Search

Numerous decameter-sized asteroids have been observed impacting Earth as fireballs. These objects can have impact energies equivalent to hundreds of kilotons of TNT, posing a hazard if they impact populated areas. Previous estimates of meteoroid flux using fireball observations have shown an Earth impact rate for decameter-size objects of about once every $2$-$3$ years. In contrast, telescopic estimates of the near-Earth asteroid population predict the impact rate of such objects to be of order $20$-$40$ years, an order-of-magnitude difference. While the cause of this discrepancy remains unclear, tidal disruption of a larger near-Earth body has been proposed as an explanation for these excess decameter-sized impactors. The release in 2022 of previously classified United States Government (USG) satellite sensor data for fireball events has provided a wealth of new information on many of these impacts. Using this newly available USG sensor data, we present the first population-level study characterizing the orbital and dynamical properties of 14 decameter-sized Earth impactors detected by USG sensors since 1994, with a particular focus on searching for evidence of tidal disruption as the cause of the impact rate discrepancy. We find there is no evidence for recent ($\lesssim 10^4$ years) tidal disruption and weak evidence for longer-term tidal disruption in the decameter impactor population, but that the latter conclusion is limited by small number statistics. We also investigate the origins of both the impactor and near-Earth asteroid populations of decameter-sized objects from the main asteroid belt. We find that both populations generally originate from the same source regions: primarily from the $\nu_6$ secular resonance ($\sim70$%) with small contributions from the Hungaria group ($\sim20$%) and the 3:1 Jupiter mean-motion resonance ($\sim10$%)., Comment: Accepted 16 December 2024 for publication in Icarus. 19 pages, 10 figures, 2 tables

RR Lyrae in ultrafaint dwarf (UFD) galaxies, if found, would be valuable to constrain the distances of these UFD. In this work, we report our search of RR Lyrae in a recently discovered UFD -- Sextans II. Based on multiband ($griz$) time-series archival DECam imaging data, we do not find any RR Lyrae in Sextans II. On the contrary, the sparse multiband DECam light curves allowed recovery of a known foreground RR Lyrae, BB Sex, which happened to fall within the DECam images, and its pulsation period was correctly identified. Therefore, it is possible that Sextans II lacks RR Lyrae, similar to several other fainter UFDs (with $M_V \gtrsim -5.0$~mag) that do not have RR Lyrae., Comment: 7 pages and 7 Figures; AJ in-press

Recent studies have indicated that effectively utilizing inference-time compute is crucial for attaining better performance from large language models (LLMs). In this work, we propose a novel inference-aware fine-tuning paradigm, in which the model is fine-tuned in a manner that directly optimizes the performance of the inference-time strategy. We study this paradigm using the simple yet effective Best-of-N (BoN) inference strategy, in which a verifier selects the best out of a set of LLM-generated responses. We devise the first imitation learning and reinforcement learning~(RL) methods for BoN-aware fine-tuning, overcoming the challenging, non-differentiable argmax operator within BoN. We empirically demonstrate that our BoN-aware models implicitly learn a meta-strategy that interleaves best responses with more diverse responses that might be better suited to a test-time input -- a process reminiscent of the exploration-exploitation trade-off in RL. Our experiments demonstrate the effectiveness of BoN-aware fine-tuning in terms of improved performance and inference-time compute. In particular, we show that our methods improve the Bo32 performance of Gemma 2B on Hendrycks MATH from 26.8% to 30.8%, and pass@32 from 60.0% to 67.0%, as well as the pass@16 on HumanEval from 61.6% to 67.1%.

The growing concern over data privacy, the benefits of utilizing data from diverse sources for model training, and the proliferation of networked devices with enhanced computational capabilities have all contributed to the rise of federated learning (FL). The clients in FL collaborate to train a global model by uploading gradients computed on their private datasets without collecting raw data. However, a new attack surface has emerged from gradient sharing, where adversaries can restore the label distribution of a victim's private data by analyzing the obtained gradients. To mitigate this privacy leakage, existing lightweight defenses restrict the sharing of gradients, such as encrypting the final-layer gradients or locally updating the parameters within. In this paper, we introduce a novel attack called Gradient Bridge (GDBR) that recovers the label distribution of training data from the limited gradient information shared in FL. GDBR explores the relationship between the layer-wise gradients, tracks the flow of gradients, and analytically derives the batch training labels. Extensive experiments show that GDBR can accurately recover more than 80% of labels in various FL settings. GDBR highlights the inadequacy of restricted gradient sharing-based defenses and calls for the design of effective defense schemes in FL.

We investigate the metric theory of Diophantine approximation on missing-digit fractals. In particular, we establish analogues of Khinchin's theorem and Gallagher's theorem, as well as inhomogeneous generalisations., Comment: 36 pages

This work concerns the zeroth-order global minimization of continuous nonconvex functions with a unique global minimizer and possibly multiple local minimizers. We formulate a theoretical framework for inexact proximal point (IPP) methods for global optimization, establishing convergence guarantees under mild assumptions when either deterministic or stochastic estimates of proximal operators are used. The quadratic regularization in the proximal operator and the scaling effect of a parameter $\delta>0$ create a concentrated landscape of an associated Gibbs measure that is practically effective for sampling. The convergence of the expectation under the Gibbs measure as $\delta\to 0^+$ is established, and the convergence rate of $\mathcal O(\delta)$ is derived under additional assumptions. These results provide a theoretical foundation for evaluating proximal operators inexactly using sampling-based methods such as Monte Carlo (MC) integration. In addition, we propose a new approach based on tensor train (TT) approximation. This approach employs a randomized TT cross algorithm to efficiently construct a low-rank TT approximation of a discretized function using a small number of function evaluations, and we provide an error analysis for the TT-based estimation. We then propose two practical IPP algorithms, TT-IPP and MC-IPP. The TT-IPP algorithm leverages TT estimates of the proximal operators, while the MC-IPP algorithm employs MC integration to estimate the proximal operators. Both algorithms are designed to adaptively balance efficiency and accuracy in inexact evaluations of proximal operators. The effectiveness of the two algorithms is demonstrated through experiments on diverse benchmark functions and various applications.

