Search

A new characterization of the Lovasz theta function is provided by relating it to the (weighted) walk-generating function, thus establishing a relationship between two seemingly quite distinct concepts in algebraic graph theory. An application of this new characterization is given by showing how it straightforwardly entails multiple natural generalizations of the Hoffman upper bound (on both the independence number and Lovasz number) to arbitrary non-regular graphs. Lastly, the derivation of the new characterization involves proving a certain analysis result which may in itself be of interest., Comment: 32 pages

Benchmarks are important tools for tracking the rapid advancements in large language model (LLM) capabilities. However, benchmarks are not keeping pace in difficulty: LLMs now achieve over 90\% accuracy on popular benchmarks like MMLU, limiting informed measurement of state-of-the-art LLM capabilities. In response, we introduce Humanity's Last Exam (HLE), a multi-modal benchmark at the frontier of human knowledge, designed to be the final closed-ended academic benchmark of its kind with broad subject coverage. HLE consists of 3,000 questions across dozens of subjects, including mathematics, humanities, and the natural sciences. HLE is developed globally by subject-matter experts and consists of multiple-choice and short-answer questions suitable for automated grading. Each question has a known solution that is unambiguous and easily verifiable, but cannot be quickly answered via internet retrieval. State-of-the-art LLMs demonstrate low accuracy and calibration on HLE, highlighting a significant gap between current LLM capabilities and the expert human frontier on closed-ended academic questions. To inform research and policymaking upon a clear understanding of model capabilities, we publicly release HLE at https://lastexam.ai., Comment: 25 pages, 6 figures

Current disposal facilities for coarse-grained waste perform manual sorting of materials with heavy machinery. Large quantities of recyclable materials are lost to coarse waste, so more effective sorting processes must be developed to recover them. Two key aspects to automate the sorting process are object detection with material classification in mixed piles of waste, and autonomous control of hydraulic machinery. Because most objects in those accumulations of waste are damaged or destroyed, object detection alone is not feasible in the majority of cases. To address these challenges, we propose a classification of materials with multispectral images of ultraviolet (UV), visual (VIS), near infrared (NIR), and short-wave infrared (SWIR) spectrums. Solution for autonomous control of hydraulic heavy machines for sorting of bulky waste is being investigated using cost-effective cameras and artificial intelligence-based controllers., Comment: Published in Proceedings of Sardinia 2023, 19th International Symposium on Waste Management, Resource Recovery and Sustainable Landfilling

Pre-trained Large Language Models (LLMs) encapsulate large amounts of knowledge and take enormous amounts of compute to train. We make use of this resource, together with the observation that LLMs are able to transfer knowledge and performance from one domain or even modality to another seemingly-unrelated area, to help with multivariate demand time series forecasting. Attention in transformer-based methods requires something worth attending to -- more than just samples of a time-series. We explore different methods to map multivariate input time series into the LLM token embedding space. In particular, our novel multivariate patching strategy to embed time series features into decoder-only pre-trained Transformers produces results competitive with state-of-the-art time series forecasting models. We also use recently-developed weight-based diagnostics to validate our findings.

We construct asymptotic foliations of asymtotically Schwarzschildean lightcones by surfaces of constant spacetime mean curvature (STCMC). Our construction is motivated by the approach of Huisken-Yau for the Riemannian setting in employing a geometric flow. We prove that initial data within a sufficient a-priori class converges exponentially to an STCMC surface under area preserving null mean curvature flow. Further, we show that the resulting STCMC surfaces form an asymptotic foliation that is unique within the a-priori class., Comment: 75 pages

Simulation is critical for safety evaluation in autonomous driving, particularly in capturing complex interactive behaviors. However, generating realistic and controllable traffic scenarios in long-tail situations remains a significant challenge. Existing generative models suffer from the conflicting objective between user-defined controllability and realism constraints, which is amplified in safety-critical contexts. In this work, we introduce the Causal Compositional Diffusion Model (CCDiff), a structure-guided diffusion framework to address these challenges. We first formulate the learning of controllable and realistic closed-loop simulation as a constrained optimization problem. Then, CCDiff maximizes controllability while adhering to realism by automatically identifying and injecting causal structures directly into the diffusion process, providing structured guidance to enhance both realism and controllability. Through rigorous evaluations on benchmark datasets and in a closed-loop simulator, CCDiff demonstrates substantial gains over state-of-the-art approaches in generating realistic and user-preferred trajectories. Our results show CCDiff's effectiveness in extracting and leveraging causal structures, showing improved closed-loop performance based on key metrics such as collision rate, off-road rate, FDE, and comfort.

Recent efforts recognize the power of scale in 3D learning (e.g. PTv3) and attention mechanisms (e.g. FlashAttention). However, current point cloud backbones fail to holistically unify geometric locality, attention mechanisms, and GPU architectures in one view. In this paper, we introduce Flash3D Transformer, which aligns geometric locality and GPU tiling through a principled locality mechanism based on Perfect Spatial Hashing (PSH). The common alignment with GPU tiling naturally fuses our PSH locality mechanism with FlashAttention at negligible extra cost. This mechanism affords flexible design choices throughout the backbone that result in superior downstream task results. Flash3D outperforms state-of-the-art PTv3 results on benchmark datasets, delivering a 2.25x speed increase and 2.4x memory efficiency boost. This efficiency enables scaling to wider attention scopes and larger models without additional overhead. Such scaling allows Flash3D to achieve even higher task accuracies than PTv3 under the same compute budget.

