Search

The Algebraic Data Type (ADT) can be used as a computational framework for molecular representation for the purpose of advancing tasks in cheminformatics. This can include generative modles in the context of Bayesian machine learning via probabilistic programming. The ADT that we put forward, implements the 'Dietz' representation for molecular constitution via multigraphs of electron valence information, and uses 3D coordinate data to provide stereochemical information, easily enabling the representation of complex molecular phenomena such as organometallics, multi-center bonds, delocalized electrons, and resonant structures. Unlike traditional string-based representations such as SMILES and SELFIES, the ADT is much more flexible, yet retains desirable qualities from type-safety to seamless integration with Bayesian Probabilistic Programming. An extensive criticism of both SMILES and SELFIES, in this article, is given, along with criticisms of the so-called Future of SELFIES. An open-source library implemented in Haskell demonstrates the ADT along with experimental extensions demonstrating its use in reaction modelling, group theoretic applications, and integration with LazyPPL, a lazy probabilistic programming library. Also provided as an extension is the ability to represent electronic structures, including shells, subshells, and orbitals, significantly expanding its representational scope compared to other representations in the literature. These features position the proposed ADT as a robust alternative to existing molecular representations, addressing limitations such as inadequate support for 3D information and syntactic invalidity while offering a platform for innovative cheminformatics research. Accompanying discussions about the meaning of a `representation' are included. The fully functioning GitHub library can be found at https://github.com/oliverjgoldstein/chemalgprog., Comment: 1 Figure

Using Gaia DR3 data, we examine the kinematics of the central core of the Sagittarius (Sgr) dwarf spheroidal galaxy using data which includes proper motions and line-of-sight velocities for member stars in addition to their projected positions. We extract a sample of bright stars that are high-probability members of Sgr. We model the distances to these stars, which is the only missing phase-space component measurement from our 5D sample, highlighting how their corresponding uncertainties propagate to affect the kinematics. Using line-of-sight velocity data only, which are not affected by the distance uncertainties, and assuming a jeans-based equilibrium analysis, we obtain a velocity anisotropy of $\beta_a = -2.24 \pm 1.99$, which implies a system with tangentially-biased orbits. With the full 5D data, we project that the data will significantly improve upon measurements of the log-slope of the dark matter density profile and the stellar velocity anisotropy. Tests with mock distance data show an improvement of anisotropy errors of approximately an order of magnitude, and log slope at the half light radius of approximately half an order of magnitude., Comment: 11 pages, 8 figures, 1 auxiliary text file

We discuss the "Infinitely Many Meanings" attacks (IMM), a category of jailbreaks that leverages the increasing capabilities of a model to handle paraphrases and encoded communications to bypass their defensive mechanisms. IMMs' viability pairs and grows with a model's capabilities to handle and bind the semantics of simple mappings between tokens and work extremely well in practice, posing a concrete threat to the users of the most powerful LLMs in commerce. We show how one can bypass the safeguards of the most powerful open- and closed-source LLMs and generate content that explicitly violates their safety policies. One can protect against IMMs by improving the guardrails and making them scale with the LLMs' capabilities. For two categories of attacks that are straightforward to implement, i.e., bijection and encoding, we discuss two defensive strategies, one in token and the other in embedding space. We conclude with some research questions we believe should be prioritised to enhance the defensive mechanisms of LLMs and our understanding of their safety.

The fastrerandomize R package provides hardware-accelerated tools for performing rerandomization and randomization testing in experimental research. Using a JAX backend, the package enables exact rerandomization inference even for large experiments with hundreds of billions of possible randomizations. Key functionalities include generating pools of acceptable rerandomizations based on covariate balance, conducting exact randomization tests, and performing pre-analysis evaluations to determine optimal rerandomization acceptance thresholds. Through batched processing and GPU acceleration, fastrerandomize achieves substantial performance gains compared to existing implementations, making previously intractable designs computationally feasible. The package therefore extends the randomization-based inference toolkit in R, allowing researchers to efficiently implement more stringent rerandomization designs and conduct valid inference even with large sample sizes or in high-dimensional settings., Comment: 35 pages, 9 figures

Recent advances in diffusion generative models have yielded remarkable progress. While the quality of generated content continues to improve, these models have grown considerably in size and complexity. This increasing computational burden poses significant challenges, particularly in resource-constrained deployment scenarios such as mobile devices. The combination of model pruning and knowledge distillation has emerged as a promising solution to reduce computational demands while preserving generation quality. However, this technique inadvertently propagates undesirable behaviors, including the generation of copyrighted content and unsafe concepts, even when such instances are absent from the fine-tuning dataset. In this paper, we propose a novel bilevel optimization framework for pruned diffusion models that consolidates the fine-tuning and unlearning processes into a unified phase. Our approach maintains the principal advantages of distillation-namely, efficient convergence and style transfer capabilities-while selectively suppressing the generation of unwanted content. This plug-in framework is compatible with various pruning and concept unlearning methods, facilitating efficient, safe deployment of diffusion models in controlled environments.

