Search

Over the past two decades, researchers have spent countless hours studying the impacts of public charter schools--independently-run, tuition-free schools of choice that serve some 3.7 million U.S. students today. Just prior to the pandemic, studies from Ohio and nationally indicated that charters on average delivered superior academic outcomes compared to traditional districts. And the very finest charters in Ohio and around the nation were driving learning gains that gave disadvantaged students the edge needed to succeed in college and career. The pandemic scrambled most everything about K-12 education. But did it upend what we know about charter school performance? The present study examines the post-pandemic performance of Ohio's brick-and-mortar charter schools, which enrolled 81,000 students--mostly from urban communities--during the 2022-23 school year. The results reveal that, in terms of student achievement growth, Ohio's charter schools remain a better educational option for the average charter student.

The conventional wisdom among educators and literacy gurus is that reading comprehension depends on the acquisition of isolatable, teachable, and generalizable skills. Consequently, many elementary and middle school English classrooms follow the "reading workshop" model, an approach to literacy instruction, with several variations that typically involve teachers spending a few minutes modeling a supposedly important skill before sending students off to practice by reading self-selected but appropriately "leveled" books. This policy brief challenges that orthodoxy. It asserts that, once students have learned to decode, reading books and other texts of any purported "level" with understanding depends more on knowledge than skills and that successful knowledge building requires explicit, carefully sequenced and paced, teacher-directed instruction across multiple subjects, including but not limited to social studies, science, and literature. Key questions asked in this report include: (1) Does reading comprehension depend on acquiring a set of teachable skills?; (2) Do students need practice with "just right" books?; (3) Does letting students choose the books they read foster the motivation necessary to improve reading comprehension?; and (4) Does extended literacy instruction enhance reading comprehension?

We investigate the effectiveness of Explainable AI (XAI) in verifying Machine Unlearning (MU) within the context of harbor front monitoring, focusing on data privacy and regulatory compliance. With the increasing need to adhere to privacy legislation such as the General Data Protection Regulation (GDPR), traditional methods of retraining ML models for data deletions prove impractical due to their complexity and resource demands. MU offers a solution by enabling models to selectively forget specific learned patterns without full retraining. We explore various removal techniques, including data relabeling, and model perturbation. Then, we leverage attribution-based XAI to discuss the effects of unlearning on model performance. Our proof-of-concept introduces feature importance as an innovative verification step for MU, expanding beyond traditional metrics and demonstrating techniques' ability to reduce reliance on undesired patterns. Additionally, we propose two novel XAI-based metrics, Heatmap Coverage (HC) and Attention Shift (AS), to evaluate the effectiveness of these methods. This approach not only highlights how XAI can complement MU by providing effective verification, but also sets the stage for future research to enhance their joint integration., Comment: ICPRW2024

Given the vast differences in interaction details, describing the dynamics of structurally disordered materials in a unified theoretical framework presents a fundamental challenge to condensed-matter physics and materials science. This paper investigates numerically a percolation scenario for the two most important relaxation processes of supercooled liquids and glasses. For nine binary glass formers we find that, as temperature is lowered from the liquid state, percolation of immobile particles takes place at the temperature locating the $\alpha$ process. Mirroring this, upon continued cooling into the glass, mobile-particle percolation pinpoints a Johari-Goldstein $\beta$ relaxation whenever it is well separated from the $\alpha$ process. For 2D systems under the same conditions, percolation of mobile and immobile particles occurs nearly simultaneously and no $\beta$ relaxation can be identified. Our findings suggest a general description of glassy dynamics based on a percolation perspective., Comment: Accepted by Nature Physics, "in principle" (this is the version originally submitted to NP)

Single-mode fibers are used in fiber-optic gyroscopes to measure the Sagnac effect and are planned to be used in forthcoming experiments on the gravitationally induced phase shift in single photons. However, current theoretical models of such experiments are limited to ray-optics approximations or, if based on wave optics, to a restricted class of fiber alignments. To overcome these shortcomings, this paper develops a comprehensive perturbative scheme to solve for electromagnetic modes, i.e., monochromatic solutions to Maxwell's equations, of arbitrarily bent step-index fibers in general stationary space-times. This leads to transport equations for the electromagnetic phase and polarization that include the gravitational redshift, the Sagnac effect, a generalization of Rytov's law to curved space, a gravitational Faraday effect in the form of shift-induced gyrotropy, as well as inverse spin Hall effects caused by fiber bending, gravitational acceleration, and space-time curvature., Comment: 12 pages, 3 figures