We consider cubic forms $\phi_{a,b}(x,y,z) = ax^3 + by^3 - z^3$ with coefficients $a,b \in \mathbb{Z}$. We give an asymptotic formula for how many of these forms are locally soluble everywhere, i.e. we give an asymptotic formula for the number of pairs of integers $(a, b)$ that satisfy $1 \leq a \leq A$, $1 \leq b \leq B$ and some mild conditions, such that $\phi_{a,b}$ has a non-zero solution in $\mathbb{Q}_p$ for all primes $p$., Comment: 21 pages

Human drivers rely on commonsense reasoning to navigate diverse and dynamic real-world scenarios. Existing end-to-end (E2E) autonomous driving (AD) models are typically optimized to mimic driving patterns observed in data, without capturing the underlying reasoning processes. This limitation constrains their ability to handle challenging driving scenarios. To close this gap, we propose VLM-AD, a method that leverages vision-language models (VLMs) as teachers to enhance training by providing additional supervision that incorporates unstructured reasoning information and structured action labels. Such supervision enhances the model's ability to learn richer feature representations that capture the rationale behind driving patterns. Importantly, our method does not require a VLM during inference, making it practical for real-time deployment. When integrated with state-of-the-art methods, VLM-AD achieves significant improvements in planning accuracy and reduced collision rates on the nuScenes dataset.

We present DriveGPT, a scalable behavior model for autonomous driving. We model driving as a sequential decision making task, and learn a transformer model to predict future agent states as tokens in an autoregressive fashion. We scale up our model parameters and training data by multiple orders of magnitude, enabling us to explore the scaling properties in terms of dataset size, model parameters, and compute. We evaluate DriveGPT across different scales in a planning task, through both quantitative metrics and qualitative examples including closed-loop driving in complex real-world scenarios. In a separate prediction task, DriveGPT outperforms a state-of-the-art baseline and exhibits improved performance by pretraining on a large-scale dataset, further validating the benefits of data scaling., Comment: 14 pages, 16 figures, 9 tables, and 1 video link

We present the first $\textit{NuSTAR}$ X-ray observation of EF Eri, a well-known polar system. The $\textit{NuSTAR}$ observation was conducted in conjunction with $\textit{NICER}$, shortly after EF Eri entered a high accretion state following an unprecedented period of low activity lasting 26 years since 1997. $\textit{NuSTAR}$ detected hard X-ray emission up to 50 keV with an X-ray flux of $1.2\times10^{-10}$ ergs s$^{-1}$ cm$^{-2}$ ($3\rm{-}50 keV$). Folded X-ray lightcurves exhibit a single peak with $\sim65\%$ spin modulation throughout the $3\rm{-}50$ keV band. We found no evidence of QPO signals at $\nu = 0.1\rm{-}100$ Hz with an upper limit on the QPO amplitude below $5\%$ ($90\%$ CL) at $\nu \sim 0.5$ Hz where the optical QPO was previously detected. Our 1-D accretion column model, called ${\tt MCVSPEC}$, was fitted to the $\textit{NuSTAR}$ spectral data, yielding an accurate WD mass measurement of $M = (0.55\rm{-}0.58) M_\odot$. $\texttt{MCVSPEC}$ accounts for radiative cooling by thermal bremsstrahlung and cyclotron emission, X-ray reflection off the WD surface, and a previously constrained range of the accretion column area. The derived WD mass range is in excellent agreement with the previous measurement of $M = (0.55\rm{-}0.60) M_\odot$ in the optical band. This demonstrates a combination of broadband X-ray spectral analysis and the ${\tt MCVSPEC}$ model that can be employed in our ongoing $\textit{NuSTAR}$ observation campaign of other polars to determine their WD masses accurately., Comment: 17 pages, 10 figures, 3 tables. Submitted to ApJ

This paper explores the application of physics-informed neural networks (PINNs) to tackle forward problems in 3D contact mechanics, focusing on small deformation elasticity. We utilize a mixed-variable formulation, enhanced with output transformations, to enforce Dirichlet and Neumann boundary conditions as hard constraints. The inherent inequality constraints in contact mechanics, particularly the Karush-Kuhn-Tucker (KKT) conditions, are addressed as soft constraints by integrating them into the network's loss function. To enforce the KKT conditions, we leverage the nonlinear complementarity problem (NCP) approach, specifically using the Fischer-Burmeister function, which is known for its advantageous properties in optimization. We investigate two benchmark examples of PINNs in 3D contact mechanics: a single contact patch test and the Hertzian contact problem.