Probabilistic programming languages, which exist in abundance, are languages that allow users to calculate probability distributions defined by probabilistic programs, by using inference algorithms. However, the underlying inference algorithms are not implemented in a modular fashion, though, the algorithms are presented as a composition of other inference components. This discordance between the theory and the practice of Bayesian machine learning, means that reasoning about the correctness of probabilistic programs is more difficult, and composing inference algorithms together in code may not necessarily produce correct compound inference algorithms. In this dissertation, I create a modular probabilistic programming library, already a nice property as its not a standalone language, called Koka Bayes, that is based off of both the modular design of Monad Bayes -- a probabilistic programming library developed in Haskell -- and its semantic validation. The library is embedded in a recently created programming language, Koka, that supports algebraic effect handlers and expressive effect types -- novel programming abstractions that support modular programming. Effects are generalizations of computational side-effects, and it turns out that fundamental operations in probabilistic programming such as probabilistic choice and conditioning are instances of effects., Comment: 47 Pages

Robustness and generalizability in medical image segmentation are often hindered by scarcity and limited diversity of training data, which stands in contrast to the variability encountered during inference. While conventional strategies -- such as domain-specific augmentation, specialized architectures, and tailored training procedures -- can alleviate these issues, they depend on the availability and reliability of domain knowledge. When such knowledge is unavailable, misleading, or improperly applied, performance may deteriorate. In response, we introduce a novel, domain-agnostic, add-on, and data-driven strategy inspired by image stacking in image denoising. Termed ``semantic stacking,'' our method estimates a denoised semantic representation that complements the conventional segmentation loss during training. This method does not depend on domain-specific assumptions, making it broadly applicable across diverse image modalities, model architectures, and augmentation techniques. Through extensive experiments, we validate the superiority of our approach in improving segmentation performance under diverse conditions. Code is available at https://github.com/ymp5078/Semantic-Stacking., Comment: AAAI2025

A key component of building safe and reliable language models is enabling the models to appropriately refuse to follow certain instructions or answer certain questions. We may want models to output refusal messages for various categories of user queries, for example, ill-posed questions, instructions for committing illegal acts, or queries which require information past the model's knowledge horizon. Engineering models that refuse to answer such questions is complicated by the fact that an individual may want their model to exhibit varying levels of sensitivity for refusing queries of various categories, and different users may want different refusal rates. The current default approach involves training multiple models with varying proportions of refusal messages from each category to achieve the desired refusal rates, which is computationally expensive and may require training a new model to accommodate each user's desired preference over refusal rates. To address these challenges, we propose refusal tokens, one such token for each refusal category or a single refusal token, which are prepended to the model's responses during training. We then show how to increase or decrease the probability of generating the refusal token for each category during inference to steer the model's refusal behavior. Refusal tokens enable controlling a single model's refusal rates without the need of any further fine-tuning, but only by selectively intervening during generation., Comment: 19 pages

The evolution of multicellularity involved the transformation of a simple cell wall of unicellular ancestors into a complex, multifunctional extracellular matrix (ECM). A suitable model organism to study the formation and expansion of an ECM during ontogenesis is the multicellular green alga $Volvox~carteri$, which, along with the related volvocine algae, produces a complex, self-organized ECM composed of multiple substructures. These self-assembled ECMs primarily consist of hydroxyproline-rich glycoproteins, a major component of which is pherophorins. To investigate the geometry of the growing ECM, we fused the $yfp$ gene with the gene for pherophorin II (PhII) in $V.~carteri$. Confocal microscopy reveals PhII:YFP localization at key structures within the ECM, including the boundaries of compartments surrounding each somatic cell and the outer surface of the organism. Image analysis during the life cycle allows the stochastic geometry of those growing compartments to be quantified. We find that their areas and aspect ratios exhibit robust gamma distributions and exhibit a transition from a tight polygonal to a looser acircular packing geometry with stable eccentricity over time, evoking parallels and distinctions with the behavior of hydrated foams. These results provide a quantitative benchmark for addressing a general, open question in biology: How do cells produce structures external to themselves in a robust and accurate manner?, Comment: 14 pages, 12 figures + SI Appendix (21 pages, 11 figures)

Image editing technologies are tools used to transform, adjust, remove, or otherwise alter images. Recent research has significantly improved the capabilities of image editing tools, enabling the creation of photorealistic and semantically informed forged regions that are nearly indistinguishable from authentic imagery, presenting new challenges in digital forensics and media credibility. While current image forensic techniques are adept at localizing forged regions produced by traditional image manipulation methods, current capabilities struggle to localize regions created by diffusion-based techniques. To bridge this gap, we present a novel framework that integrates a multimodal Large Language Model (LLM) for enhanced reasoning capabilities to localize tampered regions in images produced by diffusion model-based editing methods. By leveraging the contextual and semantic strengths of LLMs, our framework achieves promising results on MagicBrush, AutoSplice, and PerfBrush (novel diffusion-based dataset) datasets, outperforming previous approaches in mIoU and F1-score metrics. Notably, our method excels on the PerfBrush dataset, a self-constructed test set featuring previously unseen types of edits. Here, where traditional methods typically falter, achieving markedly low scores, our approach demonstrates promising performance.

