ducci cycle for $\beta > 0$, as well as prove a few cases of $n,m$ where this is guaranteed for every $\mathbf{u} \in \mathbb{Z}_m^n$.

The nilpotent graph of a group $G$ is the simple and undirected graph whose vertices are the elements of $G$ and two distinct vertices are adjacent if they generate a nilpotent subgroup of $G$. Here we discuss some topological properties of the nilpotent graph of a finite group $G$. Indeed, we characterize finite solvable groups whose closed neighborhoods are nilpotent subgroups. Moreover, we study the connectivity of the graph $\Gamma(G)$ obtained removing all universal vertices from the nilpotent graph of $G$. Some upper bounds to the diameter of $\Gamma(G)$ are provided when $G$ belongs to some classes of groups., Comment: Added a reference to the paper by Burness, Lucchini, and Nemmi

A new class of non-monotone finite difference (FD) approximation methods for approximating solutions to non-degenerate stationary Hamilton-Jacobi problems with Dirichlet boundary conditions is proposed and analyzed. The new FD methods add a high order correction to the Lax-Friedrich's method while utilizing a novel cutoff to preserve the convergence properties of the Lax-Friedrich's approximation. Since monotone methods are limited to first order accuracy by the Godunov barrier, the proposed approach provides a template for boosting the accuracy of a monotone method using a modified numerical moment stabilizer with a high-order auxiliary boundary condition. Numerical tests are provided to test the utility of the approach while a novel admissibility and stability analysis technique lays a foundation for analyzing non-monotone methods.

Astrophysical observations of the cosmos allow us to probe extreme physics and answer foundational questions on our universe. Modern astronomy is increasingly operating under a holistic approach, probing the same question with multiple diagnostics including how sources vary over time, how they appear across the electromagnetic spectrum, and through their other signatures, including gravitational waves, neutrinos, cosmic rays, and dust on Earth. Astrophysical observations are now reaching the point where approximate physics models are insufficient. Key sources of interest are explosive transients, whose understanding requires multidisciplinary studies at the intersection of astrophysics, gravity, nuclear science, plasma physics, fluid dynamics and turbulence, computation, particle physics, atomic, molecular, and optical science, condensed matter and materials science, radiation transport, and high energy density physics. This white paper provides an overview of the major scientific advances that lay at the intersection of physics and astronomy and are best probed through time-domain and multimessenger astrophysics, an exploration of how multidisciplinary science can be fostered, and introductory descriptions of the relevant scientific disciplines and key astrophysical sources of interest., Comment: This white paper is the product of the Third Time-Domain And Multimessenger Astrophysics workshop: Multidisciplinary Science in the Multimessenger Era, hosted by Louisiana State University with additional support from DOE, NSF, and NASA

Lateral Spine Image (LSI) analysis is important for medical diagnosis, treatment planning, and detailed spinal health assessments. Although modalities like Computed Tomography and Digital X-ray Imaging are commonly used, Dual Energy X-ray Absorptiometry (DXA) is often preferred due to lower radiation exposure, seamless capture, and cost-effectiveness. Accurate Vertebral Landmark Localization (VLL) on LSIs is important to detect spinal conditions like kyphosis and lordosis, as well as assessing Abdominal Aortic Calcification (AAC) using Inter-Vertebral Guides (IVGs). Nonetheless, few automated VLL methodologies have concentrated on DXA LSIs. We present VerteNet, a hybrid CNN-Transformer model featuring a novel dual-resolution attention mechanism in self and cross-attention domains, referred to as Dual Resolution Self-Attention (DRSA) and Dual Resolution Cross-Attention (DRCA). These mechanisms capture the diverse frequencies in DXA images by operating at two different feature map resolutions. Additionally, we design a Multi-Context Feature Fusion Block (MCFB) that efficiently integrates the features using DRSA and DRCA. We train VerteNet on 620 DXA LSIs from various machines and achieve superior results compared to existing methods. We also design an algorithm that utilizes VerteNet's predictions in estimating the Region of Interest (ROI) to detect potential abdominal aorta cropping, where inadequate soft tissue hinders calcification assessment. Additionally, we present a small proof-of-concept study to show that IVGs generated from VLL information can improve inter-reader correlation in AAC scoring, addressing two key areas of disagreement in expert AAC-24 scoring: IVG placement and quality control for full abdominal aorta assessment. The code for this work can be found at https://github.com/zaidilyas89/VerteNet., Comment: 10 pages with 7 figures