We address the problem of personalized, interactive text-to-image (T2I) generation, designing a reinforcement learning (RL) agent which iteratively improves a set of generated images for a user through a sequence of prompt expansions. Using human raters, we create a novel dataset of sequential preferences, which we leverage, together with large-scale open-source (non-sequential) datasets. We construct user-preference and user-choice models using an EM strategy and identify varying user preference types. We then leverage a large multimodal language model (LMM) and a value-based RL approach to suggest a personalized and diverse slate of prompt expansions to the user. Our Personalized And Sequential Text-to-image Agent (PASTA) extends T2I models with personalized multi-turn capabilities, fostering collaborative co-creation and addressing uncertainty or underspecification in a user's intent. We evaluate PASTA using human raters, showing significant improvement compared to baseline methods. We also release our sequential rater dataset and simulated user-rater interactions to support future research in personalized, multi-turn T2I generation., Comment: Link to PASTA dataset: https://www.kaggle.com/datasets/googleai/pasta-data

As Graph Neural Networks (GNNs) become increasingly popular for learning from large-scale graph data across various domains, their susceptibility to adversarial attacks when using graph reduction techniques for scalability remains underexplored. In this paper, we present an extensive empirical study to investigate the impact of graph reduction techniques, specifically graph coarsening and sparsification, on the robustness of GNNs against adversarial attacks. Through extensive experiments involving multiple datasets and GNN architectures, we examine the effects of four sparsification and six coarsening methods on the poisoning attacks. Our results indicate that, while graph sparsification can mitigate the effectiveness of certain poisoning attacks, such as Mettack, it has limited impact on others, like PGD. Conversely, graph coarsening tends to amplify the adversarial impact, significantly reducing classification accuracy as the reduction ratio decreases. Additionally, we provide a novel analysis of the causes driving these effects and examine how defensive GNN models perform under graph reduction, offering practical insights for designing robust GNNs within graph acceleration systems.

Large language models (LLMs) are increasingly used in applications where LLM inputs may span many different tasks. Recent work has found that the choice of LLM is consequential, and different LLMs may be good for different input samples. Prior approaches have thus explored how engineers might select an LLM to use for each sample (i.e. routing). While existing routing methods mostly require training auxiliary models on human-annotated data, our work explores whether it is possible to perform unsupervised routing. We propose Smoothie, a weak supervision-inspired routing approach that requires no labeled data. Given a set of outputs from different LLMs, Smoothie constructs a latent variable graphical model over embedding representations of observable LLM outputs and unknown "true" outputs. Using this graphical model, we estimate sample-dependent quality scores for each LLM, and route each sample to the LLM with the highest corresponding score. We find that Smoothie's LLM quality-scores correlate with ground-truth model quality (correctly identifying the optimal model on 9/14 tasks), and that Smoothie outperforms baselines for routing by up to 10 points accuracy., Comment: 24 pages, 8 figures, 11 tables

Emergency response times are critical in densely populated urban environments like New York City (NYC), where traffic congestion significantly impedes emergency vehicle (EMV) mobility. This study introduces an intersection-aware emergency medical service (EMS) accessibility model to evaluate and improve EMV travel times across NYC. Integrating intersection density metrics, road network characteristics, and demographic data, the model identifies vulnerable regions with inadequate EMS coverage. The analysis reveals that densely interconnected areas, such as parts of Staten Island, Queens, and Manhattan, experience significant accessibility deficits due to intersection delays and sparse medical infrastructure. To address these challenges, this study explores the adoption of EMVLight, a multi-agent reinforcement learning framework, which demonstrates the potential to reduce intersection delays by 50\%, increasing EMS accessibility to 95\% of NYC residents within the critical benchmark of 4 minutes. Results indicate that advanced traffic signal control (TSC) systems can alleviate congestion-induced delays while improving equity in emergency response. The findings provide actionable insights for urban planning and policy interventions to enhance EMS accessibility and ensure timely care for underserved populations.

We present HumanEdit, a high-quality, human-rewarded dataset specifically designed for instruction-guided image editing, enabling precise and diverse image manipulations through open-form language instructions. Previous large-scale editing datasets often incorporate minimal human feedback, leading to challenges in aligning datasets with human preferences. HumanEdit bridges this gap by employing human annotators to construct data pairs and administrators to provide feedback. With meticulously curation, HumanEdit comprises 5,751 images and requires more than 2,500 hours of human effort across four stages, ensuring both accuracy and reliability for a wide range of image editing tasks. The dataset includes six distinct types of editing instructions: Action, Add, Counting, Relation, Remove, and Replace, encompassing a broad spectrum of real-world scenarios. All images in the dataset are accompanied by masks, and for a subset of the data, we ensure that the instructions are sufficiently detailed to support mask-free editing. Furthermore, HumanEdit offers comprehensive diversity and high-resolution $1024 \times 1024$ content sourced from various domains, setting a new versatile benchmark for instructional image editing datasets. With the aim of advancing future research and establishing evaluation benchmarks in the field of image editing, we release HumanEdit at \url{https://huggingface.co/datasets/BryanW/HumanEdit}., Comment: Codes and Supplementary Material: https://github.com/viiika/HumanEdit