3D Gaussian Splatting (3D-GS) is a recent 3D scene reconstruction technique that enables real-time rendering of novel views by modeling scenes as parametric point clouds of differentiable 3D Gaussians. However, its rendering speed and model size still present bottlenecks, especially in resource-constrained settings. In this paper, we identify and address two key inefficiencies in 3D-GS, achieving substantial improvements in rendering speed, model size, and training time. First, we optimize the rendering pipeline to precisely localize Gaussians in the scene, boosting rendering speed without altering visual fidelity. Second, we introduce a novel pruning technique and integrate it into the training pipeline, significantly reducing model size and training time while further raising rendering speed. Our Speedy-Splat approach combines these techniques to accelerate average rendering speed by a drastic $6.71\times$ across scenes from the Mip-NeRF 360, Tanks & Temples, and Deep Blending datasets with $10.6\times$ fewer primitives than 3D-GS.

Activity and autonomous motion are fundamental aspects of many living and engineering systems. Here, the scale of biological agents covers a wide range, from nanomotors, cytoskeleton, and cells, to insects, fish, birds, and people. Inspired by biological active systems, various types of autonomous synthetic nano- and micromachines have been designed, which provide the basis for multifunctional, highly responsive, intelligent active materials. A major challenge for understanding and designing active matter is their inherent non-equilibrium nature due to persistent energy consumption, which invalidates equilibrium concepts such as free energy, detailed balance, and time-reversal symmetry. Furthermore, interactions in ensembles of active agents are often non-additive and non-reciprocal. An important aspect of biological agents is their ability to sense the environment, process this information, and adjust their motion accordingly. It is an important goal for the engineering of micro-robotic systems to achieve similar functionality. With many fundamental properties of motile active matter now reasonably well understood and under control, the ground is prepared for the study of physical aspects and mechanisms of motion in complex environments, of the behavior of systems with new physical features like chirality, of the development of novel micromachines and microbots, of the emergent collective behavior and swarming of intelligent self-propelled particles, and of particular features of microbial systems. The vast complexity of phenomena and mechanisms involved in the self-organization and dynamics of motile active matter poses major challenges, which can only be addressed by a truly interdisciplinary effort involving scientists from biology, chemistry, ecology, engineering, mathematics, and physics.

Technological advances in artificial intelligence (AI) have enabled the development of large vision language models (LVLMs) that are trained on millions of paired image and text samples. Subsequent research efforts have demonstrated great potential of LVLMs to achieve high performance in medical imaging use cases (e.g., radiology report generation), but there remain barriers that hinder the ability to deploy these solutions broadly. These include the cost of extensive computational requirements for developing large scale models, expertise in the development of sophisticated AI models, and the difficulty in accessing substantially large, high-quality datasets that adequately represent the population in which the LVLM solution is to be deployed. The HOPPR Medical-Grade Platform addresses these barriers by providing powerful computational infrastructure, a suite of foundation models on top of which developers can fine-tune for their specific use cases, and a robust quality management system that sets a standard for evaluating fine-tuned models for deployment in clinical settings. The HOPPR Platform has access to millions of imaging studies and text reports sourced from hundreds of imaging centers from diverse populations to pretrain foundation models and enable use case-specific cohorts for fine-tuning. All data are deidentified and securely stored for HIPAA compliance. Additionally, developers can securely host models on the HOPPR platform and access them via an API to make inferences using these models within established clinical workflows. With the Medical-Grade Platform, HOPPR's mission is to expedite the deployment of LVLM solutions for medical imaging and ultimately optimize radiologist's workflows and meet the growing demands of the field., Comment: 6 pages, 3 figures

The short gamma-ray burst (sGRB), GRB~170817A, is often considered a rare event. However, its inferred event rate, $\mathcal{O}(100s)\ \text{Gpc}^{-3}\ \text{yr}^{-1}$, exceeds cosmic sGRB rate estimates from high-redshift samples by an order of magnitude. This discrepancy can be explained by geometric effects related to the structure of the relativistic jet. We first illustrate how adopting a detector flux threshold point estimate rather than an efficiency function, can lead to a large variation in rate estimates. Simulating the Fermi-GBM sGRB detection efficiency, we then show that for a given a universal structured jet profile, one can model a geometric bias with redshift. Assuming different jet profiles, we show a geometrically scaled rate of GRB~170817A is consistent with the cosmic beaming uncorrected rate estimates of short $\gamma$-ray bursts (sGRBs) and that geometry can boost observational rates within $\mathcal{O}(100s)$\,Mpc. We find an apparent GRB~170817A rate of $303_{-300}^{+1580}$ $\mathrm{Gpc}^{-3}\, \mathrm{yr}^{-1} $ which when corrected for geometry yields $6.15_{-6.06}^{+31.2}$ $\mathrm{Gpc}^{-3}\, \mathrm{yr}^{-1} $ and $3.34_{-3.29}^{+16.7}$ $\mathrm{Gpc}^{-3}\, \mathrm{yr}^{-1} $ for two different jet profiles, consistent with pre-2017 estimates of the isotropic sGRB rate. Our study shows how jet structure can impact rate estimations and could allow one to test structured jet profiles. We finally show that modelling the maximum structured jet viewing angle with redshift can transform a cosmic beaming uncorrected rate to a representative estimate of the binary neutron star merger rate. We suggest this framework can be used to demonstrate parity with merger rates or to yield estimates of the successful jet fraction of sGRBs., Comment: Submitted to MNRAS