While large language models have facilitated breakthroughs in many applications of artificial intelligence, their inherent largeness makes them computationally expensive and challenging to deploy in resource-constrained settings. In this paper, we document the development of SmolLM2, a state-of-the-art "small" (1.7 billion parameter) language model (LM). To attain strong performance, we overtrain SmolLM2 on ~11 trillion tokens of data using a multi-stage training process that mixes web text with specialized math, code, and instruction-following data. We additionally introduce new specialized datasets (FineMath, Stack-Edu, and SmolTalk) at stages where we found existing datasets to be problematically small or low-quality. To inform our design decisions, we perform both small-scale ablations as well as a manual refinement process that updates the dataset mixing rates at each stage based on the performance at the previous stage. Ultimately, we demonstrate that SmolLM2 outperforms other recent small LMs including Qwen2.5-1.5B and Llama3.2-1B. To facilitate future research on LM development as well as applications of small LMs, we release both SmolLM2 as well as all of the datasets we prepared in the course of this project.

In this paper, we consider semi-extraspecial $p$-groups $G$ that have an automorphism of order $|G:G'| - 1$. We prove that these groups are isomorphic to Sylow $p$-subgroups of ${\rm SU}_3 (p^{2a})$ for some integer $a$. If $p$ is odd, this is equivalent to saying that $G$ is isomorphic to a Sylow $p$-subgroup of ${\rm SL}_3 (p^a)$.

Josephson junctions are the principal circuit element in numerous superconducting quantum information devices and can be readily integrated into large-scale electronics. However, device integration at the wafer scale necessarily depends on having a reliable, high-fidelity, and high-yield fabrication method for creating Josephson junctions. When creating Al/AlOx based superconducting qubits, the standard Josephson junction fabrication method relies on a sub-micron suspended resist bridge, known as a Dolan bridge, which tends to be particularly fragile and can often times fracture during the resist development process, ultimately resulting in device failure. In this work, we demonstrate a unique Josephson junction lithography mask design that incorporates stress-relief channels. Our simulation results show that the addition of stress-relief channels reduces the lateral stress in the Dolan bridge by more than 70% for all the bridge geometries investigated. In practice, our novel mask design significantly increased the survivability of the bridge during device processing, resulting in 100% yield for over 100 Josephson junctions fabricated.

We present Branch-Train-Stitch (BTS), an efficient and flexible training algorithm for combining independently trained large language model (LLM) experts into a single, capable generalist model. Following Li et al., we start with a single seed language model which is branched into domain-specific (e.g., coding or math) experts with continual pretraining. BTS combines experts into a generalist model using lightweight stitch layers, which are inserted between frozen experts and the seed LLM, and trained on a small datamix of the expert domains. Stitch layers enable the seed LLM to integrate representations from any number of experts during the forward pass, allowing it to generalize to new domains, despite remaining frozen. Because BTS does not alter the constituent LLMs, BTS provides a modular and flexible approach: experts can be easily removed and new experts can be added with only a small amount of training. Compared to alternative model merging approaches, BTS yields the best generalist performance on a variety of downstream tasks, retaining the specialized capabilities of each of the experts.

Cepheid circumstellar emissions have previously been detected via both infrared excess and infrared interferometric observations at a few stellar radii. Those studies have shown that these circumstellar emission can be produced by ionized gas, however there is no direct observational evidence to confirm this hypothesis. In this letter we explore the continuum emission and a spectrum of the bright and long-period Cepheid, $\ell$~Car ($P=35.56\,$day) at millimeter-wavelengths in order to detect possible effects of ionized gas emission. We presented ALMA observations of $\ell$ Car in two spectral setups in Band~6 (near 212 and 253\,GHz, respectively) and we compared the measured flux density to one expected for the stellar continuum. We also derived the spectral index and probed the presence of Radio Recombination Lines (RRL). We report statistically significant emission of about 3.5$\,$mJy in the two spectral ranges, which is about 2.5 times the stellar continuum emission. For the first time, we are also able to derive the spectral index of the flux density ($S_\nu \propto \nu^\alpha$), $\alpha=+1.26\pm$0.44 ($\sim$3$\sigma$ error), which is characteristic of partially optically thick ionized gas emission. Additionally, we discovered an emission line from a RRL of hydrogen H29$\alpha$ centered on the stellar rest velocity, smaller in spatial extent than about 0\farcs2 ($\lesssim 100\,$AU), with a symmetric profile with a width at half power of 55.3$\pm$7.5\,\kms (1$\sigma$ error). It confirms the presence of ionized gas emission near $\ell$~Car. The millimeter emission detected from $\ell$ Car can be attributed to ionized gas emission from the Cepheid's chromosphere. Further radio interferometric observations are necessary to confirm the occurrence of these ionized gas envelopes around Cepheids of different pulsation periods., Comment: Accepted for publication in Astronomy and Astrophysics. 7 pages, 5 figures