Retrieval-augmented Generation (RAG) enhances Large Language Models (LLMs) by integrating external knowledge to reduce hallucinations and incorporate up-to-date information without retraining. As an essential part of RAG, external knowledge bases are commonly built by extracting structured data from unstructured PDF documents using Optical Character Recognition (OCR). However, given the imperfect prediction of OCR and the inherent non-uniform representation of structured data, knowledge bases inevitably contain various OCR noises. In this paper, we introduce OHRBench, the first benchmark for understanding the cascading impact of OCR on RAG systems. OHRBench includes 350 carefully selected unstructured PDF documents from six real-world RAG application domains, along with Q&As derived from multimodal elements in documents, challenging existing OCR solutions used for RAG To better understand OCR's impact on RAG systems, we identify two primary types of OCR noise: Semantic Noise and Formatting Noise and apply perturbation to generate a set of structured data with varying degrees of each OCR noise. Using OHRBench, we first conduct a comprehensive evaluation of current OCR solutions and reveal that none is competent for constructing high-quality knowledge bases for RAG systems. We then systematically evaluate the impact of these two noise types and demonstrate the vulnerability of RAG systems. Furthermore, we discuss the potential of employing Vision-Language Models (VLMs) without OCR in RAG systems. Code: https://github.com/opendatalab/OHR-Bench

We complement the $y$-band template light curves for RR Lyrae to the well-established template light curves in the $ugriz$-band, where the latter have been adopted in the astroML/gatspy python package as one of the period-search methods for RR Lyrae. These $y$-band template light curves were constructed based on the $z$-band time series data taken from the Sloan Digital Sky Survey (SDSS), the $y$-band light curves from the Pan-STARRS1 survey, and dedicated $y$-band observations using the Lulin One-meter Telescope for RR Lyrae located in the SDSS Stripe 82 region. These $y$-band template light curves, 9 for the ab-type RR Lyrae and 3 for the c-type RR Lyrae, can be applied in conjunction to the $ugriz$-band template light curves for upcoming sky surveys involving the $y$-band, such as the Vera C. Rubin Observatory Legacy Survey of Space and Time (LSST)., Comment: AJ accepted, 7 pages with 5 Figures and 1 Table

Instruction-based image editing aims to modify specific image elements with natural language instructions. However, current models in this domain often struggle to accurately execute complex user instructions, as they are trained on low-quality data with limited editing types. We present AnyEdit, a comprehensive multi-modal instruction editing dataset, comprising 2.5 million high-quality editing pairs spanning over 20 editing types and five domains. We ensure the diversity and quality of the AnyEdit collection through three aspects: initial data diversity, adaptive editing process, and automated selection of editing results. Using the dataset, we further train a novel AnyEdit Stable Diffusion with task-aware routing and learnable task embedding for unified image editing. Comprehensive experiments on three benchmark datasets show that AnyEdit consistently boosts the performance of diffusion-based editing models. This presents prospects for developing instruction-driven image editing models that support human creativity., Comment: 41 pages, 24 figures

Foundation models that bridge vision and language have made significant progress, inspiring numerous life-enriching applications. However, their potential for misuse to introduce new threats remains largely unexplored. This paper reveals that vision-language models (VLMs) can be exploited to overcome longstanding limitations in gradient inversion attacks (GIAs) within federated learning (FL), where an FL server reconstructs private data samples from gradients shared by victim clients. Current GIAs face challenges in reconstructing high-resolution images, especially when the victim has a large local data batch. While focusing reconstruction on valuable samples rather than the entire batch is promising, existing methods lack the flexibility to allow attackers to specify their target data. In this paper, we introduce Geminio, the first approach to transform GIAs into semantically meaningful, targeted attacks. Geminio enables a brand new privacy attack experience: attackers can describe, in natural language, the types of data they consider valuable, and Geminio will prioritize reconstruction to focus on those high-value samples. This is achieved by leveraging a pretrained VLM to guide the optimization of a malicious global model that, when shared with and optimized by a victim, retains only gradients of samples that match the attacker-specified query. Extensive experiments demonstrate Geminio's effectiveness in pinpointing and reconstructing targeted samples, with high success rates across complex datasets under FL and large batch sizes and showing resilience against existing defenses.

As object detection becomes integral to many safety-critical applications, understanding its vulnerabilities is essential. Backdoor attacks, in particular, pose a significant threat by implanting hidden backdoor in a victim model, which adversaries can later exploit to trigger malicious behaviors during inference. However, current backdoor techniques are limited to static scenarios where attackers must define a malicious objective before training, locking the attack into a predetermined action without inference-time adaptability. Given the expressive output space in object detection, including object existence detection, bounding box estimation, and object classification, the feasibility of implanting a backdoor that provides inference-time control with a high degree of freedom remains unexplored. This paper introduces AnywhereDoor, a flexible backdoor attack tailored for object detection. Once implanted, AnywhereDoor enables adversaries to specify different attack types (object vanishing, fabrication, or misclassification) and configurations (untargeted or targeted with specific classes) to dynamically control detection behavior. This flexibility is achieved through three key innovations: (i) objective disentanglement to support a broader range of attack combinations well beyond what existing methods allow; (ii) trigger mosaicking to ensure backdoor activations are robust, even against those object detectors that extract localized regions from the input image for recognition; and (iii) strategic batching to address object-level data imbalances that otherwise hinders a balanced manipulation. Extensive experiments demonstrate that AnywhereDoor provides attackers with a high degree of control, achieving an attack success rate improvement of nearly 80% compared to adaptations of existing methods for such flexible control.