We present near- and mid-infrared (IR) broadband imaging observations of the edge-on protoplanetary disk around HH 30 with the James Webb Space Telescope/Near Infrared Camera (NIRCam) and the Mid-Infrared Instrument (MIRI). We combine these observations with archival optical/near-IR scattered light images obtained with the Hubble Space Telescope (HST) and a millimeter-wavelength dust continuum image obtained with the Atacama Large Millimeter/submillimeter Array (ALMA) with the highest spatial resolution ever obtained for this target. Our multiwavelength images clearly reveal the vertical and radial segregation of micron-sized and sub-mm-sized grains in the disk. In the near- and mid-IR, the images capture not only bi-reflection nebulae separated by a dark lane but also diverse dynamical processes occurring in the HH 30 disk, such as spiral- and tail-like structures, a conical outflow, and a collimated jet. In contrast, the ALMA image reveals a flat dust disk in the disk midplane. By performing radiative transfer simulations, we show that grains of about 3 $\mu$m in radius or larger are fully vertically mixed to explain the observed mid-IR scattered light flux and its morphology, whereas millimeter-sized grains are settled into a layer with a scale height of $\gtrsim1$ au at $100$ au from the central star. We also find a tension in the disk inclination angle inferred from optical/near-IR and mm observations with the latter being closer to an exactly edge-on. Finally, we report the first detection of the proper motion of an emission knot associated with the mid-IR collimated jet detected by combining two epochs of our MIRI 12.8-$\mu$m observations., Comment: 29 pages, 23 figures; Accepted for publication in ApJ

Corrosion presents a major challenge to the longevity and reliability of products across various industries, particularly in the aerospace sector. Corrosion arises from chemical processes occurring on an atomistic scale, which lead to macroscopic degradation. Addressing this issue requires multi-scale modeling approaches, which rely on microscopic parameters that are challenging to measure experimentally or model with conventional quantum chemistry techniques. In this work, we develop and demonstrate a hybrid quantum-classical workflow tailored for atomistic simulations of corrosion processes, with a specific focus on the initial step of the oxygen reduction reaction -- a critical trigger for the corrosion of aluminum alloys widely used in modern aircraft. Using a combination of classical quantum chemistry methods and quantum computing frameworks, we identify reaction geometries characterized by multi-configurational electronic structures that are ideal for exploring with quantum algorithms. For the first time in this context, we explore both noisy intermediate-scale quantum and fault-tolerant quantum algorithms for these multi-configurational system, integrating them within a workflow designed to bridge atomistic simulations with macroscopic modeling approaches, such as finite element methods. Furthermore, we conduct a detailed quantum resource estimation to assess when and how quantum computers may play a meaningful role in tackling these problems. Our results demonstrate that significant advancements in quantum hardware but also in algorithms and error correction techniques are needed to make quantum computation practically viable for this class of problems. Nevertheless, this work establishes a critical foundation for applying quantum computation to corrosion modeling and highlights its potential to address complex, business-relevant challenges in materials science.

The crossing number of a graph is the least number of crossings over all drawings of the graph in the plane. Computing the crossing number of a given graph is NP-hard, but fixed-parameter tractable (FPT) with respect to the natural parameter. Two well-known variants of the problem are 2-layer crossing minimization and circular crossing minimization, where every vertex must lie on one of two layers, namely two parallel lines, or a circle, respectively. Both variants are NP-hard, but FPT with respect to the natural parameter. Recently, a local version of the crossing number has also received considerable attention. A graph is $k$-planar if it admits a drawing with at most $k$ crossings per edge. In contrast to the crossing number, recognizing $k$-planar graphs is NP-hard even if $k=1$. In this paper, we consider the two above variants in the local setting. The $k$-planar graphs that admit a straight-line drawing with vertices on two layers or on a circle are called 2-layer $k$-planar and outer $k$-planar graphs, respectively. We study the parameterized complexity of the two recognition problems with respect to $k$. For $k=0$, both problems can easily be solved in linear time. Two groups independently showed that outer 1-planar graphs can also be recognized in linear time [Algorithmica 2015/2016]. One group asked whether outer 2-planar graphs can be recognized in polynomial time. Our main contribution consists of XP-algorithms for recognizing 2-layer $k$-planar graphs and outer $k$-planar graphs. We complement these results by showing that both recognition problems are XNLP-hard. This implies that both problems are W$[t]$-hard for every $t$ and that it is unlikely that they admit FPT-algorithms. On the other hand, we present an FPT-algorithm for recognizing 2-layer $k$-planar graphs where the order of the vertices on one layer is specified., Comment: 23 pages, 6 figures

Modern time-series forecasting models often fail to make full use of rich unstructured information about the time series themselves. This lack of proper conditioning can lead to obvious model failures; for example, models may be unaware of the details of a particular product, and hence fail to anticipate seasonal surges in customer demand in the lead up to major exogenous events like holidays for clearly relevant products. To address this shortcoming, this paper introduces a novel forecast post-processor -- which we call LLMForecaster -- that fine-tunes large language models (LLMs) to incorporate unstructured semantic and contextual information and historical data to improve the forecasts from an existing demand forecasting pipeline. In an industry-scale retail application, we demonstrate that our technique yields statistically significantly forecast improvements across several sets of products subject to holiday-driven demand surges., Comment: Presented at NeurIPS Time Series in the Age of Large Models (2024)