Improving the safety and reliability of large language models (LLMs) is a crucial aspect of realizing trustworthy AI systems. Although alignment methods aim to suppress harmful content generation, LLMs are often still vulnerable to jailbreaking attacks that employ adversarial inputs that subvert alignment and induce harmful outputs. We propose the Randomized Embedding Smoothing and Token Aggregation (RESTA) defense, which adds random noise to the embedding vectors and performs aggregation during the generation of each output token, with the aim of better preserving semantic information. Our experiments demonstrate that our approach achieves superior robustness versus utility tradeoffs compared to the baseline defenses., Comment: Presented in the Safe Generative AI Workshop at NeurIPS 2024

Snowman is the consensus protocol run by blockchains on Avalanche. Recent work established a rigorous proof of probabilistic consistency for Snowman in the \emph{synchronous} setting, under the simplifying assumption that correct processes execute sampling rounds in `lockstep'. In this paper, we describe a modification of the protocol that ensures consistency in the \emph{partially synchronous} setting, and when correct processes carry out successive sampling rounds at their own speed, with the time between sampling rounds determined by local message delays., Comment: arXiv admin note: substantial text overlap with arXiv:2404.14250

The UK has a long-established reputation for excellence in research across a broad range of fields, but in recent years, there has been greater emphasis on STEM investment and greater recognition of the UK's success in STEM. This paper examines the relative strengths of SHAPE disciplines and demonstrates that the UK's SHAPE research portfolio outperforms the UK's STEM research, for each international benchmark considered in this work. It is argued that SHAPE research is becoming increasingly important as a partner to STEM as the widespread use of technology creates societal challenges. It is also argued that the strength of UK SHAPE is the basis of a strategic advantage for UK research., Comment: 13 pages, 9 figures

Artificial intelligence is reshaping creative domains, yet its co-creative processes, especially in group settings with novice users, remain under explored. To bridge this gap, we conducted a case study in a college-level course where nine undergraduate students were tasked with creating three original music tracks using AI tools over 10 weeks. The study spanned the entire creative journey from ideation to releasing these songs on Spotify. Participants leveraged AI for music and lyric production, cover art, and distribution. Our findings highlight how AI transforms creative workflows: accelerating ideation but compressing the traditional preparation stage, and requiring novices to navigate a challenging idea selection and validation phase. We also identified a new "collaging and refinement" stage, where participants creatively combined diverse AI-generated outputs into cohesive works. Furthermore, AI influenced group social dynamics and role division among human creators. Based on these insights, we propose the Human-AI Co-Creation Stage Model and the Human-AI Agency Model, offering new perspectives on collaborative co-creation with AI.

Until now multiscale quantum problems have appeared to be out of reach at the many-body level relevant to strongly correlated materials and current quantum information devices. In fact, they can be modeled with $q$-th order fractional derivatives, as we demonstrate in this work, treating classical and quantum phase transitions in a fractional Ising model for $0 < q \leq 2$ ($q = 2$ is the usual Ising model). We show that fractional derivatives not only enable continuous tuning of critical exponents such as $\nu$, $\delta$, and $\eta$, but also define the Hausdorff dimension $H_D$ of the system tied geometrically to the anomalous dimension $\eta$. We discover that for classical systems, $H_D$ is precisely equal to the fractional order $q$. In contrast, for quantum systems, $H_D$ deviates from this direct equivalence, scaling more gradually, driven by additional degrees of freedom introduced by quantum fluctuations. These results reveal how fractional derivatives fundamentally modify the fractal geometry of many-body interactions, directly influencing the universal symmetries of the system and overcoming traditional dimensional restrictions on phase transitions. Specifically, we find that for $q < 1$ in the classical regime and $q < 2$ in the quantum regime, fractional interactions allow phase transitions in one dimension. This work establishes fractional derivatives as a powerful tool for engineering critical behavior, offering new insights into the geometry of multiscale systems and opening avenues for exploring tunable quantum materials on NISQ devices.