This paper showcases AdaptLIL, a real-time adaptive link-indented list ontology mapping visualization that uses eye gaze as the primary input source. Through a multimodal combination of real-time systems, deep learning, and web development applications, this system uniquely curtails graphical overlays (adaptations) to pairwise mappings of link-indented list ontology visualizations for individual users based solely on their eye gaze., Comment: The paper was submitted without the consent of all authors. It is being withdrawn until full consent is obtained

Image shadow removal is a typical low-level vision problem, where the presence of shadows leads to abrupt changes in brightness in certain regions, affecting the accuracy of upstream tasks. Current shadow removal methods still face challenges such as residual boundary artifacts, and capturing feature information at shadow boundaries is crucial for removing shadows and eliminating residual boundary artifacts. Recently, Mamba has achieved remarkable success in computer vision by globally modeling long-sequence information with linear complexity. However, when applied to image shadow removal, the original Mamba scanning method overlooks the semantic continuity of shadow boundaries as well as the continuity of semantics within the same region. Based on the unique characteristics of shadow images, this paper proposes a novel selective scanning method called boundary-region selective scanning. This method scans boundary regions, shadow regions, and non-shadow regions independently, bringing pixels of the same region type closer together in the long sequence, especially focusing on the local information at the boundaries, which is crucial for shadow removal. This method combines with global scanning and channel scanning to jointly accomplish the shadow removal. We name our model ShadowMamba, the first Mamba-based model for shadow removal. Extensive experimental results show that our method outperforms current state-of-the-art models across most metrics on multiple datasets. The code for ShadowMamba is available at (Code will be released upon acceptance).

State-of-the-art noisy-intermediate-scale quantum (NISQ) processors are currently implemented across a variety of hardware platforms, each with their own distinct gatesets. As such, circuit compilation should not only be aware of, but also deeply connect to, the native gateset and noise properties of each. Trapped-ion processors are one such platform that provides a gateset that can be continuously parameterized across both one- and two-qubit gates. Here we use the Quantum Scientific Computing Open User Testbed (QSCOUT) to study noise-aware compilations focused on continuously parameterized two-qubit $\mathcal{ZZ}$ gates (based on the M{\o}lmer-S{\o}rensen interaction) using $\textbf{Superstaq}$, a quantum software platform for hardware-aware circuit compiler optimizations. We discuss the realization of $\mathcal{ZZ}$ gates with arbitrary angle on the all-to-all connected trapped-ion system. Then we discuss a variety of different compiler optimizations that innately target these $\mathcal{ZZ}$ gates and their noise properties. These optimizations include moving from a restricted maximally entangling gateset to a continuously parameterized one, swap mirroring to further reduce total entangling angle of the operations, focusing the heaviest $\mathcal{ZZ}$ angle participation on the best performing gate pairs, and circuit approximation to remove the least impactful $\mathcal{ZZ}$ gates. We demonstrate these compilation approaches on the hardware with randomized quantum volume circuits, observing the potential to realize a larger quantum volume as a result of these optimizations. Using differing yet complementary analysis techniques, we observe the distinct improvements in system performance provided by these noise-aware compilations and study the role of stochastic and coherent error channels for each compilation choice., Comment: 17 pages, 13 figures

Lifelong reinforcement learning (RL) has been developed as a paradigm for extending single-task RL to more realistic, dynamic settings. In lifelong RL, the "life" of an RL agent is modeled as a stream of tasks drawn from a task distribution. We propose EPIC (\underline{E}mpirical \underline{P}AC-Bayes that \underline{I}mproves \underline{C}ontinuously), a novel algorithm designed for lifelong RL using PAC-Bayes theory. EPIC learns a shared policy distribution, referred to as the \textit{world policy}, which enables rapid adaptation to new tasks while retaining valuable knowledge from previous experiences. Our theoretical analysis establishes a relationship between the algorithm's generalization performance and the number of prior tasks preserved in memory. We also derive the sample complexity of EPIC in terms of RL regret. Extensive experiments on a variety of environments demonstrate that EPIC significantly outperforms existing methods in lifelong RL, offering both theoretical guarantees and practical efficacy through the use of the world policy.

In recent times, Vision-Language Models (VLMs) have been trained under two predominant paradigms. Generative training has enabled Multimodal Large Language Models (MLLMs) to tackle various complex tasks, yet issues such as hallucinations and weak object discrimination persist. Discriminative training, exemplified by models like CLIP, excels in zero-shot image-text classification and retrieval, yet struggles with complex scenarios requiring fine-grained semantic differentiation. This paper addresses these challenges by proposing a unified approach that integrates the strengths of both paradigms. Considering interleaved image-text sequences as the general format of input samples, we introduce a structure-induced training strategy that imposes semantic relationships between input samples and the MLLM's hidden state. This approach enhances the MLLM's ability to capture global semantics and distinguish fine-grained semantics. By leveraging dynamic sequence alignment within the Dynamic Time Warping framework and integrating a novel kernel for fine-grained semantic differentiation, our method effectively balances generative and discriminative tasks. Extensive experiments demonstrate the effectiveness of our approach, achieving state-of-the-art results in multiple generative tasks, especially those requiring cognitive and discrimination abilities. Additionally, our method surpasses discriminative benchmarks in interleaved and fine-grained retrieval tasks. By employing a retrieval-augmented generation strategy, our approach further enhances performance in some generative tasks within one model, offering a promising direction for future research in vision-language modeling.