We prove that the Kontsevich graph complex $GC_d^{2}$ and its oriented version $OGC_{d+1}^2$ are quasi-isomorphic as dg Lie algebras., Comment: 10 pages

The Coronagraphic Instrument onboard the Nancy Grace Roman Space Telescope is an important stepping stone towards the characterization of habitable, rocky exoplanets. In a technology demonstration phase conducted during the first 18 months of the mission (expected to launch in late 2026), novel starlight suppression technology may enable direct imaging of a Jupiter analog in reflected light. Here we summarize the current activities of the Observation Planning working group formed as part of the Community Participation Program. This working group is responsible for target selection and observation planning of both science and calibration targets in the technology demonstration phase of the Roman Coronagraph. We will discuss the ongoing efforts to expand target and reference catalogs, and to model astrophysical targets (exoplanets and circumstellar disks) within the Coronagraph's expected sensitivity. We will also present preparatory observations of high priority targets., Comment: Proceedings Volume 13092, Space Telescopes and Instrumentation 2024: Optical, Infrared, and Millimeter Wave; 1309255 (2024)

This article examines systematic oxygen (O)-incorporation to reduce total leakage currents in sputtered wurtzite-type ferroelectric Al0.73Sc0.27N thin films, along with its impact on the material structure and the polarity of the as-grown films. The O in the bulk Al0.73Sc0.27N was introduced through an external gas source during the reactive sputter process. In comparison to samples without doping, O-doped films showed almost a fourfold reduction of the leakage current near the coercive field. In addition, doping resulted in the reduction of the steady-state leakage currents by roughly one order of magnitude sub-coercive fields. Microstructure analysis using X-ray diffraction 1and scanning transmission electron microscopy (STEM) revealed no significant structural degradation of the bulk Al0.73Sc0.27N. In case of the maximum O-doped film, the c-axis out-of-plane texture increased by only 20% from 1.8{\deg} and chemical mapping revealed a uniform distribution of oxygen incorporation into the bulk. Our results further demonstrate the ability to control the as-deposited polarity of Al0.73Sc0.27N via the O-concentration, changing from nitrogen- to metal-polar orientation. Thus, this article presents a promising approach to mitigate the leakage current in wurtzite-type Al0.73Sc0.27N without incurring any significant structural degradation of the bulk thin film quality, thereby making ferroelectric nitrides more suitable for microelectronic applications.

Computing a Euclidean minimum spanning tree of a set of points is a seminal problem in computational geometry and geometric graph theory. We combine it with another classical problem in graph drawing, namely computing a monotone geometric representation of a given graph. More formally, given a finite set $S$ of points in the plane and a finite set $\cal D$ of directions, a geometric spanning tree $T$ with vertex set $S$ is ${\cal D}$-monotone if, for every pair $\{u,v\}$ of vertices of $T$, there exists a direction $d \in \cal D$ for which the unique path from $u$ to $v$ in $T$ is monotone with respect to $d$. We provide a characterization of ${\cal D}$-monotone spanning trees. Based on it, we show that a ${\cal D}$-monotone spanning tree of minimum length can be computed in polynomial time if the number $k=|{\cal D}|$ of directions is fixed, both when (i) the set ${\cal D}$ of directions is prescribed and when (ii) the objective is to find a minimum-length ${\cal D}$-monotone spanning tree over all sets ${\cal D}$ of $k$ directions. For $k = 2$, we describe algorithms that are much faster than those for the general case. Furthermore, in contrast to the classical Euclidean minimum spanning tree, whose vertex degree is at most six, we show that for every even integer $k$, there exists a point set $S_k$ and a set $\cal D$ of $k$ directions such that any minimum-length $\cal D$-monotone spanning tree of $S_k$ has maximum vertex degree $2k$., Comment: To appear in Proc. 50th International Conference on Current Trends in Theory and Practice of Computer Science (SOFSEM 2025)

Nanoscaling interstitial metal hydrides offers opportunities for hydrogenation applications by enhancing kinetics, increasing surface area, and allowing for tunable properties. The introduction of interfaces impacts hydrogen absorption properties and distribution heterogeneously, making it however challenging to examine the multiple concurrent mechanisms, especially at the atomic level. Here we demonstrate the effect of proximity on interstitial hydrogen in ultrathin single crystalline vanadium films, by comparing hydride formation in identically strained Fe/V- and Cr/V-superlattices. Pressure concentration and excess resistivity isotherms show higher absolute solubility of hydrogen, higher critical temperature and concentration in the Cr/V-superlattice. Direct measurements of hydrogen site location and thermal vibrations show identical occupation of octahedral z sites at room temperature with a vibrational amplitude of 0.20-0.25 {\AA} over a wide range of hydrogen concentrations. Our findings are consistent with a more extended region of hydrogen depletion in the vicinity of Fe compared to Cr, which showcases an inverse of the hydrogen spillover effect. Advancing the understanding of interface effects resolves previously puzzling differences in the hydrogen loading of Fe/V- and Cr/V-superlattices and is relevant for advancing both catalysis and storage.