The KOTO II experiment is proposed to measure the branching ratio of the decay $K_L\to\pi^0\nu\bar{\nu}$ at J-PARC. With a beamline to extract long-lived neutral kaons at 5 degrees from a production target, the single event sensitivity of the decay is $8.5\times 10^{-13}$, which is much smaller than the Standard Model prediction $3\times 10^{-11}$. This allows searches for new physics beyond the Standard Model and the first discovery of the decay with a significance exceeding $5\sigma$. As the only experiment proposed in the world dedicated to rare kaon decays, KOTO II will be indispensable in the quest for a complete understanding of flavor dynamics in the quark sector. Moreover, by combining efforts from the kaon community worldwide, we plan to develop the KOTO II detector further and expand the physics reach of the experiment to include measurements of the branching ratio of the $K_L\to\pi^0\ell^+\ell^-$ decays, studies of other $K_L$ decays, and searches for dark photons, axions, and axion-like particles. KOTO II will therefore obtain a comprehensive understanding of $K_L$ decays, providing further constraints on new physics scenarios with existing $K^+$ results., Comment: Submitted to the J-PARC PAC. arXiv admin note: substantial text overlap with arXiv:2110.04462

From $>1000$ orbits of HST imaging, we present deep homogeneous resolved star color-magnitude diagrams that reach the oldest main sequence turnoff and uniformly measured star formation histories (SFHs) of 36 dwarf galaxies ($-6 \ge M_V \ge -17$) associated with the M31 halo, and for 10 additional fields in M31, M33, and the Giant Stellar Stream. From our SFHs we find: i) the median stellar age and quenching epoch of M31 satellites correlate with galaxy luminosity and galactocentric distance. Satellite luminosity and present-day distance from M31 predict the satellite quenching epoch to within $1.8$ Gyr at all epochs. This tight relationship highlights the fundamental connection between satellite halo mass, environmental history, and star formation duration. ii) There is no difference between the median SFH of galaxies on and off the great plane of Andromeda satellites. iii) $\sim50$\% of our M31 satellites show prominent ancient star formation ($>12$ Gyr ago) followed by delayed quenching ($8-10$ Gyr ago), which is not commonly observed among the MW satellites. iv) A comparison with TNG50 and FIRE-2 simulated satellite dwarfs around M31-like hosts show that some of these trends (dependence of SFH on satellite luminosity) are reproduced in the simulations while others (dependence of SFH on galactocentric distance, presence of the delayed-quenching population) are weaker or absent. We provide all photometric catalogs and SFHs as High-Level Science Products on MAST., Comment: Accepted for publication on ApJ. 47 pages, 24 figures, 12 tables. Corresponding HLSP data can be retrieved at: https://archive.stsci.edu/hlsp/m31-satellites

Payment Channel Networks (PCNs) are a method for improving the scaling and latency of cryptocurrency transactions. For a payment to be made between two peers in a PCN, a feasible low-fee path in the network must be planned. Many PCN path planning algorithms use a search algorithm that is a variant of Dijkstra's algorithm. In this article, we prove the correctness and computational complexity of this algorithm. Specifically, we show that, if the PCN satisfies a consistency property relating to the fees charged by payment channels, the algorithm is correct and has polynomial computational complexity. However, in the general case, the algorithm is not correct and the path planning problem is NP-hard. These newly developed results can be used to inform the development of new or existing PCNs amenable to path planning. For example, we show that the Lightning Network, which is the most widely used PCN and is built on the Bitcoin cryptocurrency, currently satisfies the above consistency property. As a second contribution, we demonstrate that a small modification to the above path planning algorithm which, although having the same asymptotic computational complexity, empirically shows better performance. This modification involves the use of a bidirectional search and is empirically evaluated by simulating transactions on the Lightning Network.

Large Language Models improve with increasing amounts of high-quality training data. However, leveraging larger datasets requires balancing quality, quantity, and diversity across sources. After evaluating nine baseline methods under both compute- and data-constrained scenarios, we find token-count heuristics outperform manual and learned mixes, indicating that simple approaches accounting for dataset size and diversity are surprisingly effective. Building on this insight, we propose two complementary approaches: UtiliMax, which extends token-based heuristics by incorporating utility estimates from reduced-scale ablations, achieving up to a 10.6x speedup over manual baselines; and Model Estimated Data Utility (MEDU), which leverages LLMs to estimate data utility from small samples, matching ablation-based performance while reducing computational requirements by $\sim$200x. Together, these approaches establish a new framework for automated, compute-efficient data mixing that is robust across training regimes., Comment: 10 pages, 8 figures