Young supernova remnants (SNRs) are believed to be the origin of energetic cosmic rays (CRs) below the "knee" of their spectrum at $\sim3$ petaelectronvolt (PeV, $10^{15}$ eV). Nevertheless, the precise location, duration, and operation of CR acceleration in young SNRs are open questions. Here, we report on multi-epoch X-ray observations of Cassiopeia A (Cas A), a 350-year-old SNR, in the 15-50 keV band that probes the most energetic CR electrons. The observed X-ray flux decrease $(15\pm1\%)$, contrary to the expected $>$90\% decrease based on previous radio, X-ray, and gamma-ray observations, provides unambiguous evidence for CR electron acceleration operating in Cas A. A temporal model for the radio and X-ray data accounting for electron cooling and continuous injection finds that the freshly injected electron spectrum is significantly harder (exponential cutoff power law index $q=2.15$), and its cutoff energy is much higher ($E_{cut}=36$ TeV) than the relic electron spectrum ($q=2.44\pm0.03$, $E_{cut}=4\pm1$ TeV). Both electron spectra are naturally explained by the recently developed modified nonlinear diffusive shock acceleration (mNLDSA) mechanism. The CR protons producing the observed gamma rays are likely accelerated at the same location by the same mechanism as those for the injected electron. The Cas A observations and spectral modeling represent the first time radio, X-ray, gamma ray and CR spectra have been self-consistently tied to a specific acceleration mechanism -- mNLDSA -- in a young SNR., Comment: Accepted for publication in ApJ. 14 pages, 4 figures, 2 tables

Adversarial training can be used to learn models that are robust against perturbations. For linear models, it can be formulated as a convex optimization problem. Compared to methods proposed in the context of deep learning, leveraging the optimization structure allows significantly faster convergence rates. Still, the use of generic convex solvers can be inefficient for large-scale problems. Here, we propose tailored optimization algorithms for the adversarial training of linear models, which render large-scale regression and classification problems more tractable. For regression problems, we propose a family of solvers based on iterative ridge regression and, for classification, a family of solvers based on projected gradient descent. The methods are based on extended variable reformulations of the original problem. We illustrate their efficiency in numerical examples.

Image monitoring and guidance during medical examinations can aid both diagnosis and treatment. However, the sampling frequency is often too low, which creates a need to estimate the missing images. We present a probabilistic motion model for sequential medical images, with the ability to both estimate motion between acquired images and forecast the motion ahead of time. The core is a low-dimensional temporal process based on a linear Gaussian state-space model with analytically tractable solutions for forecasting, simulation, and imputation of missing samples. The results, from two experiments on publicly available cardiac datasets, show reliable motion estimates and an improved forecasting performance using patient-specific adaptation by online learning., Comment: Medical Image Computing and Computer Assisted Intervention (MICCAI) 2024

We study the effects of missingness on the estimation of population parameters. Moving beyond restrictive missing completely at random (MCAR) assumptions, we first formulate a missing data analogue of Huber's arbitrary $\epsilon$-contamination model. For mean estimation with respect to squared Euclidean error loss, we show that the minimax quantiles decompose as a sum of the corresponding minimax quantiles under a heterogeneous, MCAR assumption, and a robust error term, depending on $\epsilon$, that reflects the additional error incurred by departure from MCAR. We next introduce natural classes of realisable $\epsilon$-contamination models, where an MCAR version of a base distribution $P$ is contaminated by an arbitrary missing not at random (MNAR) version of $P$. These classes are rich enough to capture various notions of biased sampling and sensitivity conditions, yet we show that they enjoy improved minimax performance relative to our earlier arbitrary contamination classes for both parametric and nonparametric classes of base distributions. For instance, with a univariate Gaussian base distribution, consistent mean estimation over realisable $\epsilon$-contamination classes is possible even when $\epsilon$ and the proportion of missingness converge (slowly) to 1. Finally, we extend our results to the setting of departures from missing at random (MAR) in normal linear regression with a realisable missing response., Comment: 86 pages, 6 figures

For more than twenty-five years, public charter schools have served Ohio families and communities by providing quality educational options beyond the local school district. But it's no secret that we've also had a long-standing debate over whether increasing school choice impacts students who remain in traditional districts. In important--and sometimes impassioned--discussions such as these, rigorous research is critical to ground conversations in facts and evidence. Our latest report offers an analysis of the rapid scale-up of Ohio charter schools during the late 1990s and early 2000s. It finds that charters slightly boosted the graduation and attendance rates of traditional district students, while having no significant impacts on their state exam scores. These results follow a body of research from various locales showing that expanding educational choice--whether via public charter schools or private schools--consistently yields neutral to slightly positive impacts on traditional districts.

We present Agglomerative Token Clustering (ATC), a novel token merging method that consistently outperforms previous token merging and pruning methods across image classification, image synthesis, and object detection & segmentation tasks. ATC merges clusters through bottom-up hierarchical clustering, without the introduction of extra learnable parameters. We find that ATC achieves state-of-the-art performance across all tasks, and can even perform on par with prior state-of-the-art when applied off-the-shelf, i.e. without fine-tuning. ATC is particularly effective when applied with low keep rates, where only a small fraction of tokens are kept and retaining task performance is especially difficult., Comment: ECCV 2024. Project webpage at https://vap.aau.dk/atc/

The SoccerNet 2024 challenges represent the fourth annual video understanding challenges organized by the SoccerNet team. These challenges aim to advance research across multiple themes in football, including broadcast video understanding, field understanding, and player understanding. This year, the challenges encompass four vision-based tasks. (1) Ball Action Spotting, focusing on precisely localizing when and which soccer actions related to the ball occur, (2) Dense Video Captioning, focusing on describing the broadcast with natural language and anchored timestamps, (3) Multi-View Foul Recognition, a novel task focusing on analyzing multiple viewpoints of a potential foul incident to classify whether a foul occurred and assess its severity, (4) Game State Reconstruction, another novel task focusing on reconstructing the game state from broadcast videos onto a 2D top-view map of the field. Detailed information about the tasks, challenges, and leaderboards can be found at https://www.soccer-net.org, with baselines and development kits available at https://github.com/SoccerNet., Comment: 7 pages, 1 figure