Planetary systems exhibiting mean-motion resonances (MMRs) offer unique opportunities to study the imprint of disk-induced migration on the orbital architectures of planetary systems. The HD 45364 system, discovered via the radial velocity (RV) method to host two giant planets in a 3:2 MMR, has been the subject of several studies attempting to reconstruct the system's orbital migration history based on its present-day resonant configuration. Recently, Li et al. (2022) called into question the system's residence in the 3:2 MMR based on a revised orbital solution derived from an expanded set of RV observations that extend the time baseline of the original discovery data by over a decade. However, we show that inferences about the planets' dynamical state with respect to the 3:2 MMR are sensitive to the particular prior assumptions adopted in the orbital modeling. Using $N$-body dynamical models, we show that orbital solutions constrained to reside deep in the 3:2 MMR fit the RV data with a similar quality to unconstrained orbital solutions. We conclude that the RV observations of HD 45364 are consistent with orbital configurations produced by smooth migration and resonance capture. We further show that past convergent orbital migration can reproduce the system's present-day orbital configuration provided that the ratio of migration to eccentricity damping timescales, $K$, was in the range $10\lesssim K \lesssim 175$. We also find that dynamical interactions in the system can break the usual mass-inclination degeneracy inherent to Keplerian models of RV observations and constrain the planets' absolute masses to within a factor of $\sim2$., Comment: 10 pages, 6 figures, submitted to ApJ

Protecting intellectual property (IP) in federated learning (FL) is increasingly important as clients contribute proprietary data to collaboratively train models. Model watermarking, particularly through backdoor-based methods, has emerged as a popular approach for verifying ownership and contributions in deep neural networks trained via FL. By manipulating their datasets, clients can embed a secret pattern, resulting in non-intuitive predictions that serve as proof of participation, useful for claiming incentives or IP co-ownership. However, this technique faces practical challenges: client watermarks can collide, leading to ambiguous ownership claims, and malicious clients may exploit watermarks to inject harmful backdoors, jeopardizing model integrity. To address these issues, we propose Sanitizer, a server-side method that ensures client-embedded backdoors cannot be triggered on natural queries in harmful ways. It identifies subnets within client-submitted models, extracts backdoors throughout the FL process, and confines them to harmless, client-specific input subspaces. This approach not only enhances Sanitizer's efficiency but also resolves conflicts when clients use similar triggers with different target labels. Our empirical results demonstrate that Sanitizer achieves near-perfect success in verifying client contributions while mitigating the risks of malicious watermark use. Additionally, it reduces GPU memory consumption by 85% and cuts processing time by at least 5 times compared to the baseline.

GPT-4o is an autoregressive omni model that accepts as input any combination of text, audio, image, and video, and generates any combination of text, audio, and image outputs. It's trained end-to-end across text, vision, and audio, meaning all inputs and outputs are processed by the same neural network. GPT-4o can respond to audio inputs in as little as 232 milliseconds, with an average of 320 milliseconds, which is similar to human response time in conversation. It matches GPT-4 Turbo performance on text in English and code, with significant improvement on text in non-English languages, while also being much faster and 50\% cheaper in the API. GPT-4o is especially better at vision and audio understanding compared to existing models. In line with our commitment to building AI safely and consistent with our voluntary commitments to the White House, we are sharing the GPT-4o System Card, which includes our Preparedness Framework evaluations. In this System Card, we provide a detailed look at GPT-4o's capabilities, limitations, and safety evaluations across multiple categories, focusing on speech-to-speech while also evaluating text and image capabilities, and measures we've implemented to ensure the model is safe and aligned. We also include third-party assessments on dangerous capabilities, as well as discussion of potential societal impacts of GPT-4o's text and vision capabilities.

We characterise partition regularity for linear equations over the Piatetski-Shapiro numbers $\lfloor n^c \rfloor$ when $1 < c < c^\dag(s)$, where $s \ge 3$ is the number of variables. Here $c^\dag(3) = 12/11$ and $c^\dag(4) = 7/6$, while $c^\dag(s) = 2$ for $s \ge 5$. We also establish density results with quantitative bounds., Comment: 18 pages; added a reference to and discussion of earlier work of Saito and Yoshida

We present Meissonic, which elevates non-autoregressive masked image modeling (MIM) text-to-image to a level comparable with state-of-the-art diffusion models like SDXL. By incorporating a comprehensive suite of architectural innovations, advanced positional encoding strategies, and optimized sampling conditions, Meissonic substantially improves MIM's performance and efficiency. Additionally, we leverage high-quality training data, integrate micro-conditions informed by human preference scores, and employ feature compression layers to further enhance image fidelity and resolution. Our model not only matches but often exceeds the performance of existing models like SDXL in generating high-quality, high-resolution images. Extensive experiments validate Meissonic's capabilities, demonstrating its potential as a new standard in text-to-image synthesis. We release a model checkpoint capable of producing $1024 \times 1024$ resolution images., Comment: Codes and Supplementary Material: https://github.com/viiika/Meissonic