PDS 453 is a rare highly inclined disk where the stellar photosphere is seen at grazing incidence on the disk surface. Our goal is take advantage of this geometry to constrain the structure and composition of this disk, in particular the fact that it shows a 3.1 $\mu$m water ice band in absorption that can be related uniquely to the disk. We observed the system in polarized intensity with the VLT/SPHERE instrument, as well as in polarized light and total intensity using the HST/NICMOS camera. Infrared archival photometry and a spectrum showing the water ice band are used to model the spectral energy distribution under Mie scattering theory. Based on these data, we fit a model using the radiative transfer code MCFOST to retrieve the geometry and dust and ice content of the disk. PDS 453 has the typical morphology of a highly inclined system with two reflection nebulae where the disk partially attenuates the stellar light. The upper nebula is brighter than the lower nebula and shows a curved surface brightness profile in polarized intensity, indicating a ring-like structure. With an inclination of 80{\deg} estimated from models, the line-of-sight crosses the disk surface and a combination of absorption and scattering by ice-rich dust grains produces the water ice band. PDS 453 is seen highly inclined and is composed of a mixture of silicate dust and water ice. The radial structure of the disk includes a significant jump in density and scale height at a radius of 70 au in order to produce a ring-like image. The depth of the 3.1 $\mu$m water ice band depends on the amount of water ice, until it saturates when the optical thickness along the line-of-sight becomes too large. Therefore, quantifying the exact amount of water from absorption bands in edge-on disks requires a detailed analysis of the disk structure and tailored radiative transfer modeling., Comment: 11 pages, 11 figures

Demand forecasting faces challenges induced by Peak Events (PEs) corresponding to special periods such as promotions and holidays. Peak events create significant spikes in demand followed by demand ramp down periods. Neural networks like MQCNN and MQT overreact to demand peaks by carrying over the elevated PE demand into subsequent Post-Peak-Event (PPE) periods, resulting in significantly over-biased forecasts. To tackle this challenge, we introduce a neural forecasting model called Split Peak Attention DEcomposition, SPADE. This model reduces the impact of PEs on subsequent forecasts by modeling forecasting as consisting of two separate tasks: one for PEs; and the other for the rest. Its architecture then uses masked convolution filters and a specialized Peak Attention module. We show SPADE's performance on a worldwide retail dataset with hundreds of millions of products. Our results reveal an overall PPE improvement of 4.5%, a 30% improvement for most affected forecasts after promotions and holidays, and an improvement in PE accuracy by 3.9%, relative to current production models.

We present a provisory scattered light detection of the Vega debris disk using deep Hubble Space Telescope coronagraphy (PID 16666). At only 7.7 parsecs, Vega is immensely important in debris disk studies both for its prominence and also because it allows the highest physical resolution among all debris systems relative to temperature zones around the star. We employ the STIS coronagraph's widest wedge position and classical Reference Differential Imaging to achieve among the lowest surface brightness sensitivities to date ($\sim 4\,\mu Jy/arcsec^{2}$) at wide separations using 32 orbits in Cycle 29. We detect a halo extending from the inner edge of our effective inner working angle at $10^{\prime\prime}.5$ out to the photon noise floor at $30^{\prime\prime}$ (80 - 230 au). The face-on orientation of the system and the lack of a perfectly color-matched PSF star have provided significant challenges to the reductions, particularly regarding artifacts from the imperfect color matching. However, we find that a halo of small dust grains provides the best explanation for the observed signal. Unlike Fomalhaut (a close twin to Vega in luminosity, distance, and age), there is no clear distinction in scattered light between the parent planetesimal belt observed with ALMA and the extended dust halo. These HST observations complement JWST GTO Cycle 1 observations of the system with NIRCam and MIRI.

We present images of the Vega planetary debris disk obtained at 15.5, 23, and 25.5 microns with the Mid-Infrared Instrument (MIRI) on JWST. The debris system is remarkably symmetric and smooth, and centered accurately on the star. There is a broad Kuiper-belt-analog ring at 80 to 170 au that coincides with the planetesimal belt detected with ALMA at 1.34 mm. The interior of the broad belt is filled with warm debris that shines most efficiently at mid-infrared along with a shallow flux dip/gap at 60 au from the star. These qualitative characteristics argue against any Saturn-mass planets orbiting the star outside of about 10 au assuming the unseen planet would be embedded in the very broad planetesimal disk from a few to hundred au. We find that the distribution of dust detected interior to the broad outer belt is consistent with grains being dragged inward by the Poynting-Robertson effect. Tighter constraints can be derived for planets in specific locations, for example any planet shepherding the inner edge of the outer belt is likely to be less than 6 Earth masses. The disk surface brightness profile along with the available infrared photometry suggest a disk inner edge near 3-5 au, disconnected from the sub-au region that gives rise to the hot near-infrared excess. The gap between the hot, sub-au zone and the inner edge of the warm debris might be shepherded by a modest mass, Neptune-size planet., Comment: accepted for publication in ApJ