Diffusion models have achieved remarkable progress in generative modelling, particularly in enhancing image quality to conform to human preferences. Recently, these models have also been applied to low-level computer vision for photo-realistic image restoration (IR) in tasks such as image denoising, deblurring, dehazing, etc. In this review paper, we introduce key constructions in diffusion models and survey contemporary techniques that make use of diffusion models in solving general IR tasks. Furthermore, we point out the main challenges and limitations of existing diffusion-based IR frameworks and provide potential directions for future work., Comment: Review paper; any comments and suggestions are most welcome!

Generative diffusions are a powerful class of Monte Carlo samplers that leverage bridging Markov processes to approximate complex, high-dimensional distributions, such as those found in image processing and language models. Despite their success in these domains, an important open challenge remains: extending these techniques to sample from conditional distributions, as required in, for example, Bayesian inverse problems. In this paper, we present a comprehensive review of existing computational approaches to conditional sampling within generative diffusion models. Specifically, we highlight key methodologies that either utilise the joint distribution, or rely on (pre-trained) marginal distributions with explicit likelihoods, to construct conditional generative samplers.

We study the efficient estimation of a class of mean functionals in settings where a complete multivariate dataset is complemented by additional datasets recording subsets of the variables of interest. These datasets are allowed to have a general, in particular non-monotonic, structure. Our main contribution is to characterise the asymptotic minimal mean squared error for these problems and to introduce an estimator whose risk approximately matches this lower bound. We show that the efficient rescaled variance can be expressed as the minimal value of a quadratic optimisation problem over a function space, thus establishing a fundamental link between these estimation problems and the theory of generalised ANOVA decompositions. Our estimation procedure uses iterated nonparametric regression to mimic an approximate influence function derived through gradient descent. We prove that this estimator is approximately normally distributed, provide an estimator of its variance and thus develop confidence intervals of asymptotically minimal width. Finally we study a more direct estimator, which can be seen as a U-statistic with a data-dependent kernel, showing that it is also efficient under stronger regularity conditions., Comment: 69 pages

Uncertainty estimation is a necessary component when implementing AI in high-risk settings, such as autonomous cars, medicine, or insurances. Large Language Models (LLMs) have seen a surge in popularity in recent years, but they are subject to hallucinations, which may cause serious harm in high-risk settings. Despite their success, LLMs are expensive to train and run: they need a large amount of computations and memory, preventing the use of ensembling methods in practice. In this work, we present a novel method that allows for fast and memory-friendly training of LLM ensembles. We show that the resulting ensembles can detect hallucinations and are a viable approach in practice as only one GPU is needed for training and inference., Comment: 6 pages, 3 figures

Safe Bayesian optimization (BO) algorithms promise to find optimal control policies without knowing the system dynamics while at the same time guaranteeing safety with high probability. In exchange for those guarantees, popular algorithms require a smoothness assumption: a known upper bound on a norm in a reproducing kernel Hilbert space (RKHS). The RKHS is a potentially infinite-dimensional space, and it is unclear how to, in practice, obtain an upper bound of an unknown function in its corresponding RKHS. In response, we propose an algorithm that estimates an upper bound on the RKHS norm of an unknown function from data and investigate its theoretical properties. Moreover, akin to Lipschitz-based methods, we treat the RKHS norm as a local rather than a global object, and thus reduce conservatism. Integrating the RKHS norm estimation and the local interpretation of the RKHS norm into a safe BO algorithm yields PACSBO, an algorithm for probably approximately correct safe Bayesian optimization, for which we provide numerical and hardware experiments that demonstrate its applicability and benefits over popular safe BO algorithms., Comment: Accepted to the Symposium on Systems Theory in Data and Optimization (SysDO 2024). This is a preprint of the final version, which is to appear in Lecture Notes in Control and Information Sciences - Proceedings

The study of meteorites and their microstructures is a topic which spans multiple fields of research, such as meteoritics and materials science. For materials scientists and engineers, the extreme and unusual conditions which these microstructures form allow for insight into materials which would exist at the edge of our thermomechanical processing abilities. One such microstructure found in low-shock event iron meteorites is Neumann bands. These bands are an array of lenticular deformation twins that form throughout the Fe-Ni matrix with numerous intersections, resulting in many high stress and strain regions within the material's surface. The existence of these regions and the shocks that formed them encourage atypical strain accommodating mechanisms and structural changes of the material. However, direct investigation of the deformation twin intersections and the microstructural behaviour in and around these regions has been limited. In this work, investigation of these regions in a Campo del Cielo meteorite fragment, with electron backscatter diffraction (EBSD) and forescatter electron (FSE) imaging, revealed two primary findings: high-intensity pattern doubling mirrored across the {110} band at twin-twin intersection and microband formation across the sample surface, which suggest multilayer twinning and constraint of the crystal structure at points of twin-twin intersection. Microbands were found to form along the {110} plane and in regions near Neumann bands. The simultaneous existence of Neumann bands (microtwins) and microbands is presented here for a BCC material, and it is believed the Neumann band and microbands formed during different types and/or shock events. The presence of both Neumann bands and microbands within a BCC iron meteorite is previously unreported and may be valuable in furthering our understanding of shock deformation within iron-based materials., Comment: As submitted pre-print