Many data sets cannot be accurately described by standard probability distributions due to the excess number of zero values present. For example, zero-inflation is prevalent in microbiome data and single-cell RNA sequencing data, which serve as our real data examples. Several models have been proposed to address zero-inflated datasets including the zero-inflated negative binomial, hurdle negative binomial model, and the truncated latent Gaussian copula model. This study aims to compare various models and determine which one performs optimally under different conditions using both simulation studies and real data analyses. We are particularly interested in investigating how dependence among the variables, level of zero-inflation or deflation, and variance of the data affects model selection.

Inflammation of the umbilical cord can be seen as a result of ascending intrauterine infection or other inflammatory stimuli. Acute fetal inflammatory response (FIR) is characterized by infiltration of the umbilical cord by fetal neutrophils, and can be associated with neonatal sepsis or fetal inflammatory response syndrome. Recent advances in deep learning in digital pathology have demonstrated favorable performance across a wide range of clinical tasks, such as diagnosis and prognosis. In this study we classified FIR from whole slide images (WSI). We digitized 4100 histological slides of umbilical cord stained with hematoxylin and eosin(H&E) and extracted placental diagnoses from the electronic health record. We build models using attention-based whole slide learning models. We compared strategies between features extracted by a model (ConvNeXtXLarge) pretrained on non-medical images (ImageNet), and one pretrained using histopathology images (UNI). We trained multiple iterations of each model and combined them into an ensemble. The predictions from the ensemble of models trained using UNI achieved an overall balanced accuracy of 0.836 on the test dataset. In comparison, the ensembled predictions using ConvNeXtXLarge had a lower balanced accuracy of 0.7209. Heatmaps generated from top accuracy model appropriately highlighted arteritis in cases of FIR 2. In FIR 1, the highest performing model assigned high attention to areas of activated-appearing stroma in Wharton's Jelly. However, other high-performing models assigned attention to umbilical vessels. We developed models for diagnosis of FIR from placental histology images, helping reduce interobserver variability among pathologists. Future work may examine the utility of these models for identifying infants at risk of systemic inflammatory response or early onset neonatal sepsis.

A detailed description of the asymptotic behaviour in the Belinski-Khalatnikov-Lifshitz (BKL) scenario is presented through a simple geometric picture illustrating the geometry of their ordinary differential equations (ODE), which describe a neighbourhood of the cosmic singularity. The Lagrangian version of the dynamics governed by these equations is described in terms of trajectories inside a conical subset of the corresponding space of the generalised velocities. The calculations confirm that the initial conditions of decreasing volume inevitably result in eventual total collapse, while oscillations along paths reflecting from a hyperboloid, similar to those predicted by Kasner's solutions, occur on the way. The exact solution, found in our previous work, proves to be the only one that shrinks to a point along a differentiable path. Therefore, its instability means that the collapse is always chaotic. It is also shown that the BKL equations are not satisfied by the Kasner solutions exactly, even in the asymptotic regime, although the precision of their approximation may be high., Comment: 21 pages, 3 figures

Neural networks has recently attracted much interest as useful representations of quantum many body ground states, which might help address the infamous sign problem. Most attention was directed at their representability properties, while possible limitations on finding the desired optimal state have not been suitably explored. By leveraging well-established results applicable in the context of infinite width, specifically regarding the renowned neural tangent kernel and conjugate kernel, a comprehensive analysis of the convergence and initialization characteristics of the method is conducted. We reveal the dependence of these characteristics on the interplay among these kernels, the Hamiltonian, and the basis used for its representation. We introduce and motivate novel performance metrics and explore the condition for their optimization. By leveraging these findings, we elucidate a substantial dependence of the effectiveness of this approach on the selected basis, demonstrating that so-called stoquastic Hamiltonians are more amenable to solution through neural networks than those suffering from a sign problem.

