Search

Open-source multimodal large language models (MLLMs) have shown significant potential in a broad range of multimodal tasks. However, their reasoning capabilities remain constrained by existing instruction-tuning datasets, which were predominately repurposed from academic datasets such as VQA, AI2D, and ChartQA. These datasets target simplistic tasks, and only provide phrase-level answers without any intermediate rationales. To address these challenges, we introduce a scalable and cost-effective method to construct a large-scale multimodal instruction-tuning dataset with rich intermediate rationales designed to elicit CoT reasoning. Using only open models, we create a dataset containing 12M instruction-response pairs to cover diverse, reasoning-intensive tasks with detailed and faithful rationales. Experiments demonstrate that training MLLMs on this dataset significantly improves reasoning capabilities, achieving state-of-the-art performance on benchmarks such as MathVerse (+8.1%), MMMU-Pro (+7%), and MuirBench (+13.3%). Additionally, the model demonstrates notable improvements of up to 4% on non-reasoning-based benchmarks. Ablation studies further highlight the importance of key components, such as rewriting and self-filtering, in the dataset construction process.

Anomalies such as redundant, inconsistent, contradictory, and deficient values in a Knowledge Graph (KG) are unavoidable, as these graphs are often curated manually, or extracted using machine learning and natural language processing techniques. Therefore, anomaly detection is a task that can enhance the quality of KGs. In this paper, we propose SEKA (SEeking Knowledge graph Anomalies), an unsupervised approach for the detection of abnormal triples and entities in KGs. SEKA can help improve the correctness of a KG whilst retaining its coverage. We propose an adaption of the Path Rank Algorithm (PRA), named the Corroborative Path Rank Algorithm (CPRA), which is an efficient adaptation of PRA that is customized to detect anomalies in KGs. Furthermore, we also present TAXO (TAXOnomy of anomaly types in KGs), a taxonomy of possible anomaly types that can occur in a KG. This taxonomy provides a classification of the anomalies discovered by SEKA with an extensive discussion of possible data quality issues in a KG. We evaluate both approaches using the four real-world KGs YAGO-1, KBpedia, Wikidata, and DSKG to demonstrate the ability of SEKA and TAXO to outperform the baselines.

Developments in machine learning and computing power suggest that artificial general intelligence is within reach. This raises the question of artificial consciousness: if a computer were to be functionally equivalent to a human, being able to do all we do, would it experience sights, sounds, and thoughts, as we do when we are conscious? Answering this question in a principled manner can only be done on the basis of a theory of consciousness that is grounded in phenomenology and that states the necessary and sufficient conditions for any system, evolved or engineered, to support subjective experience. Here we employ Integrated Information Theory (IIT), which provides principled tools to determine whether a system is conscious, to what degree, and the content of its experience. We consider pairs of systems constituted of simple Boolean units, one of which -- a basic stored-program computer -- simulates the other with full functional equivalence. By applying the principles of IIT, we demonstrate that (i) two systems can be functionally equivalent without being phenomenally equivalent, and (ii) that this conclusion is not dependent on the simulated system's function. We further demonstrate that, according to IIT, it is possible for a digital computer to simulate our behavior, possibly even by simulating the neurons in our brain, without replicating our experience. This contrasts sharply with computational functionalism, the thesis that performing computations of the right kind is necessary and sufficient for consciousness.

A simple model-free and distribution-free statistic, the functional relationship between the number of "effective" degrees of freedom and portfolio size, or N*(N), is used to discriminate between two alternative models for the correlation of daily cryptocurrency returns within a retail universe of defined by the list of tradable assets available to account holders at the Robinhood brokerage. The average pairwise correlation between daily cryptocurrency returns is found to be high (of order 60%) and the data collected supports description of the cross-section of returns by a simple isotropic correlation model distinct from a decomposition into a linear factor model with additive noise with high confidence. This description appears to be relatively stable through time., Comment: 15 pages, 4 figures, 2 tables

Given the increasing use of synthetic data in language model (LM) post-training, an LM's ability to generate high-quality data has become nearly as crucial as its ability to solve problems directly. While prior works have focused on developing effective data generation methods, they lack systematic comparison of different LMs as data generators in a unified setting. To address this gap, we propose AgoraBench, a benchmark that provides standardized settings and metrics to evaluate LMs' data generation abilities. Through synthesizing 1.26 million training instances using 6 LMs and training 99 student models, we uncover key insights about LMs' data generation capabilities. First, we observe that LMs exhibit distinct strengths. For instance, GPT-4o excels at generating new problems, while Claude-3.5-Sonnet performs better at enhancing existing ones. Furthermore, our analysis reveals that an LM's data generation ability doesn't necessarily correlate with its problem-solving ability. Instead, multiple intrinsic features of data quality-including response quality, perplexity, and instruction difficulty-collectively serve as better indicators. Finally, we demonstrate that strategic choices in output format and cost-conscious model selection significantly impact data generation effectiveness., Comment: Work in Progress

We present numerous aspects of the evolution of the LP 890-9 (SPECULOOS-2/TOI-4306) planetary system, focusing on the likelihood that planet c can support life. We find that the host star reaches the main sequence in 1 Gyr and that planet c lies close to the inner boundary of the habitable zone. We find the magma ocean stage can last up to 50 Myr, remove 8 Earth-oceans of water, and leave up to 2000 bars of oxygen in the atmosphere. However, if the planet forms with a hydrogen envelope as small as 0.1 Earth-masses, no water will be lost during the star's pre-main sequence phase from thermal escape processes. We find that the planets are unlikely to be in a 3:1 mean motion resonance and that both planets tidally circularize within 0.5 Gyr when tidal dissipation is held constant. However, if tidal dissipation is a function of mantle temperature and rheology, then we find that planet c's orbit may require more than 7 Gyr to circularize, during which time tidal heating may reach hundreds of terawatts. We thus conclude that the habitability of planet c depends most strongly on the initial volatile content and internal properties, but no data yet preclude the viability of an active biosphere on the planet., Comment: 16 pages, 9 figures, accepted to PSJ

Cross-device Federated Analytics (FA) is a distributed computation paradigm designed to answer analytics queries about and derive insights from data held locally on users' devices. On-device computations combined with other privacy and security measures ensure that only minimal data is transmitted off-device, achieving a high standard of data protection. Despite FA's broad relevance, the applicability of existing FA systems is limited by compromised accuracy; lack of flexibility for data analytics; and an inability to scale effectively. In this paper, we describe our approach to combine privacy, scalability, and practicality to build and deploy a system that overcomes these limitations. Our FA system leverages trusted execution environments (TEEs) and optimizes the use of on-device computing resources to facilitate federated data processing across large fleets of devices, while ensuring robust, defensible, and verifiable privacy safeguards. We focus on federated analytics (statistics and monitoring), in contrast to systems for federated learning (ML workloads), and we flag the key differences.