In this report we investigate the use of the Tustin neural network architecture (Tustin-Net) for the identification of a physical rotary inverse pendulum. This physics-based architecture is of particular interest as it builds on the known relationship between velocities and positions. We here aim at discussing the advantages, limitations and performance of Tustin-Nets compared to first-principles grey-box models on a real physical apparatus, showing how, with a standard training procedure, the former can hardly achieve the same accuracy as the latter. To address this limitation, we present a training strategy based on transfer learning that yields Tustin-Nets that are competitive with the first-principles model, without requiring extensive knowledge of the setup as the latter.

The hypersphere model is a simple one-parameter model of the potential energy landscape of viscous liquids, which consists of a percolating system of hyperspheres of equal sizes randomly distributed in $R^{3N}$ where $N$ is the number of particles. We study random walks within overlapping hyperspheres in 12 to 45 dimensions, utilizing an algorithm for on-the-fly placement of the hyperspheres in conjunction with the kinetic Monte Carlo method. We find behavior typical of viscous liquids; thus decreasing the hypersphere density (corresponding to decreasing the temperature) leads to a slowing down of the dynamics by many orders of magnitude. The shape of the mean-square displacement as a function of time is found to be very similar to that of the Kob-Andersen binary Lennard-Jones mixture and the Random Barrier Model, which predicts well the frequency-dependent fluidity of nine glass-forming liquids of different chemistry [Bierwirth et al., Phys. Rev. Lett. 119, 248001 (2017)].

The probabilities for gamma-ray and particle emission as a function of the excitation energy of a decaying nucleus are valuable observables for constraining the ingredients of the models that describe the de-excitation of nuclei near the particle emission threshold. These models are essential in nuclear astrophysics and applications. In this work, we have for the first time simultaneously measured the gamma-ray and neutron emission probabilities of 208Pb. The measurement was performed in inverse kinematics at the Experimental Storage Ring (ESR) of the GSI/FAIR facility, where a 208Pb beam interacted through the 208Pb(p,p') reaction with a hydrogen gas jet target. Instead of detecting the gamma-rays and neutrons emitted by 208Pb, we detected the heavy beam-like residues produced after gamma and neutron emission. These heavy residues were fully separated by a dipole magnet of the ESR and were detected with outstanding efficiencies. The comparison of the measured probabilities with model calculations has allowed us to test and select different descriptions of the gamma-ray strength function and the nuclear level density available in the literature., Comment: 23 pages, 14 figures

We combine a new Galactic plane survey of Hydrogen Radio Recombination Lines (RRLs) with far-infrared (FIR) surveys of ionized Nitrogen, N+, to determine Nitrogen abundance across Galactic radius. RRLs were observed with NASA DSS-43 70m antenna and the Green Bank Telescope in 108 lines-of-sight spanning -135 degrees < l < 60 degrees, at b=0 degrees. These positions were also observed in [N II] 122 um and 205 um lines with the Herschel Space Observatory. Combining RRL and [N II] 122 um and 205 um observations in 41 of 108 samples with high signal-to-noise ratio, we studied ionized Nitrogen abundance distribution across Galactocentric distances of 0-8 kpc. Combined with existing Solar neighborhood and Outer galaxy N/H abundance determinations, we studied this quantity's distribution within the Milky Way's inner 17 kpc for the first time. We found a Nitrogen abundance gradient extending from Galactocentric radii of 4-17 kpc in the Galactic plane, while within 0-4 kpc, the N/H distribution remained flat. The gradient observed at large Galactocentric distances supports inside-out galaxy growth with the additional steepening resulting from variable star formation efficiency and/or radial flows in the Galactic disk, while the inner 4 kpc flattening, coinciding with the Galactic bar's onset, may be linked to radial flows induced by the bar potential. Using SOFIA/FIFI-LS and Herschel/PACS, we observed the [N III] 57 um line to trace doubly ionized gas contribution in a sub-sample of sightlines. We found negligible N++ contributions along these sightlines, suggesting mostly singly ionized Nitrogen originating from low ionization H II region outskirts., Comment: Accepted for publication at the Astrophysical Journal. 25 pages, 13 figures

Generative AI (genAI) is being used in education for different purposes. From the teachers' perspective, genAI can support activities such as learning design. However, there is a need to study the impact of genAI on the teachers' agency. While GenAI can support certain processes of idea generation and co-creation, GenAI has the potential to negatively affect professional agency due to teachers' limited power to (i) act, (ii) affect matters, and (iii) make decisions or choices, as well as the possibility to (iv) take a stance. Agency is identified in the learning sciences studies as being one of the factors in teachers' ability to trust AI. This paper aims to introduce a dual perspective. First, educational technology, as opposed to other computer-mediated communication (CMC) tools, has two distinctly different user groups and different user needs, in the form of learners and teachers, to cater for. Second, the design of educational technology often prioritises learner agency and engagement, thereby limiting the opportunities for teachers to influence the technology and take action. This study aims to analyse the way GenAI is influencing teachers' agency. After identifying the current limits of GenAI, a solution based on the combination of human intelligence and artificial intelligence through a hybrid intelligence approach is proposed. This combination opens up the discussion of a collaboration between teacher and genAI being able to open up new practices in learning design in which they HI support the extension of the teachers' activity.