Many human interactions, such as political debates, are carried out in group settings, where there are arbitrarily many participants, each with different views and agendas. To explore such complex social settings, we present SAUCE: a customizable Python platform, allowing researchers to plug-and-play various LLMs participating in discussions on any topic chosen by the user. Our platform takes care of instantiating the models, scheduling their responses, managing the discussion history, and producing a comprehensive output log, all customizable through configuration files, requiring little to no coding skills. A novel feature of SAUCE is our asynchronous communication feature, where models decide when to speak in addition to what to say, thus modeling an important facet of human communication. We show SAUCE's attractiveness in two initial experiments, and invite the community to use it in simulating various group simulations., Comment: https://github.com/Deep-Cognition-Lab/SAUCE

The placenta forms a critical barrier to infection through pregnancy, labor and, delivery. Inflammatory processes in the placenta have short-term, and long-term consequences for offspring health. Digital pathology and machine learning can play an important role in understanding placental inflammation, and there have been very few investigations into methods for predicting and understanding Maternal Inflammatory Response (MIR). This work intends to investigate the potential of using machine learning to understand MIR based on whole slide images (WSI), and establish early benchmarks. To that end, we use Multiple Instance Learning framework with 3 feature extractors: ImageNet-based EfficientNet-v2s, and 2 histopathology foundation models, UNI and Phikon to investigate predictability of MIR stage from histopathology WSIs. We also interpret predictions from these models using the learned attention maps from these models. We also use the MIL framework for predicting white blood cells count (WBC) and maximum fever temperature ($T_{max}$). Attention-based MIL models are able to classify MIR with a balanced accuracy of up to 88.5% with a Cohen's Kappa ($\kappa$) of up to 0.772. Furthermore, we found that the pathology foundation models (UNI and Phikon) are both able to achieve higher performance with balanced accuracy and $\kappa$, compared to ImageNet-based feature extractor (EfficientNet-v2s). For WBC and $T_{max}$ prediction, we found mild correlation between actual values and those predicted from histopathology WSIs. We used MIL framework for predicting MIR stage from WSIs, and compared effectiveness of foundation models as feature extractors, with that of an ImageNet-based model. We further investigated model failure cases and found them to be either edge cases prone to interobserver variability, examples of pathologist's overreach, or mislabeled due to processing errors.

Contrastive learning methods, such as CLIP, leverage naturally paired data-for example, images and their corresponding text captions-to learn general representations that transfer efficiently to downstream tasks. While such approaches are generally applied to two modalities, domains such as robotics, healthcare, and video need to support many types of data at once. We show that the pairwise application of CLIP fails to capture joint information between modalities, thereby limiting the quality of the learned representations. To address this issue, we present Symile, a simple contrastive learning approach that captures higher-order information between any number of modalities. Symile provides a flexible, architecture-agnostic objective for learning modality-specific representations. To develop Symile's objective, we derive a lower bound on total correlation, and show that Symile representations for any set of modalities form a sufficient statistic for predicting the remaining modalities. Symile outperforms pairwise CLIP, even with modalities missing in the data, on cross-modal classification and retrieval across several experiments including on an original multilingual dataset of 33M image, text and audio samples and a clinical dataset of chest X-rays, electrocardiograms, and laboratory measurements. All datasets and code used in this work are publicly available at https://github.com/rajesh-lab/symile., Comment: NeurIPS 2024

Artificial intelligence may assist healthcare systems in meeting increasing demand for pathology services while maintaining diagnostic quality and reducing turnaround time and costs. We aimed to investigate the performance of an institutionally developed system for prostate cancer detection, grading, and workflow optimization and to contrast this with commercial alternatives. From August 2021 to March 2023, we scanned 21,396 slides from 1,147 patients with positive biopsies. We developed models for cancer detection, grading, and screening of equivocal cases for IHC ordering. We compared a task-specific model trained using the PANDA dataset of prostate cancer biopsies with one built using features extracted by the general-purpose histology foundation model, UNI and compare their performance in an unfiltered prospectively collected dataset that reflects our patient population (1737 slides,95 patients). We evaluated the contributions of a bespoke model designed to improve sensitivity in detecting small cancer foci and scoring of broader patterns observed at lower resolution. We found high concordance between the developed systems and pathologist reference in detection (AUC 98.5, sensitivity 95.0, and specificity 97.8), ISUP grading (quadratic Cohen's kappa 0.869), grade group 3 or higher (AUC 97.5, sensitivity 94.9, specificity 96.6) and comparable to published data from commercial systems. Screening could reduce IHC ordering for equivocal cases by 44.5% with an overall error rate of 1.8% (1.4% false positive, 0.4% false negative rates). Institutions like academic medical centers that have high scanning volumes and report abstraction capabilities can develop accurate computational pathology models for internal use. These models have the potential to aid in quality control role and to improve workflow in the pathology lab to help meet future challenges in prostate cancer diagnosis.