We show that the use of 1,3-diaminopropane (DAP) as a chemical modifier at the perovskite/electron-transport layer (ETL) interface enhances the power conversion efficiency (PCE) of 1.7 eV bandgap FACs mixed-halide perovskite single-junction cells, primarily by boosting the open-circuit voltage (VOC) from 1.06 V to 1.15 V. Adding a post-processing annealing step after C60 evaporation, further improves the fill factor (FF) by 20% from the control to the DAP + post-annealing devices. Using hyperspectral photoluminescence microscopy, we demonstrate that annealing helps improve compositional homogeneity at the top and bottom interfaces of the solar cell, which prevents detrimental bandgap pinning in the devices and improves C60 adhesion. Using time-of-flight secondary ion mass spectrometry, we show that DAP reacts with formamidinium present near the surface of the perovskite lattice to form a larger molecular cation, 1,4,5,6-tetrahydropyrimidinium (THP) that remains at the interface. Combining the use of DAP and the annealing of C60 interface, we fabricate Si-perovskite tandems with PCE of 25.29%, compared to 23.26% for control devices. Our study underscores the critical role of chemical reactivity and thermal post-processing of the C60/Lewis-base passivator interface in minimizing device losses and advancing solar-cell performance of wide-bandgap mixed-cation mixed-halide perovskite for tandem application.

The experimental realisation of unconventional superconductivity and charge order in kagome systems \textit{A}V$_3$Sb$_5$ is of critical importance. We conducted a highly systematic study of Cs(V$_{1-x}$Nb$_x$)$_3$Sb$_5$ with $x$=0.07 (Nb$_{0.07}$-CVS) by employing a unique combination of tuning parameters such as doping, hydrostatic pressure, magnetic fields, and depth, using muon spin rotation, AC susceptibility, and STM. We uncovered tunable magnetism in the normal state of Nb$_{0.07}$-CVS, which transitions to a time-reversal symmetry (TRS) breaking superconducting state under pressure. Specifically, our findings reveal that the bulk of Nb$_{0.07}$-CVS (at depths greater than 20 nm from the surface) experiences TRS breaking below $T^*=40~$K, lower than the charge order onset temperature, $T_\mathrm{CO}$ = 58 K. However, near the surface (within 20 nm from the surface), the TRS breaking signal doubles and onsets at $T_\mathrm{CO}$, indicating that Nb-doping decouples TRS breaking from charge order in the bulk but synchronises them near the surface. Additionally, Nb-doping raises the superconducting critical temperature $T_\mathrm{C}$ from 2.5 K to 4.4 K. Applying hydrostatic pressure enhances both $T_\mathrm{C}$ and the superfluid density by a factor of two, with a critical pressure $p_\mathrm{cr}$ ${\simeq}$ 0.85 GPa, suggesting competition with charge order. Notably, above $p_\mathrm{cr}$, we observe nodeless electron pairing and weak internal fields below $T_\mathrm{C}$, indicating broken TRS in the superconducting state. Overall, these results demonstrate a highly unconventional normal state with a depth-tunable onset of TRS breaking at ambient pressure, a transition to TRS-breaking superconductivity under low hydrostatic pressure, and an unconventional scaling between $T_\mathrm{C}$ and the superfluid density.

Recent clinical results indicate that aberrant erythrocyte aggregation in hematological disorders is accompanied by endothelial damage and glycocalyx disruption, but the underlying biophysical mechanisms remain unclear. This study uses direct computational modeling to explore how red blood cell (RBC) aggregation impacts shear stress in small blood vessels, highlighting the increased risk of vascular damage. RBC aggregation creates a heterogeneous distribution, leading to variations in the cell-free layer thickness and fluctuating wall shear stress, especially near vessel walls. This effect aligns with experimental findings on endothelial disruption linked to RBC clustering near the wall, potentially reducing the protective glycocalyx layer. The power spectral density analysis of wall shear stress fluctuations reveals that, with RBC aggregation, there is a distinct peak near frequency f = 0.04, indicating increased fluctuations due to aggregated RBC clusters traveling close to the vessel wall. The presence of aberrant cells in blood disorders, modeled here by sickle cells, further amplifies these effects, as aggregation-enhanced margination drives sickle cells closer to vessel walls, exacerbating shear stress fluctuations and increasing the likelihood of vascular injury and inflammation. Simulations show that curved vascular geometry, with curvature accentuating RBC clustering near vessel walls, intensifies aggregation-induced wall shear stress fluctuations and increases the risk of vascular damage, particularly in sickle cell disease where sickle cells marginate closer to the wall.

A key task in managing distributed, sensitive data is to measure the extent to which a distribution changes. Understanding this drift can effectively support a variety of federated learning and analytics tasks. However, in many practical settings sharing such information can be undesirable (e.g., for privacy concerns) or infeasible (e.g., for high communication costs). In this work, we describe novel algorithmic approaches for estimating the KL divergence of data across federated models of computation, under differential privacy. We analyze their theoretical properties and present an empirical study of their performance. We explore parameter settings that optimize the accuracy of the algorithm catering to each of the settings; these provide sub-variations that are applicable to real-world tasks, addressing different context- and application-specific trust level requirements. Our experimental results confirm that our private estimators achieve accuracy comparable to a baseline algorithm without differential privacy guarantees., Comment: 28 pages, 5 figures

Space exploration to have the biosignatures of extraterrestrial life on different planets with oceans in our solar system and beyond requires the design and manufacturing of robust and reliable electronic systems that can be used for sensing, data processing, controlling motor/actuators, and communication while surviving an extreme environment. Commercial off the shelf (COTS) components cannot survive a long time in such harsh environments after being housed in a Warm Electronics Box, and any electronic system designed for such extreme conditions must be tailored to suit such operation. The presence of extremely cold temperatures and high radiation adversely affects the device parameters over time, i.e. the operation of electronic systems.