In this study, we formulate the task of Video Anomaly Detection as a probabilistic analysis of object bounding boxes. We hypothesize that the representation of objects via their bounding boxes only, can be sufficient to successfully identify anomalous events in a scene. The implied value of this approach is increased object anonymization, faster model training and fewer computational resources. This can particularly benefit applications within video surveillance running on edge devices such as cameras. We design our model based on human reasoning which lends itself to explaining model output in human-understandable terms. Meanwhile, the slowest model trains within less than 7 seconds on a 11th Generation Intel Core i9 Processor. While our approach constitutes a drastic reduction of problem feature space in comparison with prior art, we show that this does not result in a reduction in performance: the results we report are highly competitive on the benchmark datasets CUHK Avenue and ShanghaiTech, and significantly exceed on the latest State-of-the-Art results on StreetScene, which has so far proven to be the most challenging VAD dataset., Comment: Accepted for publication at GCPR 2024, after peer review. Use of this Accepted Version is subject to the publisher's Accepted Manuscript terms of use https://www.springer-nature.com/gp/open-research/policies/accepted-manuscript-terms. Code available on GitHub: https://github.com/milestonesys-research/VAD-with-PGMs/

For many years, the conventional wisdom in the field was that grade retention was a bad idea. A 1997 opinion piece in "Education Week" titled "Grade retention doesn't work" reflected the prevailing sentiment in the education community and the available research evidence at that time: retained students performed worse than their promoted peers in the years that followed. This brief challenges that notion, based on more recent studies that do a better job of isolating the causal effect of retention. Key Questions: (1) Can grade retention be beneficial for students?; (2) What risks are associated with retention?; and (3) Is grade retention too costly for school systems?.

Thanks to rapid increases in the state's Hispanic and Asian populations, the number of Texas students classified as English Learners has surged in the past decade, from approximately 830,000 in 2010 to more than 1.1 million today. In the charter sector, change has been even more rapid. In 2010, 16 percent of charter students in Texas were classified as English Learners. Yet by 2021, that figure had nearly doubled to just under 30 percent, even as the share of all Texas students who enrolled in charters "also" doubled (and then some). Consequently, the total number of English Learners in Texas charters has quintupled in the past decade, from less than 25,000 in 2010 to nearly 120,000 in 2021. The rapid growth in Texas's English-Learner population puts a premium on analyzing the educational experiences of this student group, and the disproportionate share of this growth that has occurred in the state's charter sector makes it particularly important to understand how English Learners in that sector are faring. Accordingly, this report addresses three research questions: (1) How do the English Learners who enroll in Texas charter schools compare to their peers in traditional public schools?; (2) How do the academic and economic outcomes of English Learners in Texas charter schools compare to those of their peers in traditional public schools?; and (3) How have the academic outcomes of English Learners in Texas charter schools evolved as the number of English Learners in charters has increased? To address these questions, this report uses student-level data spanning much of the first two decades of this century and a wide range of analytic techniques. In so doing, it brings important new evidence to bear on the educational experiences of one of the fastest-growing student populations in one of the fastest-growing sectors in one of the fastest-growing states in the country. [This report was written with David Griffith. The foreword was written by Michael J. Petrilli and David Griffith.]

Magnetic skyrmions are promising candidates for information and storage technologies. In the last years, magnetic multilayer systems have been tuned to enable room-temperature skyrmions, stable even in the absence of external magnetic field. There are several models describing the properties of an isolated skyrmion in a homogeneous background for single repetition multilayer stack, however, the description on how the equilibrium skyrmion size in lattices scales with increasing the number of repetitions of the stack remains unaddressed. This question is essential for fundamental and practical perspectives, as the behaviour of an ensemble of skyrmions differs from the isolated case. Based on a multilayer stack hosting a skyrmion lattice, we have carried out a series of imaging experiments scaling up the dipolar interaction by repeating $n$ times the multilayer unit, from $n =1$ up to $n=30$. We have developed an analytical description for the skyrmion radius in the whole multilayer regime, $i.e.$, from thin to thick film limits. Furthermore, we provide insight on how nucleation by an externally applied field can give rise to a lattice with more skyrmions (thus, overfilled) than the predicted by the calculations., Comment: 9 pages, 3 figures