We explore the evolving relationship between galaxies and their dark matter halos from $z \sim 0.1$ to $z \sim 12$ using mass-limited angular clustering measurements in the 0.54 deg$^2$ of the COSMOS-Web survey. This study provides the first measurements of the mass-limited two-point correlation function at $z \ge 10$ and a consistent analysis spanning 13.4 Gyr of cosmic history, setting new benchmarks for future simulations and models. Using a halo occupation distribution (HOD) framework, we derive characteristic halo masses and the stellar-to-halo mass relationship (SHMR) across redshifts and stellar mass bins. Our results first indicate that HOD models fit data at $z \ge 2.5$ best when incorporating a non-linear scale-dependent halo bias, boosting clustering at non-linear scales (r = 10-100 kpc). We find that galaxies at z > 10.5 with $\log(M_\star / M_\odot) \ge 8.85$ are hosted by halos with $M_{\rm h} \sim 10^{10.5}\,M_\odot$, achieving a star formation efficiency (SFE) $M_\star / (f_b M_{\rm h}) $ up to 1 dex higher than at $z \le 1$. The high galaxy bias at $z \ge 8$ suggests that these galaxies reside in massive halos with intrinsic high SFE. Our SHMR evolves significantly with redshift, starting high at $z \ge 10.5$, decreasing until $z \sim 2 - 3$, then increasing again until the present. Current simulations fail to reproduce both massive high-$z$ galaxies and this evolution, while semi-empirical models linking SFE to halo mass, accretion rates, and redshift align with our findings. We propose that $z > 8$ galaxies experience bursty star formation without significant feedback altering their growth, driving the rapid growth of massive galaxies observed by JWST. Over time, increasing feedback efficiency and exponential halo growth suppress star formation. At $z \sim 2 - 3$ and after, halo growth slows down while star formation continues, supported by gas reservoirs in halos., Comment: Submitted to Astronomy & Astrophysics

A current challenge in atomistic machine learning is that of efficiently predicting the response of the electron density under electric fields. We address this challenge with symmetry-adapted kernel functions that are specifically derived to account for the rotational symmetry of a three-dimensional vector field. We demonstrate the equivariance of the method on a set of rotated water molecules and show its high efficiency with respect to number of training configurations and features for liquid water and naphthalene crystals. We conclude showcasing applications for relaxed configurations of gold nanoparticles, reproducing the scaling law of the electronic polarizability with size, up to systems with more than 2000 atoms. By deriving a natural extension to equivariant learning models of the electron density, our method provides an accurate and inexpensive strategy to predict the electrostatic response of molecules and materials., Comment: 5 pages, 3 figures

Highly frequency-stable lasers are a ubiquitous tool for optical frequency metrology, precision interferometry, and quantum information science. While making a universally applicable laser is unrealistic, spectral noise can be tailored for specific applications. Here we report a high-power clock laser with minimized high-frequency noise, together with long-term instability of $3.5 \times 10^{-17}$ at one to thousands of seconds. The laser frequency noise is precisely characterized with atom-based spectral analysis that employs a pulse sequence designed to suppress sensitivity to intensity noise. This method provides universally applicable tunability of the spectral response and analysis of quantum sensors over a wide frequency range. With the optimized laser system characterized by this technique, we achieve an average single-qubit Clifford gate fidelity of up to $F_1^2 = 0.99964(3)$ when simultaneously driving 3000 optical qubits with a homogeneous Rabi frequency ranging from 10 Hz to $\sim$ 1 kHz. This result represents the highest single optical-qubit gate fidelity for large number of atoms.

Many AI companies are training their large language models (LLMs) on data without the permission of the copyright owners. The permissibility of doing so varies by jurisdiction: in countries like the EU and Japan, this is allowed under certain restrictions, while in the United States, the legal landscape is more ambiguous. Regardless of the legal status, concerns from creative producers have led to several high-profile copyright lawsuits, and the threat of litigation is commonly cited as a reason for the recent trend towards minimizing the information shared about training datasets by both corporate and public interest actors. This trend in limiting data information causes harm by hindering transparency, accountability, and innovation in the broader ecosystem by denying researchers, auditors, and impacted individuals access to the information needed to understand AI models. While this could be mitigated by training language models on open access and public domain data, at the time of writing, there are no such models (trained at a meaningful scale) due to the substantial technical and sociological challenges in assembling the necessary corpus. These challenges include incomplete and unreliable metadata, the cost and complexity of digitizing physical records, and the diverse set of legal and technical skills required to ensure relevance and responsibility in a quickly changing landscape. Building towards a future where AI systems can be trained on openly licensed data that is responsibly curated and governed requires collaboration across legal, technical, and policy domains, along with investments in metadata standards, digitization, and fostering a culture of openness.