A likelihood analysis of the observables in deeply virtual exclusive photoproduction off a proton target, $ep \rightarrow e' p' \gamma'$, is presented. Two processes contribute to the reaction: deeply virtual Compton scattering, where the photon is produced at the proton vertex, and the Bether-Heitler process, where the photon is radiated from the electron. We consider the unpolarized process for which the largest amount of data with all the kinematic dependences are available from corresponding datasets with unpolarized beams and unpolarized targets from Jefferson Lab. We provide and use a method which derives a joint likelihood of the Compton form factors, which parametrize the deeply virtual Compton scattering amplitude in QCD, for each observed combination of the kinematic variables defining the reaction. The unpolarized twist-two cross section likelihood fully constrains only three of the Compton form factors (CFFs). The impact of the twist-three corrections to the analysis is also explored. The derived likelihoods are explored using Markov chain Monte Carlo (MCMC) methods. Using our proposed method we derive CFF error bars and covariances. Additionally, we explore methods which may reduce the magnitude of error bars/contours in the future., Comment: 22 pages, 11 figures

Instantons, semi-classical trajectories of quantum tunneling in imaginary time, have long been used to study thermodynamic and transport properties in a myriad of condensed matter and high energy systems. A recent experiment in superconducting circuits [Phys. Rev. Lett. 126, 197701, (2021)] provided first evidence for direct dynamical signatures of instantons (phase slips), manifested by order-unity inelastic decay probabilities for photons with which they interact, motivating the development of a scattering theory of instantons [Phys. Rev. Lett. 126, 137701, (2021)]. While this framework successfully predicted the measured inelastic decay rates of the photons for several experimental devices, it is valid only if the tunneling time of the instantons is much shorter than the relaxation time of the environment in which they are embedded, and requires a closed analytical expression for the instanton trajectory. Here, we amend these issues by incorporating numerical methods that lift some of the previously applied approximations. Our results agree with the experimental measurements, also for devices with shorter relaxation times, without fitting parameters. This framework should be useful in many other quantum field theoretical contexts., Comment: 15 pages, 4 figures. The first two authors contributed equally

Despite being a pillar of quantum mechanics, little attention has been paid to the onset of Fermi's golden rule as a discrete microscopic bath of modes approaches the macroscopic thermodynamic limit and forms a continuum. Motivated by recent experiments in circuit quantum electrodynamics, we tackle this question through the lens of single-photon decay in a finite transmission line coupled to a qubit ("quantum impurity"). We consider a single-photon state, coupled via the nonlinear impurity to several baths formed by multi-photon states with different number of photons, which are inherently discrete due to the finite size of the line. We focus on the late-time dynamics of the single-photon, and uncover the conditions under which the photon's decoherence rate approaches the decay rate predicted by Fermi's golden rule. We show that it is necessary to keep a small but finite escape rate (unrelated to the impurity) for each single-photon mode to obtain a finite long-time decay rate. We analyze the contribution of the baths formed by many-body states with different number of photons, and illustrate how the decay rate induced by some bath of $n$ photon states is enhanced by the presence of other baths of $m \neq n$ photon states, highlighting the contribution of cascade photon decay processes. Our formalism could be used to analyze recent experiments in superconducting circuits., Comment: 14 pages, 5 figures

Understanding the inner workings of Transformers is crucial for achieving more accurate and efficient predictions. In this work, we analyze the computation performed by Transformers in the layers after the top-1 prediction has become fixed, which has been previously referred to as the "saturation event". We expand the concept of saturation events for top-k tokens, demonstrating that similar saturation events occur across language, vision, and speech models. We find that these saturation events happen in order of the corresponding tokens' ranking, i.e., the model first decides on the top ranking token, then the second highest ranking token, and so on. This phenomenon seems intrinsic to the Transformer architecture, occurring across different architectural variants (decoder-only, encoder-only, and to a lesser extent full-Transformer), and even in untrained Transformers. We propose an underlying mechanism of task transition for this sequential saturation, where task k corresponds to predicting the k-th most probable token, and the saturation events are in fact discrete transitions between the tasks. In support of this we show that it is possible to predict the current task from hidden layer embedding. Furthermore, using an intervention method we demonstrate that we can cause the model to switch from one task to the next. Finally, leveraging our findings, we introduce a novel token-level early-exit strategy, which surpasses existing methods in balancing performance and efficiency.

Background/Objectives: Cataract surgery, a very common and critical procedure for restoring vision, has outcomes that can vary based on patient demographics. This study aimed to elucidate the effects of age and sex on the risk factors, intraoperative complications, and postoperative outcomes of cataract surgery. Subjects/Methods: Conducted as a single-center retrospective cohort study, it analyzed 691 eyes from 589 individuals who underwent surgery at a tertiary referral center, utilizing data from electronic medical records to assess preoperative risk factors, intraoperative complications, and pre- and post-operative best corrected visual acuity (BCVA) along with demographic data. Results: The main results highlighted that males aged 65-75 years exhibited significantly higher rates of functional postoperative BCVA (91% for males vs. 79% for females, p=0.007), a disparity that is not explained by differences in surgical complications or risk factor prevalence. Furthermore, the study identified age-specific thresholds where BCVA improvements significantly declined beyond 65 years for females and 75 years for males. The likelihood of worsened BCVA post-surgery increased with age for both sexes, with a significant decline in BCVA improvement transitioning from 55-65 years to 65-75 years age groups. Conclusions: The findings underscore the critical influence of both sex and age on cataract surgery outcomes, revealing significant sex-specific age thresholds that signal lesser improvements in postoperative BCVA. These insights advocate for the integration of patient age and sex into preoperative evaluations to better tailor the timing and planning of cataract surgery, ultimately aiming to optimize clinical outcomes.