Gravitational wave memory is said to arise when a gravitational wave burst produces changes in a physical system that persist even after that wave has passed. This paper analyzes gravitational wave bursts in plane wave spacetimes, deriving memory effects on timelike and null geodesics, massless scalar fields, and massless spinning particles whose motion is described by the spin Hall equations. All associated memory effects are found to be characterized by four ``memory tensors,'' three of which are independent. These tensors form a scattering matrix for the transverse components of geodesics. However, unlike for the ``classical'' memory effect involving initially comoving pairs of timelike geodesics, one of our results is that memory effects for null geodesics can have strong longitudinal components. When considering massless particles with spin, we solve the spin Hall equations analytically by showing that there exists a conservation law associated with each conformal Killing vector field. These solutions depend only on the same four memory tensors that control geodesic scattering. For massless scalar fields, we show that given any solution in flat spacetime, a weak-field solution in a plane wave spacetime can be generated just by differentiation. Precisely which derivatives are involved depend on the same four memory tensors. These effects are illustrated for scalar plane waves and higher-order Gaussian beams. Furthermore, we also present a numerical comparison between the dynamics of localized wave packets carrying angular momentum and the spin Hall equations.

Isomorphs are curves in the thermodynamic phase diagram of invariant excess entropy, structure, and dynamics, while pseudoisomorphs are curves of invariant structure and dynamics, but not of the excess entropy. The latter curves have been shown to exist in molecular models with flexible bonds [A. E. Olsen et al., J. Chem. Phys. 145, 241103 (2016)]. We here present three force-based methods to trace out pseudoisomorphs based on a single configuration and test them on the asymmetric dumbbell and 10-bead Lennard-Jones chain models with bonds modeled as harmonic springs. The three methods are based on requiring that particle forces, center-of-mass forces, and torques, respectively, are invariant in reduced units. For each of the two investigated models we identify a method that works well for tracing out pseudoisomorphs, but these methods are not the same. Overall, it appears that the more internal degrees of freedom there are in the molecule studied, the less they affect the gross dynamical behavior. Moreover, the "internal" degrees of freedom (including rotation) do not appear to significantly affect the scaling behavior of the dynamical/transport coefficients provided some "quenching" is performed.

Can pretrained models generalize to new datasets without any retraining? We deploy pretrained image models on datasets they were not trained for, and investigate whether their embeddings form meaningful clusters. Our suite of benchmarking experiments use encoders pretrained solely on ImageNet-1k with either supervised or self-supervised training techniques, deployed on image datasets that were not seen during training, and clustered with conventional clustering algorithms. This evaluation provides new insights into the embeddings of self-supervised models, which prioritize different features to supervised models. Supervised encoders typically offer more utility than SSL encoders within the training domain, and vice-versa far outside of it, however, fine-tuned encoders demonstrate the opposite trend. Clustering provides a way to evaluate the utility of self-supervised learned representations orthogonal to existing methods such as kNN. Additionally, we find the silhouette score when measured in a UMAP-reduced space is highly correlated with clustering performance, and can therefore be used as a proxy for clustering performance on data with no ground truth labels. Our code implementation is available at \url{https://github.com/scottclowe/zs-ssl-clustering/}.

Given an unconditional diffusion model $\pi(x, y)$, using it to perform conditional simulation $\pi(x \mid y)$ is still largely an open question and is typically achieved by learning conditional drifts to the denoising SDE after the fact. In this work, we express conditional simulation as an inference problem on an augmented space corresponding to a partial SDE bridge. This perspective allows us to implement efficient and principled particle Gibbs and pseudo-marginal samplers marginally targeting the conditional distribution $\pi(x \mid y)$. Contrary to existing methodology, our methods do not introduce any additional approximation to the unconditional diffusion model aside from the Monte Carlo error. We showcase the benefits and drawbacks of our approach on a series of synthetic and real data examples., Comment: 24 pages, 12 figures

Most of the literature on differential privacy considers the item-level case where each user has a single observation, but a growing field of interest is that of user-level privacy where each of the $n$ users holds $T$ observations and wishes to maintain the privacy of their entire collection. In this paper, we derive a general minimax lower bound, which shows that, for locally private user-level estimation problems, the risk cannot, in general, be made to vanish for a fixed number of users even when each user holds an arbitrarily large number of observations. We then derive matching, up to logarithmic factors, lower and upper bounds for univariate and multidimensional mean estimation, sparse mean estimation and non-parametric density estimation. In particular, with other model parameters held fixed, we observe phase transition phenomena in the minimax rates as $T$ the number of observations each user holds varies. In the case of (non-sparse) mean estimation and density estimation, we see that, for $T$ below a phase transition boundary, the rate is the same as having $nT$ users in the item-level setting. Different behaviour is however observed in the case of $s$-sparse $d$-dimensional mean estimation, wherein consistent estimation is impossible when $d$ exceeds the number of observations in the item-level setting, but is possible in the user-level setting when $T \gtrsim s \log (d)$, up to logarithmic factors. This may be of independent interest for applications as an example of a high-dimensional problem that is feasible under local privacy constraints., Comment: 57 pages, 1 figure, 4 tables