Scientific progress depends on researchers' ability to synthesize the growing body of literature. Can large language models (LMs) assist scientists in this task? We introduce OpenScholar, a specialized retrieval-augmented LM that answers scientific queries by identifying relevant passages from 45 million open-access papers and synthesizing citation-backed responses. To evaluate OpenScholar, we develop ScholarQABench, the first large-scale multi-domain benchmark for literature search, comprising 2,967 expert-written queries and 208 long-form answers across computer science, physics, neuroscience, and biomedicine. On ScholarQABench, OpenScholar-8B outperforms GPT-4o by 5% and PaperQA2 by 7% in correctness, despite being a smaller, open model. While GPT4o hallucinates citations 78 to 90% of the time, OpenScholar achieves citation accuracy on par with human experts. OpenScholar's datastore, retriever, and self-feedback inference loop also improves off-the-shelf LMs: for instance, OpenScholar-GPT4o improves GPT-4o's correctness by 12%. In human evaluations, experts preferred OpenScholar-8B and OpenScholar-GPT4o responses over expert-written ones 51% and 70% of the time, respectively, compared to GPT4o's 32%. We open-source all of our code, models, datastore, data and a public demo.

This work explores the relationship between state space methods and Koopman operator-based methods for predicting the time-evolution of nonlinear dynamical systems. We demonstrate that extended dynamic mode decomposition with dictionary learning (EDMD-DL), when combined with a state space projection, is equivalent to a neural network representation of the nonlinear discrete-time flow map on the state space. We highlight how this projection step introduces nonlinearity into the evolution equations, enabling significantly improved EDMD-DL predictions. With this projection, EDMD-DL leads to a nonlinear dynamical system on the state space, which can be represented in either discrete or continuous time. This system has a natural structure for neural networks, where the state is first expanded into a high dimensional feature space followed by a linear mapping which represents the discrete-time map or the vector field as a linear combination of these features. Inspired by these observations, we implement several variations of neural ordinary differential equations (ODEs) and EDMD-DL, developed by combining different aspects of their respective model structures and training procedures. We evaluate these methods using numerical experiments on chaotic dynamics in the Lorenz system and a nine-mode model of turbulent shear flow, showing comparable performance across methods in terms of short-time trajectory prediction, reconstruction of long-time statistics, and prediction of rare events. We also show that these methods provide comparable performance to a non-Markovian approach in terms of prediction of extreme events.

We report the discovery of three ultracompact binary white dwarf systems hosting accretion disks, with orbital periods of 7.95, 8.68, and 13.15 minutes. This significantly augments the population of mass-transferring binaries at the shortest periods, and provides the first evidence that accretors in ultracompacts can be dense enough to host accretion disks even below 10 minutes (where previously only direct-impact accretors were known). In the two shortest-period systems, we measured changes in the orbital periods driven by the combined effect of gravitational wave emission and mass transfer; we find $\dot{P}$ is negative in one case, and positive in the other. This is only the second system measured with a positive $\dot{P}$, and it the most compact binary known that has survived a period minimum. Using these systems as examples, we show how the measurement of $\dot{P}$ is a powerful tool in constraining the physical properties of binaries, e.g. the mass and mass-radius relation of the donor stars. We find that the chirp masses of ultracompact binaries at these periods seem to cluster around $\mathcal{M}_c \sim 0.3 M_\odot$, perhaps suggesting a common origin for these systems or a selection bias in electromagnetic discoveries. Our new systems are among the highest-amplitude known gravitational wave sources in the millihertz regime, providing exquisite opportunity for multi-messenger study with future space-based observatories such as \textit{LISA} and TianQin; we discuss how such systems provide fascinating laboratories to study the unique regime where the accretion process is mediated by gravitational waves., Comment: Accepted for publication in ApJ

A semiring generalises the notion of a ring, replacing the additive abelian group structure with that of a commutative monoid. In this paper, we study a notion positioned between a ring and a semiring -- a semiring whose additive monoid is a commutative inverse semigroup. These inverse semirings include some important classes of semirings, as well as some new motivating examples. We devote particular attention to the inverse semiring of bounded polynomials and argue for their computational significance. We then prove a number of fundamental results about inverse semirings, their modules and their ideals. Parts of the theory show strong similarities with rings, while other parts are akin to the theory of idempotent semirings or distributive lattices. We note in particular that downward-closed submodules are precisely kernels. We end by exploring a connection to the theory of E-unitary inverse semigroups., Comment: 16 pages

Transitioning from quantum computation on physical qubits to quantum computation on encoded, logical qubits can improve the error rate of operations, and will be essential for realizing valuable quantum computational advantages. Using a neutral atom quantum processor with 256 qubits, each an individual Ytterbium atom, we demonstrate the entanglement of 24 logical qubits using the distance-two [[4,2,2]] code, simultaneously detecting errors and correcting for lost qubits. We also implement the Bernstein-Vazirani algorithm with up to 28 logical qubits encoded in the [[4,1,2]] code, showing better-than-physical error rates. We demonstrate fault-tolerant quantum computation in our approach, guided by the proposal of Gottesman (2016), by performing repeated loss correction for both structured and random circuits encoded in the [[4,2,2]] code. Finally, since distance-two codes can correct qubit loss, but not other errors, we show repeated loss and error correction using the distance-three [[9,1,3]] Bacon-Shor code. These results begin to clear a path for achieving scientific quantum advantage with a programmable neutral atom quantum processor., Comment: 17 pages, 16 figures

Cardiovascular diseases (CVDs) remain the leading cause of mortality and morbidity worldwide. Both diagnosis and prognosis of these diseases benefit from high-quality imaging, which cardiac magnetic resonance imaging provides. CMR imaging requires lengthy acquisition times and multiple breath-holds for a complete exam, which can lead to patient discomfort and frequently results in image artifacts. In this work, we present a Low-rank tensor U-Net method (LowRank-CGNet) that rapidly reconstructs highly undersampled data with a variety of anatomy, contrast, and undersampling artifacts. The model uses conjugate gradient data consistency to solve for the spatial and temporal bases and employs a U-Net to further regularize the basis vectors. Currently, model performance is superior to a standard U-Net, but inferior to conventional compressed sensing methods. In the future, we aim to further improve model performance by increasing the U-Net size, extending the training duration, and dynamically updating the tensor rank for different anatomies., Comment: 11 pages, 6 figures

This white paper is the output of a multidisciplinary workshop in Nairobi (Nov 2023). Led by a cross-organisational team including Microsoft Research, NEPAD, Lelapa AI, and University of Oxford. The workshop brought together diverse thought-leaders from various sectors and backgrounds to discuss the implications of Generative AI for the future of work in Africa. Discussions centred around four key themes: Macroeconomic Impacts; Jobs, Skills and Labour Markets; Workers' Perspectives and Africa-Centris AI Platforms. The white paper provides an overview of the current state and trends of generative AI and its applications in different domains, as well as the challenges and risks associated with its adoption and regulation. It represents a diverse set of perspectives to create a set of insights and recommendations which aim to encourage debate and collaborative action towards creating a dignified future of work for everyone across Africa.