It is generally assumed that existing artificial systems are not phenomenally conscious, and that the construction of phenomenally conscious artificial systems would require significant technological progress if it is possible at all. We challenge this assumption by arguing that if Global Workspace Theory (GWT) - a leading scientific theory of phenomenal consciousness - is correct, then instances of one widely implemented AI architecture, the artificial language agent, might easily be made phenomenally conscious if they are not already. Along the way, we articulate an explicit methodology for thinking about how to apply scientific theories of consciousness to artificial systems and employ this methodology to arrive at a set of necessary and sufficient conditions for phenomenal consciousness according to GWT.

Many real-world applications of tabular data involve using historic events to predict properties of new ones, for example whether a credit card transaction is fraudulent or what rating a customer will assign a product on a retail platform. Existing approaches to event prediction include costly, brittle, and application-dependent techniques such as time-aware positional embeddings, learned row and field encodings, and oversampling methods for addressing class imbalance. Moreover, these approaches often assume specific use-cases, for example that we know the labels of all historic events or that we only predict a pre-specified label and not the data's features themselves. In this work, we propose a simple but flexible baseline using standard autoregressive LLM-style transformers with elementary positional embeddings and a causal language modeling objective. Our baseline outperforms existing approaches across popular datasets and can be employed for various use-cases. We demonstrate that the same model can predict labels, impute missing values, or model event sequences., Comment: 10 pages, 6 pages of references+appendix

We give an overview of the science objectives and mission design of the Spectroscopic Time-Resolving Observatory for Broadband Energy X-rays (STROBE-X) observatory, which has been proposed as a NASA probe-class (~$1.5B) mission in response to the Astro2020 recommendation for an X-ray probe., Comment: 11 pages, 5 figures, accepted for publication in JATIS

In this work we study the Kondo impurity problem - at a finite concentration of impurities. We identify two parameter regimes for the Kondo impurity problem. 1) The single impurity limit, where the concentration of Kondo impurities is so low that the background scattering mechanisms (non-magnetic impurities, Umklapp scattering, etc.) of the metal considered are the dominant conduction electron scattering mechanisms at zero temperature. 2) The dilute impurity system limit where the concentration of magnetic impurities is such that they form the dominant mechanism of conduction electron scattering at zero temperature of the metal in question (this is accompanied by a variety of easily detectable Kondo signatures (resistance minimum, specific heat measurements, magnetization as a function of external magnetic field, conduction electron dephasing rates as well as ARPES, RIXS and NMR spectroscopies)) while still being very dilute. Most theoretical efforts are currently in regime where a single isolated impurity is considered - regime 1) while most experimental efforts are in regime 2). We present analytical evidence that this explains the well known discrepancy between experiment and theory as to the value of the Kondo temperature. We find that the ratio between the two Kondo temperatures in regime 1) and regime 2) is given by: $\mathcal{R}=\exp\left[\frac{\pi^{2}\rho v_{F}}{2k_{F}^{2}Vol}\right]$ where $\rho$ is the density of states, $v_{F}$ is the fermi velocity, and $k_{F}$ is the Fermi wavevector and $Vol$ is the volume of a unit cell. We note that there is no dependence on the impurity concentration in this ratio so it is possible to define a single Kondo temperature for limit 2) for the dilute Kondo impurity system. In this work we present results within the Reed-Newns Kondo meanfield approximation and to leading order of the linked cluster expansion., Comment: Comments welcome

The prompt spectra of gamma-ray bursts are known to follow broadband continuum behavior over decades in energy. GRB 221009A, given the moniker the brightest of all time (BOAT), is the brightest gamma-ray burst identified in half a century of observations, and was first identified by the Fermi Gamma-ray Burst Monitor (GBM). On behalf of the Fermi-GBM Team, Lesage et al. (2023) described the initial GBM analysis. Ravasio et al. (2024) report the identification of a spectral line in part of the prompt emission of this burst, which they describe as evolving over 80 s from $\sim$12 MeV to 6 MeV. We report a GBM Team analysis on the Ravasio Line: 1) We cannot identify an instrumental effect that could have produced this signal, and 2) our method of calculating the statistical significance of the line shows it easily exceeds the 5$\sigma$ discovery threshold. We additionally comment on the claim of the line beginning at earlier time intervals, up to 37 MeV, as reported in Zhang et al. (2024). We find that it is reasonable to utilize these measurements for characterization of the line evolution, with caution. We encourage theoretical studies exploring this newly discovered gamma-ray burst spectral feature, unless any rigorous alternative explanation unrelated to the emission from GRB 221009A is identified.