Dielectric media are very promising for near-field radiative heat transfer (NFRHT) applications as these materials can thermally emit surface phonon polaritons (SPhPs) resulting in large and quasi-monochromatic heat fluxes. Near-field radiative heat flux between dissimilar dielectric media is much smaller than that between similar dielectric media and is also not quasi-monochromatic. This is due to the mismatch of the SPhP frequencies of the two heat-exchanging dielectric media. Here, we experimentally demonstrate that NFRHT between dissimilar dielectric media increases substantially when a graphene sheet is deposited on the medium with the smaller SPhP frequency. An enhancement of 2.7 to 3.2 folds is measured for the heat flux between SiC and LiF separated by a vacuum gap of size 100 to 140 nm when LiF is covered by a graphene sheet. This enhancement is due to the coupling of SPhPs and surface plasmon polaritons (SPPs). The SPPs of graphene are coupled to the SPhPs of LiF resulting in coupled SPhP-SPPs with a dispersion branch monotonically increasing with the wavevector. This monotonically increasing branch of dispersion relation intersects the dispersion branch of the SPhPs of SiC causing the coupling of the surface modes across the vacuum gap, which resonantly increases the heat flux at the SPhP frequency of SiC. This surface phonon-plasmon coupling also makes NFRHT quasi-monochromatic, which is highly desired for applications such as near-field thermophotovoltaics and thermophotonics. This study experimentally demonstrates that graphene is a very promising material for tuning the magnitude and spectrum of NFRHT between dissimilar dielectric media.

After large models (LMs) have gained widespread acceptance in code-related tasks, their superior generative capacity has greatly promoted the application of the code LM. Nevertheless, the security of the generated code has raised attention to its potential damage. Existing secure code generation methods have limited generalizability to unseen test cases and poor robustness against the attacked model, leading to safety failures in code generation. In this paper, we propose a generalizable and robust secure code generation method SecCoder by using in-context learning (ICL) and the safe demonstration. The dense retriever is also used to select the most helpful demonstration to maximize the improvement of the generated code's security. Experimental results show the superior generalizability of the proposed model SecCoder compared to the current secure code generation method, achieving a significant security improvement of an average of 7.20% on unseen test cases. The results also show the better robustness of SecCoder compared to the current attacked code LM, achieving a significant security improvement of an average of 7.74%. Our analysis indicates that SecCoder enhances the security of LMs in generating code, and it is more generalizable and robust., Comment: To Appear in the 2024 Conference on Empirical Methods in Natural Language Processing (EMNLP)

The discovery of superconductivity in the Ba-La-Cu-O system (the cuprate) at the 30 K range in 1986 marked a significant breakthrough, as it far exceeded the highest known critical temperature ($T_c$) at the time and surpassed the predicted 30 K limit, which was thought to be the maximum before phonon-mediated electron pairing would break down due to thermal excitation. Despite recent successful observations of superconductivity in nickel-oxide-based compounds (the nickelate), superconductivity above 30 K at ambient pressure in a system that is isostructural and isoelectronic to the cuprate but without copper has remained elusive. Here, we report a superconducting $T_c$ above 35 K under ambient pressure in hole-doped, late rare-earth infinite-layer nickel oxide (Sm-Eu-Ca-Sr)NiO$_2$ thin films. Electron microscopy reveals a small thickness of ~2 nm of infinite-layer phase stabilised at present, which indicates a higher temperature superconductivity should be observable in clean bulk crystals., Comment: Main: 14 pages, 6 figures; Supplementary: 13 pages, 9 figures

With the rise of data volume and computing power, seismological research requires more advanced skills in data processing, numerical methods, and parallel computing. We present the experience of conducting training workshops over various forms of delivery to support the adoption of large-scale High-Performance Computing and Cloud computing to advance seismological research. The seismological foci were on earthquake source parameter estimation in catalogs, forward and adjoint wavefield simulations in 2 and 3 dimensions at local, regional, and global scales, earthquake dynamics, ambient noise seismology, and machine learning. This contribution describes the series of workshops, the learning outcomes of the participants, and lessons learned by the instructors. Our curriculum was grounded on open and reproducible science, large-scale scientific computing and data mining, and computing infrastructure (access and usage) for HPC and the cloud. We also describe the types of teaching materials that have proven beneficial to the instruction and the sustainability of the program. We propose guidelines to deliver future workshops on these topics.

Given $\boldsymbol{\alpha} \in [0,1]^d$, we estimate the smooth discrepancy of the Kronecker sequence $(n \boldsymbol{\alpha} \,\mathrm{mod}\, 1)_{n\geq 1}$. We find that it can be smaller than the classical discrepancy of $\textbf{any}$ sequence when $d \le 2$, and can even be bounded in the case $d=1$. To achieve this, we establish a novel deterministic analogue of Beck's local-to-global principle (Ann. of Math. 1994), which relates the discrepancy of a Kronecker sequence to multiplicative diophantine approximation. This opens up a new avenue of attack for Littlewood's conjecture., Comment: 17 pages; comments welcome

Multi-modal large language models (MLLMs) have enabled numerous advances in understanding and reasoning in domains like vision, but we have not yet seen this broad success for time-series. Although prior works on time-series MLLMs have shown promising performance in time-series forecasting, very few works show how an LLM could be used for time-series reasoning in natural language. We propose a novel multi-modal time-series LLM approach that learns generalizable information across various domains with powerful zero-shot performance. First, we train a lightweight time-series encoder on top of an LLM to directly extract time-series information. Then, we fine-tune our model with chain-of-thought augmented time-series tasks to encourage the model to generate reasoning paths. We show that our model learns a latent representation that reflects specific time-series features (e.g. slope, frequency), as well as outperforming GPT-4o on a set of zero-shot reasoning tasks on a variety of domains., Comment: Oral Presentation at 2024 NeurIPS Workshop on Time Series in the Age of Large Models