We construct a version of geodesic normal coordinates adapted to a submanifold of a pseudo-Riemannian manifold and show that the Taylor coefficients of the metric in these coordinates can be expressed as universal polynomials in the components of the covariant derivatives of the background curvature tensor and the covariant derivatives of the second fundamental form. We formulate a definition of natural submanifold tensors and show that these are linear combinations of contractions of covariant derivatives of the background curvature tensor and covariant derivatives of the second fundamental form. We also describe how this result gives a similar characterization of natural submanifold differential operators., Comment: 12 pages

In this short note the theory for multivariate asset allocation with elliptically symmetric distributions of returns, as developed in the author's prior work, is specialized to the case of returns drawn from a multivariate Laplace distribution. This analysis delivers a result closely, but not perfectly, consistent with the conjecture presented in the author's article Thinking Differently About Asset Allocation. The principal differences are due to the introduction of a term in the dimensionality of the problem, which was omitted from the conjectured solution, and a rescaling of the variance due to varying parameterizations of the univariate Laplace distribution., Comment: 9 pages, 1 figure

This note discusses some of the aspects of a model for the covariance of equity returns based on a simple "isotropic" structure in which all pairwise correlations are taken to be the same value. The effect of the structure on feasible values for the common correlation of returns and on the "effective degrees of freedom" within the equity cross-section are discussed, as well as the impact of this constraint on the asymptotic Normality of portfolio returns. An eigendecomposition of the covariance matrix is presented and used to partition variance into that from a common "market" factor and "non-diversifiable" idiosyncratic risk. A empirical analysis of the recent history of the returns of S&P 500 Index members is presented and compared to the expectations from both this model and linear factor models. This analysis supports the isotropic covariance model and does not seem to provide evidence in support of linear factor models. Analysis of portfolio selection under isotropic correlation is presented using mean-variance optimization for both heteroskedastic and homoskedastic cases. Portfolio selection for negative exponential utility maximizers is also discussed for the general case of distributions of returns with elliptical symmetry. The fact that idiosyncratic risk may not be removed by diversification in a model that the data supports undermines the basic premises of structures such as the C.A.P.M. and A.P.T. If the cross-section of equity returns is more accurately described by this structure then an inevitable consequence is that picking stocks is not a "pointless" activity, as the returns to residual risk would be non-zero., Comment: 24 pages, 5 figures, code is available on the author's personal GitHub repository, code executes in Google's Colab system and generates figures from live data downloaded from Yahoo! Finance

In the past few years, the improved sensitivity and cadence of wide-field optical surveys have enabled the discovery of several afterglows without associated detected gamma-ray bursts (GRBs). We present the identification, observations, and multiwavelength modeling of a recent such afterglow (AT2023lcr), and model three literature events (AT2020blt, AT2021any, and AT2021lfa) in a consistent fashion. For each event, we consider the following possibilities as to why a GRB was not observed: 1) the jet was off-axis; 2) the jet had a low initial Lorentz factor; and 3) the afterglow was the result of an on-axis classical GRB (on-axis jet with physical parameters typical of the GRB population), but the emission was undetected by gamma-ray satellites. We estimate all physical parameters using afterglowpy and Markov Chain Monte Carlo methods from emcee. We find that AT2023lcr, AT2020blt, and AT2021any are consistent with on-axis classical GRBs, and AT2021lfa is consistent with both on-axis low Lorentz factor ($\Gamma_0 \approx 5 - 13$) and off-axis ($\theta_\text{obs}=2\theta_\text{jet}$) high Lorentz factor ($\Gamma_0 \approx 100$) jets., Comment: 40 pages, 18 figures, 20 tables

We demonstrate an end-to-end technique for observing and characterizing massive black hole binary signals before they merge with the LISA space-based gravitational-wave observatory. Our method uses a zero-latency whitening filter, originally designed for rapidly observing compact binary mergers in ground-based observatories, to be able to observe signals with no additional latency due to filter length. We show that with minimal computational cost, we are able to reliably observe signals as early as 14 days premerger as long as the signal has accrued a signal-to-noise ratio of at least 8 in the LISA data. We also demonstrate that this method can be used to characterize the source properties, providing early estimates of the source's merger time, chirp mass, and sky localization. Early observation and characterization of massive black holes is crucial to enable the possibility of rapid multimessenger observations, and to ensure that LISA can enter a protected operating period when the merger signal arrives., Comment: 16 pages, 11 figures

In this second paper on the DECam deep drilling field (DDF) program we release 2,020 optical gri-band light curves for transients and variables in the extragalactic COSMOS and ELAIS fields based on time series observations with a 3-day cadence from semester 2021A through 2023A. In order to demonstrate the wide variety of time domain events detected by the program and encourage others to use the data set, we characterize the sample by presenting a brief analysis of the light curve parameters such as time span, amplitude, and peak brightness. We also present preliminary light curve categorizations, and identify potential stellar variables, active galactic nuclei, tidal disruption events, supernovae (such as Type Ia, Type IIP, superluminous, and gravitationally lensed supernovae), and fast transients. Where relevant, the number of identified transients is compared to the predictions of the original proposal. We also discuss the challenges of analyzing DDF data in the context of the upcoming Vera C. Rubin Observatory and its Legacy Survey of Space and Time, which will include DDFs. Images from the DECam DDF program are available without proprietary period and the light curves presented in this work are publicly available for analysis., Comment: 17 pages, 18 figures, 1 table; accepted for publication in MNRAS

We study the superconducting critical fields of two-dimensional electron systems at (111)$\mathrm{LaAlO_3/KTaO_3}$ interfaces as a function of electrostatic back-gating. Our work reveals inplane critical fields of unprecedented magnitudes at oxide interfaces. By comparing the critical fields in-plane and out-of-plane we discover an extremely anisotropic superconductor with an effective thickness below 1 nm and a 12-fold violation of the Chandrasekhar-Clogston paramagnetic limit. The analysis of magneto-transport indicates that the enhancement of the critical fields is due to an exceptionally thin superconducting layer and to a paramagnetic susceptibility suppressed by spin-orbit scattering.