Video Foundation Models (ViFMs) aim to learn a general-purpose representation for various video understanding tasks. Leveraging large-scale datasets and powerful models, ViFMs achieve this by capturing robust and generic features from video data. This survey analyzes over 200 video foundational models, offering a comprehensive overview of benchmarks and evaluation metrics across 14 distinct video tasks categorized into 3 main categories. Additionally, we offer an in-depth performance analysis of these models for the 6 most common video tasks. We categorize ViFMs into three categories: 1) Image-based ViFMs, which adapt existing image models for video tasks, 2) Video-Based ViFMs, which utilize video-specific encoding methods, and 3) Universal Foundational Models (UFMs), which combine multiple modalities (image, video, audio, and text etc.) within a single framework. By comparing the performance of various ViFMs on different tasks, this survey offers valuable insights into their strengths and weaknesses, guiding future advancements in video understanding. Our analysis surprisingly reveals that image-based foundation models consistently outperform video-based models on most video understanding tasks. Additionally, UFMs, which leverage diverse modalities, demonstrate superior performance on video tasks. We share the comprehensive list of ViFMs studied in this work at: \url{https://github.com/NeeluMadan/ViFM_Survey.git}

We present extensive observations of the Type II supernova (SN II) 2023ufx which is likely the most metal-poor SN II observed to-date. It exploded in the outskirts of a low-metallicity ($Z_{\rm host} \sim 0.1~Z_\odot$) dwarf ($M_g = -13.23\pm0.15$~mag; $r_e\sim 1$~kpc) galaxy. The explosion is luminous, peaking at $M_g\approx -18.5~$mag, and shows rapid evolution. The $r$-band (pseudo-bolometric) light curve has a shock-cooling phase lasting 20 (17) days followed by a 19 (23)-day plateau. The entire optically-thick phase lasts only $\approx 55~$days following explosion, indicating that the red supergiant progenitor had a thinned H envelope prior to explosion. The early spectra obtained during the shock-cooling phase show no evidence for narrow emission features and limit the pre-explosion mass-loss rate to $\dot{M} \lesssim 10^{-3}~\rm M_\odot$/yr. The photospheric-phase spectra are devoid of prominent metal absorption features, indicating a progenitor metallicity of $\lesssim 0.1~Z_\odot$. The semi-nebular ($\sim 60-130~$d) spectra reveal weak Fe II, but other metal species typically observed at these phases (Ti II, Sc II, Ba II) are conspicuously absent. The late-phase optical and near-infrared spectra also reveal broad ($\approx 10^4~\rm{km}~\rm s^{-1}$) double-peaked H$\alpha$, P$\beta$, and P$\gamma$ emission profiles suggestive of a fast outflow launched during the explosion. Outflows are typically attributed to rapidly-rotating progenitors which also prefer metal-poor environments. This is only the second SN II with $\lesssim 0.1~Z_\odot$ and both exhibit peculiar evolution, suggesting a sizable fraction of metal-poor SNe II have distinct properties compared to nearby metal-enriched SNe II. These observations lay the groundwork for modeling the metal-poor SNe II expected in the early Universe., Comment: 17 pages, 15 figures and 3 tables in main text, an additional 5 pages, 4 figures, and 2 tables in the appendix. Accepted by ApJ, spectra and photometry are included as ancillary data

As part of the R2D2 (Rare Decays with Radial Detector) R&D, the use of a gas detector with a spherical or cylindrical cathode, equipped with a single anode and operating at high pressure, was studied for the search of rare phenomena such as neutrinoless double-beta decay. The presented measurements were obtained with a cylindrical detector, covering gas pressures ranging from 1 to 10 bar in argon and 1 to 6 bar in xenon, using both a point-like source of $^{210}$Po (5.3 MeV $\alpha$ ) and a diffuse source of $^{222}$Rn (5.5 MeV $\alpha$). Analysis and interpretation of the data were developed using the anodic current waveform. Similar detection performances were achieved with both gases, and comparable energy resolutions were measured with both sources. As long as the purity of the gas was sufficient, no significant degradation of the measured energy was observed by increasing the pressure. At the highest operating pressure, an energy resolution better than 1.5% full-width at half-maximum (FWHM) was obtained for both gaseous media, although optimal noise conditions were not reached.

Though diffusion models have been successfully applied to various image restoration (IR) tasks, their performance is sensitive to the choice of training datasets. Typically, diffusion models trained in specific datasets fail to recover images that have out-of-distribution degradations. To address this problem, this work leverages a capable vision-language model and a synthetic degradation pipeline to learn image restoration in the wild (wild IR). More specifically, all low-quality images are simulated with a synthetic degradation pipeline that contains multiple common degradations such as blur, resize, noise, and JPEG compression. Then we introduce robust training for a degradation-aware CLIP model to extract enriched image content features to assist high-quality image restoration. Our base diffusion model is the image restoration SDE (IR-SDE). Built upon it, we further present a posterior sampling strategy for fast noise-free image generation. We evaluate our model on both synthetic and real-world degradation datasets. Moreover, experiments on the unified image restoration task illustrate that the proposed posterior sampling improves image generation quality for various degradations., Comment: CVPRW 2024; Code: https://github.com/Algolzw/daclip-uir

Identifying and classifying underground utilities is an important task for efficient and effective urban planning and infrastructure maintenance. We present OpenTrench3D, a novel and comprehensive 3D Semantic Segmentation point cloud dataset, designed to advance research and development in underground utility surveying and mapping. OpenTrench3D covers a completely novel domain for public 3D point cloud datasets and is unique in its focus, scope, and cost-effective capturing method. The dataset consists of 310 point clouds collected across 7 distinct areas. These include 5 water utility areas and 2 district heating utility areas. The inclusion of different geographical areas and main utilities (water and district heating utilities) makes OpenTrench3D particularly valuable for inter-domain transfer learning experiments. We provide benchmark results for the dataset using three state-of-the-art semantic segmentation models, PointNeXt, PointVector and PointMetaBase. Benchmarks are conducted by training on data from water areas, fine-tuning on district heating area 1 and evaluating on district heating area 2. The dataset is publicly available. With OpenTrench3D, we seek to foster innovation and progress in the field of 3D semantic segmentation in applications related to detection and documentation of underground utilities as well as in transfer learning methods in general.