Deep learning methods have been widely used as an end-to-end modeling strategy of electrical energy systems because of their conveniency and powerful pattern recognition capability. However, due to the "black-box" nature, deep learning methods have long been blamed for their poor interpretability when modeling a physical system. In this paper, we introduce a novel neural network structure, Kolmogorov-Arnold Network (KAN), to achieve "white-box" modeling for electrical energy systems to enhance the interpretability. The most distinct feature of KAN lies in the learnable activation function together with the sparse training and symbolification process. Consequently, KAN can express the physical process with concise and explicit mathematical formulas while remaining the nonlinear-fitting capability of deep neural networks. Simulation results based on three electrical energy systems demonstrate the effectiveness of KAN in the aspects of interpretability, accuracy, robustness and generalization ability.

Body composition analysis is vital in assessing health conditions such as obesity, sarcopenia, and metabolic syndromes. MRI provides detailed images of skeletal muscle (SKM), visceral adipose tissue (VAT), and subcutaneous adipose tissue (SAT), but their manual segmentation is labor-intensive and limits clinical applicability. This study validates an automated tool for MRI-based 2D body composition analysis- (Data Analysis Facilitation Suite (DAFS) Express), comparing its automated measurements with expert manual segmentations using UK Biobank data. A cohort of 399 participants from the UK Biobank dataset was selected, yielding 423 single L3 slices for analysis. DAFS Express performed automated segmentations of SKM, VAT, and SAT, which were then manually corrected by expert raters for validation. Evaluation metrics included Jaccard coefficients, Dice scores, Intraclass Correlation Coefficients (ICCs), and Bland-Altman Plots to assess segmentation agreement and reliability. High agreements were observed between automated and manual segmentations with mean Jaccard scores: SKM 99.03%, VAT 95.25%, and SAT 99.57%; and mean Dice scores: SKM 99.51%, VAT 97.41%, and SAT 99.78%. Cross-sectional area comparisons showed consistent measurements with automated methods closely matching manual measurements for SKM and SAT, and slightly higher values for VAT (SKM: Auto 132.51 cm^2, Manual 132.36 cm^2; VAT: Auto 137.07 cm^2, Manual 134.46 cm^2; SAT: Auto 203.39 cm^2, Manual 202.85 cm^2). ICCs confirmed strong reliability (SKM: 0.998, VAT: 0.994, SAT: 0.994). Bland-Altman plots revealed minimal biases, and boxplots illustrated distribution similarities across SKM, VAT, and SAT areas. On average DAFS Express took 18 seconds per DICOM. This underscores its potential to streamline image analysis processes in research and clinical settings, enhancing diagnostic accuracy and efficiency.

Monitoring carbon dioxide (CO2) injected and stored in subsurface reservoirs is critical for avoiding failure scenarios and enables real-time optimization of CO2 injection rates. Sequential Bayesian data assimilation (DA) is a statistical method for combining information over time from multiple sources to estimate a hidden state, such as the spread of the subsurface CO2 plume. An example of scalable and efficient sequential Bayesian DA is the ensemble Kalman filter (EnKF). We improve upon existing DA literature in the seismic-CO2 monitoring domain by applying this scalable DA algorithm to a high-dimensional CO2 reservoir using two-phase flow dynamics and time-lapse full waveform seismic data with a realistic surface-seismic survey design. We show more accurate estimates of the CO2 saturation field using the EnKF compared to using either the seismic data or the fluid physics alone. Furthermore, we test a range of values for the EnKF hyperparameters and give guidance on their selection for seismic CO2 reservoir monitoring., Comment: This work has been submitted to the IEEE for possible publication

Deep learning (DL) models have shown significant potential in Alzheimer's Disease (AD) classification. However, understanding and interpreting these models remains challenging, which hinders the adoption of these models in clinical practice. Techniques such as saliency maps have been proven effective in providing visual and empirical clues about how these models work, but there still remains a gap in understanding which specific brain regions DL models focus on and whether these brain regions are pathologically associated with AD. To bridge such gap, in this study, we developed a quantitative disease-focusing strategy to first enhance the interpretability of DL models using saliency maps and brain segmentations; then we propose a disease-focus (DF) score that quantifies how much a DL model focuses on brain areas relevant to AD pathology based on clinically known MRI-based pathological regions of AD. Using this strategy, we compared several state-of-the-art DL models, including a baseline 3D ResNet model, a pretrained MedicalNet model, and a MedicalNet with data augmentation to classify patients with AD vs. cognitive normal patients using MRI data; then we evaluated these models in terms of their abilities to focus on disease-relevant regions. Our results show interesting disease-focusing patterns with different models, particularly characteristic patterns with the pretrained models and data augmentation, and also provide insight into their classification performance. These results suggest that the approach we developed for quantitatively assessing the abilities of DL models to focus on disease-relevant regions may help improve interpretability of these models for AD classification and facilitate their adoption for AD diagnosis in clinical practice. The code is publicly available at https://github.com/Liang-lt/ADNI.