Statistical heterogeneity is a measure of how skewed the samples of a dataset are. It is a common problem in the study of differential privacy that the usage of a statistically heterogeneous dataset results in a significant loss of accuracy. In federated scenarios, statistical heterogeneity is more likely to happen, and so the above problem is even more pressing. We explore the three most promising ways to measure statistical heterogeneity and give formulae for their accuracy, while simultaneously incorporating differential privacy. We find the optimum privacy parameters via an analytic mechanism, which incorporates root finding methods. We validate the main theorems and related hypotheses experimentally, and test the robustness of the analytic mechanism to different heterogeneity levels. The analytic mechanism in a distributed setting delivers superior accuracy to all combinations involving the classic mechanism and/or the centralized setting. All measures of statistical heterogeneity do not lose significant accuracy when a heterogeneous sample is used., Comment: 26 pages, 6 tables, 1 figure

We construct four-variable $p$-adic $L$-functions for the spin Galois representation of a Siegel modular form of genus 2 twisted by the Galois representation of a cuspidal modular form as the modular forms vary in Coleman families. The main ingredient is the construction of a space of nearly overconvergent modular forms in the coherent cohomology of the Siegel threefold, extending the spaces of overconvergent modular forms appearing in higher Coleman theory. In addition to this, we construct $p$-adic distributions interpolating the Gan-Gross-Prasad automorphic periods for $(\mathrm{GSpin}(5), \mathrm{GSpin}(4))$ which, conditional on the local and global Gan-Gross-Prasad conjectures for this pair of groups, provides a construction of "square-root" $p$-adic $L$-functions for $\mathrm{GSp}(4) \times \mathrm{GL}(2) \times \mathrm{GL}(2)$ as the automorphic forms vary in Coleman families., Comment: 54 pages

This article proposes a new integration of linguistic anthropology and machine learning (ML) around convergent interests in both the underpinnings of language and making language technologies more socially responsible. While linguistic anthropology focuses on interpreting the cultural basis for human language use, the ML field of interpretability is concerned with uncovering the patterns that Large Language Models (LLMs) learn from human verbal behavior. Through the analysis of a conversation between a human user and an LLM-powered chatbot, we demonstrate the theoretical feasibility of a new, conjoint field of inquiry, cultural interpretability (CI). By focusing attention on the communicative competence involved in the way human users and AI chatbots co-produce meaning in the articulatory interface of human-computer interaction, CI emphasizes how the dynamic relationship between language and culture makes contextually sensitive, open-ended conversation possible. We suggest that, by examining how LLMs internally "represent" relationships between language and culture, CI can: (1) provide insight into long-standing linguistic anthropological questions about the patterning of those relationships; and (2) aid model developers and interface designers in improving value alignment between language models and stylistically diverse speakers and culturally diverse speech communities. Our discussion proposes three critical research axes: relativity, variation, and indexicality., Comment: Accepted for publication in Big Data & Society, November 2, 2024

The Legacy Survey of Space and Time (LSST) at Vera C. Rubin Observatory is planned to begin in the Fall of 2025. The LSST survey cadence has been designed via a community-driven process regulated by the Survey Cadence Optimization Committee (SCOC), which recommended up to 3% of the observing time to carry out Target of Opportunity (ToO) observations. Experts from the scientific community, Rubin Observatory personnel, and members of the SCOC were brought together to deliver a recommendation for the implementation of the ToO program during a workshop held in March 2024. Four main science cases were identified: gravitational wave multi-messenger astronomy, high energy neutrinos, Galactic supernovae, and small potentially hazardous asteroids possible impactors. Additional science cases were identified and briefly addressed in the documents, including lensed or poorly localized gamma-ray bursts and twilight discoveries. Trigger prioritization, automated response, and detailed strategies were discussed for each science case. This document represents the outcome of the Rubin ToO 2024 workshop, with additional contributions from members of the Rubin Science Collaborations. The implementation of the selection criteria and strategies presented in this document has been endorsed in the SCOC Phase 3 Recommendations document (PSTN-056). Although the ToO program is still to be finalized, this document serves as a baseline plan for ToO observations with the Rubin Observatory.

When comparing the linguistic capabilities of language models (LMs) with humans using LM probabilities, factors such as the length of the sequence and the unigram frequency of lexical items have a significant effect on LM probabilities in ways that humans are largely robust to. Prior works in comparing LM and human acceptability judgments treat these effects uniformly across models, making a strong assumption that models require the same degree of adjustment to control for length and unigram frequency effects. We propose MORCELA, a new linking theory between LM scores and acceptability judgments where the optimal level of adjustment for these effects is estimated from data via learned parameters for length and unigram frequency. We first show that MORCELA outperforms a commonly used linking theory for acceptability--SLOR (Pauls and Klein, 2012; Lau et al. 2017)--across two families of transformer LMs (Pythia and OPT). Furthermore, we demonstrate that the assumed degrees of adjustment in SLOR for length and unigram frequency overcorrect for these confounds, and that larger models require a lower relative degree of adjustment for unigram frequency, though a significant amount of adjustment is still necessary for all models. Finally, our subsequent analysis shows that larger LMs' lower susceptibility to frequency effects can be explained by an ability to better predict rarer words in context.

A spectral minimal partition of a manifold is a decomposition into disjoint open sets that minimizes a spectral energy functional. While it is known that bipartite minimal partitions correspond to nodal partitions of Courant-sharp Laplacian eigenfunctions, the non-bipartite case is much more challenging. In this paper, we unify the bipartite and non-bipartite settings by defining a modified Laplacian operator and proving that the nodal partitions of its eigenfunctions are exactly the critical points of the spectral energy functional. Moreover, we prove that the Morse index of a critical point equals the nodal deficiency of the corresponding eigenfunction. Some striking consequences of our main result are: 1) in the bipartite case, every local minimum of the energy functional is in fact a global minimum; 2) in the non-bipartite case, every local minimum of the energy functional minimizes within a certain topological class of partitions. Our results are valid for partitions with non-smooth boundaries; this introduces considerable technical challenges, which are overcome using delicate approximation arguments in the Sobolev space $H^{1/2}$.