The most prominent of the IRAS debris disk systems, $\alpha$ Lyrae (Vega), at a distance of 7.7 pc, has been observed by both the NIRCam and MIRI instruments on the James Webb Space Telescope (JWST). This paper describes NIRCam coronagraphic observations which have achieved F444W contrast levels of 3$\times10^{-7}$ at 1\arcsec\ (7.7 au), 1$\times10^{-7}$ at 2\arcsec\ (15 au) and few $\times 10^{-8}$ beyond 5\arcsec\ (38 au), corresponding to masses of $<$ 3, 2 and 0.5 MJup for a system age of 700 Myr. Two F444W objects are identified in the outer MIRI debris disk, around 48 au. One of these is detected by MIRI, appears to be extended and has a spectral energy distribution similar to those of distant extragalactic sources. The second one also appears extended in the NIRCam data suggestive of an extragalactic nature.The NIRCam limits within the inner disk (1\arcsec\ --10\arcsec) correspond to a model-dependent masses of 2$\sim$3 \mj. \citet{Su2024} argue that planets larger even 0.3 MJup would disrupt the smooth disk structure seen at MIRI wavelengths. Eight additional objects are found within 60\arcsec\ of Vega, but none has astrometric properties or colors consistent with planet candidates. These observations reach a level consistent with expected Jeans Mass limits. Deeper observations achieving contrast levels $<10^{-8}$ outside of $\sim$4\arcsec\ and reaching masses below that of Saturn are possible, but may not reveal a large population of new objects., Comment: 19 pages, 12 figures. accepted for Astronomical Journal

We report on the low-temperature fabrication (300$\deg$C) of ultrathin 2D amorphous carbon layers on III-V semiconductors by plasma-enhanced chemical vapor deposition as a universal template for remote epitaxy. We present growth and detailed characterization of 2D amorphous carbon layers on various host substrates and their subsequent remote epitaxial overgrowth by solid-source molecular beam epitaxy. We present the fabrication of ultra-smooth monolayer thick amorphous carbon layers with roughness $\leq 0.3$ nm determined by atomic-force microscopy and X-ray reflectivity measurement. We show that precisely tailoring the carbon layer thickness allows superior tunability of the substrate-layer interaction. Further, X-ray photoelectron and Raman spectroscopy measurements reveal predominantly sp$^2$-hybridised carbon in the amorphous layers. We observe that a low-temperature nucleation step is favorable for nucleation of III-V material growth on substrates coated with thin amorphous carbon layers. Under optimized preparation conditions, we obtain high-quality, single-crystalline, (001)-oriented GaAs, cubic-AlN, cubic-GaN and In(x)Ga(1-x)As, respectively, and various carbon-coated (001)-oriented substrates as GaAs, InP and 3C-SiC. Transmission electron microscopy images of the substrate-carbon-layer interface reveal a stretching of the atomic bonds at the interface and high-resolution X-ray diffraction measurements reveal high crystal quality and low dislocation densities $<1\times 10^7 \text{cm}^{-2}$. Our results show the universality of our carbon deposition process to fabricate templates for remote epitaxy, e. g., for remote epitaxy on temperature sensitive substrates like GaAs or InP and growth of metastable phases. Lift-off of layers from their substrates is demonstrated by employing a Ni stressor.

The celebrated Dvoretzky theorem asserts that every $N$-dimensional convex body admits central sections of dimension $d = \Omega(\log N)$, which is nearly spherical. For many instances of convex bodies, typically unit balls with respect to some norm, much better lower bounds on $d$ have been obtained, with most research focusing on such lower bounds and on the degree of approximation of the section by a $k$-dimensional Euclidean ball. In this note we focus on another parameter, namely the radius of the approximating ball. We focus on the case of the unit ball of the space $\ell_1^N$ (the so-called cross-polytope), which is relevant to various questions of interest in theoretical computer science. We will also survey other instances where similar questions for other normed spaces (most often $\ell_p$-spaces or their non-commutative analogues) were found relevant to problems in various areas of mathematics and its applications, and state some open problems. Finally, in view of the computer science ramifications, we will comment on the algorithmic aspects of finding nearly spherical sections., Comment: 12 pages

We study the Fubini-Study currents and equilibrium metrics of continuous Hermitian metrics on sequences of holomorphic line bundles over a fixed compact K\"ahler manifold. We show that the difference between the Fubini-Study currents and the curvature of the equilibrium metric, when appropriately scaled, converges to 0 in the sense of currents. As a consequence, we obtain sufficient conditions for the scaled Fubini-Study currents to converge weakly.