In this paper, we introduce T-DEED, a Temporal-Discriminability Enhancer Encoder-Decoder for Precise Event Spotting in sports videos. T-DEED addresses multiple challenges in the task, including the need for discriminability among frame representations, high output temporal resolution to maintain prediction precision, and the necessity to capture information at different temporal scales to handle events with varying dynamics. It tackles these challenges through its specifically designed architecture, featuring an encoder-decoder for leveraging multiple temporal scales and achieving high output temporal resolution, along with temporal modules designed to increase token discriminability. Leveraging these characteristics, T-DEED achieves SOTA performance on the FigureSkating and FineDiving datasets. Code is available at https://github.com/arturxe2/T-DEED.

In this paper, we introduce ASTRA, a Transformer-based model designed for the task of Action Spotting in soccer matches. ASTRA addresses several challenges inherent in the task and dataset, including the requirement for precise action localization, the presence of a long-tail data distribution, non-visibility in certain actions, and inherent label noise. To do so, ASTRA incorporates (a) a Transformer encoder-decoder architecture to achieve the desired output temporal resolution and to produce precise predictions, (b) a balanced mixup strategy to handle the long-tail distribution of the data, (c) an uncertainty-aware displacement head to capture the label variability, and (d) input audio signal to enhance detection of non-visible actions. Results demonstrate the effectiveness of ASTRA, achieving a tight Average-mAP of 66.82 on the test set. Moreover, in the SoccerNet 2023 Action Spotting challenge, we secure the 3rd position with an Average-mAP of 70.21 on the challenge set.

The ability to detect unfamiliar or unexpected images is essential for safe deployment of computer vision systems. In the context of classification, the task of detecting images outside of a model's training domain is known as out-of-distribution (OOD) detection. While there has been a growing research interest in developing post-hoc OOD detection methods, there has been comparably little discussion around how these methods perform when the underlying classifier is not trained on a clean, carefully curated dataset. In this work, we take a closer look at 20 state-of-the-art OOD detection methods in the (more realistic) scenario where the labels used to train the underlying classifier are unreliable (e.g. crowd-sourced or web-scraped labels). Extensive experiments across different datasets, noise types & levels, architectures and checkpointing strategies provide insights into the effect of class label noise on OOD detection, and show that poor separation between incorrectly classified ID samples vs. OOD samples is an overlooked yet important limitation of existing methods. Code: https://github.com/glhr/ood-labelnoise, Comment: Accepted at CVPR 2024

"Excellence gaps" are the disparities in advanced academic performance that exist between student groups. These gaps have important implications for both academic equity and American economic competitiveness, as the most lucrative jobs often go to those who perform at the highest levels. Although considerable work has evaluated how and why these excellence gaps occur, what's not been examined closely is what excellence gaps look like for students of different races/ethnicities within the same socioeconomic group. This new report uses National Assessment of Educational Progress (NAEP) data on eighth graders over the last two decades to trace the performance of America's highest-achieving students by both race/ethnicity and socioeconomic status (SES). Specifically, Fordham's research associate Meredith Coffey and national research director Adam Tyner examine: (1) The extent to which racial/ethnic excellence gaps can be explained by differences in SES; (2) Whether excellence gaps still exist when racial/ethnic groups are compared within the same socioeconomic groups; and (3) How excellence gaps by race/ethnicity and SES have shifted over the past two decades, including since the COVID-19 pandemic. This new analysis reveals five key findings and several implications for education leaders and policymakers who want to increase the numbers of students from all backgrounds whose academic performance rises to the top. [Foreword written by Michael J. Petrilli and Amber M. Northern.]

Historically, many American students from poor families have been trapped in sorely underfunded public schools. The conventional wisdom suggests that school funding remains unequal across low- and high-income schools and that equal funding equates to equitable resources for students. This brief challenges the notion that economically disadvantaged students receive less funding than other students, with implications for equalizing classroom resources and optimizing other social policies. Specifically, this brief addresses the following ideas that have become conventional wisdom in some quarters: (1) "Students from traditionally disadvantaged backgrounds attend poorly funded schools."; (2) "Better school funding makes a difference to students from low-income families."; (3) "School funding remains unequal across low- and high-income schools."; (4) "Eliminating the SES gaps means racial/ethnic funding gaps disappear, too."; (5) "Equal funding means equitable resources for students."; and (6) "Even if school funding is equal, it's not adequate to meet students' needs."