Class imbalance and label noise are pervasive in large-scale datasets, yet much of machine learning research assumes well-labeled, balanced data, which rarely reflects real world conditions. Existing approaches typically address either label noise or class imbalance in isolation, leading to suboptimal results when both issues coexist. In this work, we propose Conformal-in-the-Loop (CitL), a novel training framework that addresses both challenges with a conformal prediction-based approach. CitL evaluates sample uncertainty to adjust weights and prune unreliable examples, enhancing model resilience and accuracy with minimal computational cost. Our extensive experiments include a detailed analysis showing how CitL effectively emphasizes impactful data in noisy, imbalanced datasets. Our results show that CitL consistently boosts model performance, achieving up to a 6.1% increase in classification accuracy and a 5.0 mIoU improvement in segmentation. Our code is publicly available: CitL., Comment: Under Review

The worlds of Data Science (including big and/or federated data, machine learning, etc) and Astrophysics started merging almost two decades ago. For instance, around 2005, international initiatives such as the Virtual Observatory framework rose to standardize the way we publish and transfer data, enabling new tools such as VOSA (SED Virtual Observatory Analyzer) to come to existence and remain relevant today. More recently, new facilities like the Vera Rubin Observatory, serve as motivation to develop efficient and extremely fast (very often deep learning based) methodologies in order to fully exploit the informational content of the vast Legacy Survey of Space and Time (LSST) dataset. However, fundamental changes in the way we explore and analyze data cannot permeate in the "astrophysical sociology and idiosyncrasy" without adequate training. In this talk, I will focus on one specific initiative that has been extremely successful and is based on "learning by doing": the La Serena School for Data Science. I will also briefly touch on a different successful approach: a series of schools organized by the Spanish Virtual Observatory. The common denominator among the two kinds of schools is to present the students with real scientific problems that benefit from the concepts / methodologies taught. On the other hand, the demographics targeted by both initiatives vary significantly and can represent examples of two "flavours" to be followed by others., Comment: Invited talk IAU-GA 2024, FM 7

We use Magnetospheric Multiscale (MMS) mission data to investigate electron-scale energy transfer due to lower hybrid drift waves during magnetopause reconnection. We analyze waves observed in an electron-scale plasma mixing layer at the edge of the magnetospheric outflow. Using high-resolution 7.5 ms electron moments, we obtain an electron current density with a Nyquist frequency of ~66 Hz, sufficient to resolve most of the lower hybrid drift wave power observed in the event. We then employ wavelet analysis to evaluate dJ.dE, which accounts for the phase differences between the fluctuating quantities. The analysis shows that the energy exchange is localized within the plasma mixing layer, and it is highly fluctuating, with energy bouncing between waves and electrons throughout the analyzed time and frequency range. However, the cumulative sum over time indicates a net energy transfer from the waves to electrons. We observe an anomalous electron flow toward the magnetosphere, consistent with diffusion and electron mixing. These results indicate that waves and electrons interact dynamically to dissipate the excess internal energy accumulated by sharp density gradients. We conclude that the electron temperature profile within the plasma mixing layer is produced by a combination of electron diffusion across the layer, as well as heating by large-scale parallel potential and lower hybrid drift waves., Comment: 23 pages, 6 figures. Submitted to Journal of Geophysical Research: Space Physics

We identify symmetries in a broad class of vector-like confining dark sectors that forbid the leading electromagnetic moments that would ordinarily mediate dark baryon scattering with the Standard Model. The absence of these operators implies dark baryon dark matter has much smaller cross sections for elastic scattering off nuclei, leading to suppressed direct detection signals. In the confined description, we identify an ``$\mathcal{H}$-parity'' symmetry that exists in any dark sector with dark quarks transforming under a vector-like representation of a new confining SU($N_c$) gauge theory as well as a vector-like representation of the electroweak group SU(2)$_L$. The parity is independent of $N_c$ and $N_f$, though it is essential that the dark quarks are neutral under hypercharge. This parity forbids dark hadron electric and magnetic dipole moments, charge radius, and anapole moment, while permitting dimension-7 operators that include polarizability, electroweak loop-induced interactions, as well as lower dimensional electromagnetic $\textit{transition}$ moments between different neutral dark baryon states. We work out an explicit example, $N_c=N_f=3$, that is the most minimal theory with fermionic dark baryons. In this specific model, we use the non-relativistic quark model to show the magnetic dipole moment and charge radius vanish while the transition moments are non-zero, consistent with $\mathcal{H}$-parity. We discuss the implications of a suppressed direct detection signal, emphasizing that this broad class of models provide a well-motivated target for future colliders., Comment: 28+14 pages, 1+1 figures, 2 tables

The ability to resolve photon numbers is crucial in quantum information science and technology, driving the development of detectors with intrinsic photon-number resolving (PNR) capabilities. Although transition edge sensors represent the state-of-the-art in PNR performance, superconducting nanowire single-photon detectors (SNSPDs) offer superior efficiency, speed, noise reduction, and timing precision. Directly inferring photon numbers, however, has only recently become feasible due to advances in readout technology. Despite this, photon-number discrimination remains constrained by the nanowire's electrical properties and readout jitter. In this work, we employ waveguide-integrated SNSPDs and time-resolved measurements to explore how the nanowire kinetic inductance and system jitter affect PNR capabilities. By analyzing the latency time of the photon detection, we can resolve changes in the rising edge of the detection pulse. We find that lower jitter as well as increased kinetic inductance enhances the pulse separation for different photon numbers and improves the PNR capability. Enhancing the kinetic inductance from 165 nH to 872 nH improves PNR quality by 12%, 31% and 23% over the first three photon numbers, though at the cost of reducing the detector's count rate from 165 Mcps to 19 Mcps. Our findings highlight the trade-off between PNR resolution and detector speed., Comment: 21 pages, 7 figures

A strong first-order electroweak phase transition (SFOEWPT) is essential for explaining baryogenesis and for potentially generating observable gravitational waves. This study investigates the potential of a high-energy muon collider to examine the occurrence of SFOEWPT within the context of a Standard Model extended by a real scalar singlet (xSM). We analyzed all possible decay modes of the singlet to constrain the valid parameter space of SFOEWPT, which was extracted numerically at different renormalization scales to account for theoretical uncertainties, thereby determining the sensitivity of a muon collider to the production and decay channels of novel heavy scalar particles that emerge in the xSM. The findings demonstrate that a 3 TeV muon collider can directly examine the nature of electroweak symmetry breaking by efficiently detecting novel scalar particles associated with a first-order electroweak phase transition through jet-rich final states, thus complementing the indirect constraints from gravitational wave experiments.