XTE J1829-098 is a transient X-ray pulsar with a period of ~7.8 s. It is a candidate Be star system, although the evidence for this is not yet definitive. We investigated the twenty-year long X-ray light curve using the Rossi X-ray Timing Explorer Proportional Counter Array (PCA), Neil Gehrels Swift Observatory Burst Alert Telescope (BAT), and the Monitor of All-sky X-ray Image (MAXI). We find that all three light curves are clearly modulated on the ~244 day orbital period previously reported from PCA monitoring observations, with outbursts confined to a narrow phase range. The light curves also show that XTE J1829-098 was in an inactive state between approximately December 2008 and April 2018 and no strong outbursts occurred. Such behavior is typical of Be X-ray binary systems, with the absence of outbursts likely related to the dissipation of the Be star's decretion disk. The mean outburst shapes can be approximated with a triangular profile and, from a joint fit of this to all three light curves, we refine the orbital period to 243.95 +/- 0.04 days. The mean outburst profile does not show any asymmetry and has a total phase duration of 0.140 +/- 0.007. However, the PCA light curve shows that there is considerable cycle-to-cycle variability of the individual outbursts. We compare the properties of XTE J1829-098 with other sources that show short phase-duration outbursts, in particular GS 1843-02 (2S 1845-024) which has a very similar orbital period, but longer pulse period, and whose orbit is known to be highly eccentric., Comment: Accepted for publication in the Astrophysical Journal. 21 pages

We present JWST broadband images of the highly inclined protoplanetary disk SSTc2d J163131.2-242627 (Oph163131) from 2.0 to 21$\mu$m. The images show a remarkable evolution in disk structure with wavelength, quite different from previous JWST observations of other edge-on disks. At 2.0 and 4.4$\mu$m, Oph163131 shows two scattering surfaces separated by a dark lane, typical of highly inclined disks. Starting at 7.7$\mu$m however, 1) the two linear nebulosities flanking the dark lane disappear; 2) the brighter nebula tracing the disk upper surface transitions into a compact central source distinctly larger than the JWST PSF and whose intrinsic size increases with wavelength; and 3) patches of extended emission appear at low latitudes, and at surprisingly large radii nearly twice that of the scattered light seen with $HST$ and NIRCam, and of the gas. We interpret the compact central source as thermal emission from the star and the inner disk that is not seen directly, but which instead is able to progressively propagate to greater distances at longer wavelengths. The lack of sharp-edged structures in the extended patchy emission argues against the presence of shocks and suggests photoexcitation or stochastic heating of material smoothly flowing away from the star along the disk surface. Finally, the dark lane thickness decreases significantly between 0.6$\mu$m and 4.4$\mu$m which indicates that the surface layers of Oph163131 lack grains larger than 1$\mu$m., Comment: Accepted for publication in the Astrophysical Journal. 11 pages, 7 figures

Ensuring robust performance on long-tail examples is an important problem for many real-world applications of machine learning, such as autonomous driving. This work focuses on the problem of identifying rare examples within a corpus of unlabeled data. We propose a simple and scalable data mining approach that leverages the knowledge contained within a large vision language model (VLM). Our approach utilizes a VLM to summarize the content of an image into a set of keywords, and we identify rare examples based on keyword frequency. We find that the VLM offers a distinct signal for identifying long-tail examples when compared to conventional methods based on model uncertainty. Therefore, we propose a simple and general approach for integrating signals from multiple mining algorithms. We evaluate the proposed method on two diverse tasks: 2D image classification, in which inter-class variation is the primary source of data diversity, and on 3D object detection, where intra-class variation is the main concern. Furthermore, through the detection task, we demonstrate that the knowledge extracted from 2D images is transferable to the 3D domain. Our experiments consistently show large improvements (between 10\% and 50\%) over the baseline techniques on several representative benchmarks: ImageNet-LT, Places-LT, and the Waymo Open Dataset.

Recent constraints on neutron star mass and radius have advanced our understanding of the equation of state (EOS) of cold dense matter. Some of them have been obtained by modeling the pulses of three millisecond X-ray pulsars observed by the Neutron Star Interior Composition Explorer (NICER). Here, we present a Bayesian parameter inference for a fourth pulsar, PSR J1231$-$1411, using the same technique with NICER and XMM-Newton data. When applying a broad mass-inclination prior from radio timing measurements and the emission region geometry model that can best explain the data, we find likely converged results only when using a limited radius prior. If limiting the radius to be consistent with the previous observational constraints and EOS analyses, we infer the radius to be $12.6 \pm 0.3$ km and the mass to be $1.04_{-0.03}^{+0.05}$ $M_\odot$, each reported as the posterior credible interval bounded by the $16\,\%$ and $84\,\%$ quantiles. If using an uninformative prior but limited between $10$ and $14$ km, we find otherwise similar results, but $R_{\mathrm{eq}} = 13.5_{-0.5}^{+0.3}$ km for the radius. In both cases, we find a nonantipodal hot region geometry where one emitting spot is at the equator or slightly above, surrounded by a large colder region, and where a noncircular hot region lies close to southern rotational pole. If using a wider radius prior, we only find solutions that fit the data significantly worse. We discuss the challenges in finding the better fitting solutions, possibly related to the weak interpulse feature in the pulse profile., Comment: 23 pages, 13 figures (2 of which are figure sets), 3 tables, published in ApJ