Imaging tasks are typically tackled using a structured optimization framework. This paper delves into a class of algorithms for difference-of-convex (DC) structured optimization, focusing on minimizing a DC function along with a possibly nonconvex function. Existing DC algorithm (DCA) versions often fail to effectively handle nonconvex functions or exhibit slow convergence rates. We propose a novel inertial proximal DC algorithm in Bregman geometry, named iBPDCA, designed to address nonconvex terms and enhance convergence speed through inertial techniques. We provide a detailed theoretical analysis, establishing both subsequential and global convergence of iBPDCA via the Kurdyka-{\L}ojasiewicz property. Additionally, we introduce a Plug-and-Play variant, PnP-iBPDCA, which employs a deep neural network-based prior for greater flexibility and robustness while ensuring theoretical convergence. We also establish that the Gaussian gradient step denoiser used in our method is equivalent to evaluating the Bregman proximal operator for an implicitly weakly convex functional. We extensively validate our method on Rician noise and phase retrieval. We demonstrate that iBPDCA surpasses existing state-of-the-art methods.

Interest in the study and analysis of dynamic processes in the social, behavioral, and health sciences has burgeoned in recent years due to the increased availability of intensive longitudinal data. However, how best to model and account for the persistent heterogeneity characterizing such processes remains an open question. The multi-VAR framework, a recent methodological development built on the vector autoregressive model, accommodates heterogeneous dynamics in multiple-subject time series through structured penalization. In the original multi-VAR proposal, individual-level transition matrices are decomposed into common and unique dynamics, allowing for generalizable and person-specific features. The current project extends this framework to allow additionally for the identification and penalized estimation of subgroup-specific dynamics; that is, patterns of dynamics that are shared across subsets of individuals. The performance of the proposed subgrouping extension is evaluated in the context of both a simulation study and empirical application, and results are compared to alternative methods for subgrouping multiple-subject, multivariate time series.

We present an advanced $\textit{ab initio}$ quantum Monte Carlo (QMC) calculation of the ground state of undoped CaCuO$_2$. We extend the traditional single-determinant Slater-Jastrow approach to include multi-determinant wave functions, inhomogeneous Jastrow factors, and orbital optimization. Our results demonstrate not only an improvement in the variational bound of the ground state energy -- 2.3 eV per formula unit lower than previous state of the art techniques -- but also confirm the presence of spin fluctuations in multi-determinant wave functions in a strongly correlated cuprate system, which is integral to understanding high-$T_c$ superconductivity. This is the first demonstration of capturing spin fluctuations in QMC wave functions on a cuprate, establishing the groundwork for new studies on doped cuprates in the superconducting state, where spin fluctuations require more accurate characterization., Comment: 9 pages, 6 figures

As an IT-enabled multi-passenger mobility service, microtransit has the potential to improve accessibility, reduce congestion, and enhance flexibility in transportation options. However, due to its heterogeneous impacts on different communities and population segments, there is a need for better tools in microtransit forecast and revenue management, especially when actual usage data are limited. We propose a novel framework based on an agent-based mixed logit model estimated with microtransit usage data and synthetic trip data. The framework involves estimating a lower-branch mode choice model with synthetic trip data, combining lower-branch parameters with microtransit data to estimate an upper-branch ride pass subscription model, and applying the nested model to evaluate microtransit pricing and subsidy policies. The framework enables further decision-support analysis to consider diverse travel patterns and heterogeneous tastes of the total population. We test the framework in a case study with synthetic trip data from Replica Inc. and microtransit data from Arlington Via. The lower-branch model result in a rho-square value of 0.603 on weekdays and 0.576 on weekends. Predictions made by the upper-branch model closely match the marginal subscription data. In a ride pass pricing policy scenario, we show that a discount in weekly pass (from $25 to $18.9) and monthly pass (from $80 to $71.5) would surprisingly increase total revenue by $102/day. In an event- or place-based subsidy policy scenario, we show that a 100% fare discount would reduce 80 car trips during peak hours at AT&T Stadium, requiring a subsidy of $32,068/year.

Let $\Gamma$ be a Zariski dense $\Theta$-Anosov subgroup of a connected semisimple real algebraic group for some nonempty subset of simple roots $\Theta$. In the Anosov setting, there is a natural compact metric space $\mathcal{X}$ equipped with a family of translation flows $\{a^{\mathsf{v}}_t\}_{t \in \mathbb R}$, parameterized by vectors $\mathsf{v}$ in the interior of the $\Theta$-limit cone $\mathcal{L}_\Theta$ of $\Gamma$, and they are H\"{o}lder conjugate to H\"{o}lder reparametrizations of the Gromov geodesic flow. We prove that for all $\mathsf{v}$ outside an exceptional cone $\mathscr{E} \subset \operatorname{int}\mathcal{L}_\Theta$, which is a smooth image of the walls of the Weyl chamber, the translation flow is exponentially mixing with respect to the Bowen-Margulis-Sullivan measure associated to $\mathsf{v}$. Moreover, the exponential rate is uniform for a compact set of such $\mathsf{v}$. We also obtain an essential spectral gap formulated in terms of the Selberg zeta function and a prime orbit theorem with a power saving error term. Our proof relies on Lie theoretic techniques to prove the crucial local non-integrability condition (LNIC) for the translation flows and thereby implement Dolgopyat's method in a uniform fashion. The exceptional cone $\mathscr{E}$ arises from the failure of LNIC for those vectors., Comment: 45 pages

We determine the generic multiplicative approximation rate on a hypersurface. There are four regimes, according to convergence or divergence and curved or flat, and we address all of them. Using geometry and arithmetic in Fourier space, we develop a general framework of moment transference principles, which convert Lebesgue data into data for some other measure.