Trust and interpretability are crucial for the use of Artificial Intelligence (AI) in scientific research, but current models often operate as black boxes offering limited transparency and justifications for their outputs. We introduce AiSciVision, a framework that specializes Large Multimodal Models (LMMs) into interactive research partners and classification models for image classification tasks in niche scientific domains. Our framework uses two key components: (1) Visual Retrieval-Augmented Generation (VisRAG) and (2) domain-specific tools utilized in an agentic workflow. To classify a target image, AiSciVision first retrieves the most similar positive and negative labeled images as context for the LMM. Then the LMM agent actively selects and applies tools to manipulate and inspect the target image over multiple rounds, refining its analysis before making a final prediction. These VisRAG and tooling components are designed to mirror the processes of domain experts, as humans often compare new data to similar examples and use specialized tools to manipulate and inspect images before arriving at a conclusion. Each inference produces both a prediction and a natural language transcript detailing the reasoning and tool usage that led to the prediction. We evaluate AiSciVision on three real-world scientific image classification datasets: detecting the presence of aquaculture ponds, diseased eelgrass, and solar panels. Across these datasets, our method outperforms fully supervised models in low and full-labeled data settings. AiSciVision is actively deployed in real-world use, specifically for aquaculture research, through a dedicated web application that displays and allows the expert users to converse with the transcripts. This work represents a crucial step toward AI systems that are both interpretable and effective, advancing their use in scientific research and scientific discovery.

Quasars are bright active galactic nuclei powered by the accretion of matter around supermassive black holes at the center of galaxies. Their stochastic brightness variability depends on the physical properties of the accretion disk and black hole. The upcoming Rubin Observatory Legacy Survey of Space and Time (LSST) is expected to observe tens of millions of quasars, so there is a need for efficient techniques like machine learning that can handle the large volume of data. Quasar variability is believed to be driven by an X-ray corona, which is reprocessed by the accretion disk and emitted as UV/optical variability. We are the first to introduce an auto-differentiable simulation of the accretion disk and reprocessing. We use the simulation as a direct component of our neural network to jointly model the driving variability and reprocessing to fit simulated LSST 10-year quasar light curves. The driving variability is reconstructed using a latent stochastic differential equation, a physically motivated, generative deep learning method that can model continuous-time stochastic dynamics. By embedding these physical processes into our network, we achieve a model that is more robust and interpretable. We also use transformers to scale our model to tens of millions of parameters. We demonstrate how our model outperforms a Gaussian process regression baseline and can infer accretion disk parameters and time delays between wavebands, even for out-of-distribution driving signals. Our approach provides a powerful and scalable framework that can be adapted to solve other inverse problems in multivariate time series with irregular sampling., Comment: 31 pages, 19 figures

Recent advances in multi-agent reinforcement learning (MARL) have created opportunities to solve complex real-world tasks. Cybersecurity is a notable application area, where defending networks against sophisticated adversaries remains a challenging task typically performed by teams of security operators. In this work, we explore novel MARL strategies for building autonomous cyber network defenses that address challenges such as large policy spaces, partial observability, and stealthy, deceptive adversarial strategies. To facilitate efficient and generalized learning, we propose a hierarchical Proximal Policy Optimization (PPO) architecture that decomposes the cyber defense task into specific sub-tasks like network investigation and host recovery. Our approach involves training sub-policies for each sub-task using PPO enhanced with domain expertise. These sub-policies are then leveraged by a master defense policy that coordinates their selection to solve complex network defense tasks. Furthermore, the sub-policies can be fine-tuned and transferred with minimal cost to defend against shifts in adversarial behavior or changes in network settings. We conduct extensive experiments using CybORG Cage 4, the state-of-the-art MARL environment for cyber defense. Comparisons with multiple baselines across different adversaries show that our hierarchical learning approach achieves top performance in terms of convergence speed, episodic return, and several interpretable metrics relevant to cybersecurity, including the fraction of clean machines on the network, precision, and false positives on recoveries., Comment: 9 pages, 7 figures, AAMAS preprint

Accelerating research on Large Multimodal Models (LMMs) in non-English languages is crucial for enhancing user experiences across broader populations. In this paper, we introduce JMMMU (Japanese MMMU), the first large-scale Japanese benchmark designed to evaluate LMMs on expert-level tasks based on the Japanese cultural context. To facilitate comprehensive culture-aware evaluation, JMMMU features two complementary subsets: (i) culture-agnostic (CA) subset, where the culture-independent subjects (e.g., Math) are selected and translated into Japanese, enabling one-to-one comparison with its English counterpart MMMU; and (ii) culture-specific (CS) subset, comprising newly crafted subjects that reflect Japanese cultural context. Using the CA subset, we observe performance drop in many LMMs when evaluated in Japanese, which is purely attributable to language variation. Using the CS subset, we reveal their inadequate Japanese cultural understanding. Further, by combining both subsets, we identify that some LMMs perform well on the CA subset but not on the CS subset, exposing a shallow understanding of the Japanese language that lacks depth in cultural understanding. We hope this work will not only help advance LMM performance in Japanese but also serve as a guideline to create high-standard, culturally diverse benchmarks for multilingual LMM development. The project page is https://mmmu-japanese-benchmark.github.io/JMMMU/., Comment: Project page: https://mmmu-japanese-benchmark.github.io/JMMMU/

We present the results of a search for gravitational-wave transients associated with core-collapse supernova SN 2023ixf, which was observed in the galaxy Messier 101 via optical emission on 2023 May 19th, during the LIGO-Virgo-KAGRA 15th Engineering Run. We define a five-day on-source window during which an accompanying gravitational-wave signal may have occurred. No gravitational waves have been identified in data when at least two gravitational-wave observatories were operating, which covered $\sim 14\%$ of this five-day window. We report the search detection efficiency for various possible gravitational-wave emission models. Considering the distance to M101 (6.7 Mpc), we derive constraints on the gravitational-wave emission mechanism of core-collapse supernovae across a broad frequency spectrum, ranging from 50 Hz to 2 kHz where we assume the GW emission occurred when coincident data are available in the on-source window. Considering an ellipsoid model for a rotating proto-neutron star, our search is sensitive to gravitational-wave energy $1 \times 10^{-5} M_{\odot} c^2$ and luminosity $4 \times 10^{-5} M_{\odot} c^2/\text{s}$ for a source emitting at 50 Hz. These constraints are around an order of magnitude more stringent than those obtained so far with gravitational-wave data. The constraint on the ellipticity of the proto-neutron star that is formed is as low as $1.04$, at frequencies above $1200$ Hz, surpassing results from SN 2019ejj., Comment: Main paper: 6 pages, 4 figures and 1 table. Total with appendices: 20 pages, 4 figures, and 1 table