This study investigates the automated classification of Calls to Action (CTAs) within the 2021 German Instagram election campaign to advance the understanding of mobilization in social media contexts. We analyzed over 2,208 Instagram stories and 712 posts using fine-tuned BERT models and OpenAI's GPT-4 models. The fine-tuned BERT model incorporating synthetic training data achieved a macro F1 score of 0.93, demonstrating a robust classification performance. Our analysis revealed that 49.58% of Instagram posts and 10.64% of stories contained CTAs, highlighting significant differences in mobilization strategies between these content types. Additionally, we found that FDP and the Greens had the highest prevalence of CTAs in posts, whereas CDU and CSU led in story CTAs., Comment: Accepted Archival Paper for the CPSS Workshop at KONVENS 2024. Camera Ready Submission

We introduce the Tidal Tales Plugin, a Firefox extension for efficiently collecting and archiving of Instagram stories, addressing the challenges of ephemeral data in social media research. It enables an automated collection of story metadata and media files without risking account bans. It contributes to Web Science by facilitating expansive, long-term studies with enhanced data access and integrity.

Not every directed acyclic graph (DAG) whose underlying undirected graph is planar admits an upward planar drawing. We are interested in pushing the notion of upward drawings beyond planarity by considering upward $k$-planar drawings of DAGs in which the edges are monotonically increasing in a common direction and every edge is crossed at most $k$ times for some integer $k \ge 1$. We show that the number of crossings per edge in a monotone drawing is in general unbounded for the class of bipartite outerplanar, cubic, or bounded pathwidth DAGs. However, it is at most two for outerpaths and it is at most quadratic in the bandwidth in general. From the computational point of view, we prove that upward-$k$-planarity testing is NP-complete already for $k =1$ and even for restricted instances for which upward planarity testing is polynomial. On the positive side, we can decide in linear time whether a single-source DAG admits an upward $1$-planar drawing in which all vertices are incident to the outer face., Comment: This is the extended version, with full appendix, of a paper to appear in the Proc. 32nd Int. Symp. Graph Drawing & Network Visualization (GD 2024)

The optimal rate of reliable communication over a quantum channel can be enhanced by pre-shared entanglement. Whereas the enhancement may be unbounded in infinite-dimensional settings even when the input power is constrained, a long-standing conjecture asserts that the ratio between the entanglement-assisted and unassisted classical capacities is bounded in finite-dimensional settings [Bennett et al., IEEE Trans. Inf. Theory 48, 2637 (2002)]. In this work, we prove this conjecture by showing that their ratio is upper bounded by $o(d^2)$, where $d$ is the input dimension of the channel. An application to quantum communication with noisy encoders and decoders is given., Comment: 7 pages, 1 figure

For visual recognition, knowledge distillation typically involves transferring knowledge from a large, well-trained teacher model to a smaller student model. In this paper, we introduce an effective method to distill knowledge from an off-the-shelf vision-language model (VLM), demonstrating that it provides novel supervision in addition to those from a conventional vision-only teacher model. Our key technical contribution is the development of a framework that generates novel text supervision and distills free-form text into a vision encoder. We showcase the effectiveness of our approach, termed VLM-KD, across various benchmark datasets, showing that it surpasses several state-of-the-art long-tail visual classifiers. To our knowledge, this work is the first to utilize knowledge distillation with text supervision generated by an off-the-shelf VLM and apply it to vanilla randomly initialized vision encoders.

Emergent language research has made significant progress in recent years, but still largely fails to explore how communication emerges in more complex and situated multi-agent systems. Existing setups often employ a reference game, which limits the range of language emergence phenomena that can be studied, as the game consists of a single, purely language-based interaction between the agents. In this paper, we address these limitations and explore the emergence and utility of token-based communication in open-ended multi-agent environments, where situated agents interact with the environment through movement and communication over multiple time-steps. Specifically, we introduce two novel cooperative environments: Multi-Agent Pong and Collectors. These environments are interesting because optimal performance requires the emergence of a communication protocol, but moderate success can be achieved without one. By employing various methods from explainable AI research, such as saliency maps, perturbation, and diagnostic classifiers, we are able to track and interpret the agents' language channel use over time. We find that the emerging communication is sparse, with the agents only generating meaningful messages and acting upon incoming messages in states where they cannot succeed without coordination., Comment: 10 pages, 4 figures, 4 tables, preprint

Memory safety is an essential correctness property of software systems. For programs operating on linked heap-allocated data structures, the problem of proving memory safety boils down to analyzing the possible shapes of data structures, leading to the field of shape analysis. This paper presents a novel reduction-based approach to memory safety analysis that relies on two forms of abstraction: flow abstraction, representing global properties of the heap graph through local flow equations; and view abstraction, which enable verification tools to reason symbolically about an unbounded number of heap objects. In combination, the two abstractions make it possible to reduce memory-safety proofs to proofs about heap-less imperative programs that can be discharged using off-the-shelf software verification tools without built-in support for heap reasoning. Using an empirical evaluation on a broad range of programs, the paper shows that the reduction approach can effectively verify memory safety for sequential and concurrent programs operating on different kinds of linked data structures, including singly-linked, doubly-linked, and nested lists as well as trees.