Web browsers are a portal to the internet, where much of human activity is undertaken. Thus, there has been significant research work in AI agents that interact with the internet through web browsing. However, there is also another interface designed specifically for machine interaction with online content: application programming interfaces (APIs). In this paper we ask -- what if we were to take tasks traditionally tackled by browsing agents, and give AI agents access to APIs? To do so, we propose two varieties of agents: (1) an API-calling agent that attempts to perform online tasks through APIs only, similar to traditional coding agents, and (2) a Hybrid Agent that can interact with online data through both web browsing and APIs. In experiments on WebArena, a widely-used and realistic benchmark for web navigation tasks, we find that API-based agents outperform web browsing agents. Hybrid Agents out-perform both others nearly uniformly across tasks, resulting in a more than 20.0% absolute improvement over web browsing alone, achieving a success rate of 35.8%, achiving the SOTA performance among task-agnostic agents. These results strongly suggest that when APIs are available, they present an attractive alternative to relying on web browsing alone., Comment: 24 pages, 6 figures

Despite recent advances in multimodal large language models (MLLMs), their development has predominantly focused on English- and western-centric datasets and tasks, leaving most of the world's languages and diverse cultural contexts underrepresented. This paper introduces Pangea, a multilingual multimodal LLM trained on PangeaIns, a diverse 6M instruction dataset spanning 39 languages. PangeaIns features: 1) high-quality English instructions, 2) carefully machine-translated instructions, and 3) culturally relevant multimodal tasks to ensure cross-cultural coverage. To rigorously assess models' capabilities, we introduce PangeaBench, a holistic evaluation suite encompassing 14 datasets covering 47 languages. Results show that Pangea significantly outperforms existing open-source models in multilingual settings and diverse cultural contexts. Ablation studies further reveal the importance of English data proportions, language popularity, and the number of multimodal training samples on overall performance. We fully open-source our data, code, and trained checkpoints, to facilitate the development of inclusive and robust multilingual MLLMs, promoting equity and accessibility across a broader linguistic and cultural spectrum., Comment: 52 pages, 27 figures

Vision-language models (VLMs) have made significant progress in recent visual-question-answering (VQA) benchmarks that evaluate complex visio-linguistic reasoning. However, are these models truly effective? In this work, we show that VLMs still struggle with natural images and questions that humans can easily answer, which we term natural adversarial samples. We also find it surprisingly easy to generate these VQA samples from natural image-text corpora using off-the-shelf models like CLIP and ChatGPT. We propose a semi-automated approach to collect a new benchmark, NaturalBench, for reliably evaluating VLMs with 10,000 human-verified VQA samples. Crucially, we adopt a $\textbf{vision-centric}$ design by pairing each question with two images that yield different answers, preventing blind solutions from answering without using the images. This makes NaturalBench more challenging than previous benchmarks that can be solved with commonsense priors. We evaluate 53 state-of-the-art VLMs on NaturalBench, showing that models like LLaVA-OneVision, Cambrian-1, Llama3.2-Vision, Molmo, Qwen2-VL, and even GPT-4o lag 50%-70% behind human performance (over 90%). We analyze why NaturalBench is hard from two angles: (1) Compositionality: Solving NaturalBench requires diverse visio-linguistic skills, including understanding attribute bindings, object relationships, and advanced reasoning like logic and counting. To this end, unlike prior work that uses a single tag per sample, we tag each NaturalBench sample with 1 to 8 skill tags for fine-grained evaluation. (2) Biases: NaturalBench exposes severe biases in VLMs, as models often choose the same answer regardless of the image. Lastly, we apply our benchmark curation method to diverse data sources, including long captions (over 100 words) and non-English languages like Chinese and Hindi, highlighting its potential for dynamic evaluations of VLMs., Comment: Accepted to NeurIPS 24; We open-source our dataset at: https://huggingface.co/datasets/BaiqiL/NaturalBench ; Project page at: https://linzhiqiu.github.io/papers/naturalbench/

A dedicated gamma-ray burst (GRB) afterglow observing program was performed between 2007 and 2016 with GROND, a seven-channel optical and near-infrared imager at the 2.2m telescope of the Max-Planck Society at ESO/La Silla. In this first of a series of papers, we describe the GRB observing plan, providing first readings of all so far unpublished GRB afterglow measurements and some observing statistics. In total, we observed 514 GRBs with GROND, including 434 Swift-detected GRBs, representing 81\% of the observable Swift sample. For GROND-observations within 30 min of the GRB trigger, the optical/NIR afterglow detection rate is 81\% for long- and 57\% for short-duration GRBs. We report the discovery of ten new GRB afterglows plus one candidate, along with redshift estimates (partly improved) for four GRBs and new host detections for seven GRBs. We identify the (already known) afterglow of GRB 140209A as the sixth GRB exhibiting a 2175 Angstroem dust feature. As a side result, we identified two blazars, with one at a redshift of z=3.8 (in the GRB 131209A field)., Comment: 20 pages, 31 figures, accepted for publication in Astronomy & Astrophysics

Text-rich visual understanding-the ability to process environments where dense textual content is integrated with visuals-is crucial for multimodal large language models (MLLMs) to interact effectively with structured environments. To enhance this capability, we propose synthesizing general multimodal instructions from webpage UIs using text-based large language models (LLMs). Despite lacking direct visual input, text-based LLMs are able to process structured text representations from webpage accessibility trees. These instructions are then paired with UI screenshots to train multimodal models. We introduce MultiUI, a dataset containing 7.3 million samples from 1 million websites, covering diverse multimodal tasks and UI layouts. Models trained on MultiUI not only excel in web UI tasks-achieving up to a 48% improvement on VisualWebBench and a 19.1% boost in element accuracy on a web agent dataset Mind2Web-but also generalize surprisingly well to non-web UI tasks and even to non-UI domains, such as document understanding, OCR, and chart interpretation. These results highlight the broad applicability of web UI data for advancing text-rich visual understanding across various scenarios.