We present a quantum alternating operator ansatz (QAOA$^+$) that solves a class of linearly constrained optimization problems by evolving a quantum state within a Hilbert subspace of feasible problem solutions. Our main focus is on a class of problems with a linear constraint containing sequential integer coefficients. For problems in this class, we devise QAOA$^+$ circuits that provably converge to the optimal solution as the number of circuit layers increases, generalizing previous guarantees for solving unconstrained problems or problems with symmetric constraints. Our approach includes asymmetric ``mixing" Hamiltonians that drive transitions between feasible states, as well as a method to incorporate an arbitrary known feasible solution as the initial state, each of which can be applied beyond the specific linear constraints considered here. This analysis extends QAOA$^+$ performance guarantees to a more general set of linearly-constrained problems and provides tools for future generalizations., Comment: 15+9 pages, 4+6 figures

While generalization over tasks from easy to hard is crucial to profile language models (LLMs), the datasets with fine-grained difficulty annotations for each problem across a broad range of complexity are still blank. Aiming to address this limitation, we present Easy2Hard-Bench, a consistently formatted collection of 6 benchmark datasets spanning various domains, such as mathematics and programming problems, chess puzzles, and reasoning questions. Each problem within these datasets is annotated with numerical difficulty scores. To systematically estimate problem difficulties, we collect abundant performance data on attempts to each problem by humans in the real world or LLMs on the prominent leaderboard. Leveraging the rich performance data, we apply well-established difficulty ranking systems, such as Item Response Theory (IRT) and Glicko-2 models, to uniformly assign numerical difficulty scores to problems. Moreover, datasets in Easy2Hard-Bench distinguish themselves from previous collections by a higher proportion of challenging problems. Through extensive experiments with six state-of-the-art LLMs, we provide a comprehensive analysis of their performance and generalization capabilities across varying levels of difficulty, with the aim of inspiring future research in LLM generalization. The datasets are available at https://huggingface.co/datasets/furonghuang-lab/Easy2Hard-Bench., Comment: NeurIPS 2024 Datasets and Benchmarks Track

The resonant conversion of cosmic microwave background (CMB) photons into axions within large-scale structure induces an anisotropic spectral distortion in CMB temperature maps. Applying state-of-the-art foreground cleaning techniques to $\textit{Planck}$ CMB observations, we construct maps of axion-induced "patchy screening" of the CMB. We cross-correlate these maps with data from the $\textit{unWISE}$ galaxy survey and find no evidence of axions. We constrain the axion-photon coupling, $g_{a\gamma\gamma} \lesssim 2 \times 10^{-12}~{\rm GeV}^{-1}$, at the 95% confidence level for axion masses in the range $10^{-13}~{\rm eV} \lesssim m_a \lesssim 10^{-12}~{\rm eV}$. These constraints are competitive with the tightest astrophysical axion limits in this mass range and are inferred from robust population-level statistics, which makes them complementary to existing searches that rely on modeling of individual systems., Comment: 5+15 pages; 3+15 figures

We present the detection and analysis of GRB 231115A, a candidate extragalactic magnetar giant flare (MGF) observed by Fermi/GBM and localized by INTEGRAL to the starburst galaxy M82. This burst exhibits distinctive temporal and spectral characteristics that align with known MGFs, including a short duration and a high peak energy. Gamma-ray analyses reveal significant insights into this burst, supporting conclusions already established in the literature: our time-resolved spectral studies provide further evidence that GRB 231115A is indeed a MGF. Significance calculations also suggest a robust association with M82, further supported by a high Bayes factor that minimizes the probability of chance alignment with a neutron star merger. Despite extensive follow-up efforts, no contemporaneous gravitational wave or radio emissions were detected. The lack of radio emission sets stringent upper limits on possible radio luminosity. Constraints from our analysis show no fast radio bursts (FRBs) associated with two MGFs. X-ray observations conducted post-burst by Swift/XRT and XMM/Newton provided additional data, though no persistent counterparts were identified. Our study underscores the importance of coordinated multi-wavelength follow-up and highlights the potential of MGFs to enhance our understanding of short GRBs and magnetar activities in the cosmos. Current MGF identification and follow-up implementation are insufficient for detecting expected counterparts; however, improvements in these areas may allow for the recovery of follow-up signals with existing instruments. Future advancements in observational technologies and methodologies will be crucial in furthering these studies.

Using Parker Solar Probe data from orbits 8 through 17, we examine fluctuation amplitudes throughout the critical region where the solar wind flow speed approaches and then exceeds the Alfv\'en wave speed, taking account of various exigencies of the plasma data. In contrast to WKB theory for non-interacting Alfv\'en waves streaming away from the Sun, the magnetic and kinetic fluctuation energies per unit volume are not monotonically decreasing. Instead, there is clear violation of conservation of standard WKB wave action, which is consistent with previous indications of strong in-situ fluctuation energy input in the solar wind near the Alfv\'en critical region. This points to strong violations of WKB theory due to nonlinearity (turbulence) and major energy input near the critical region, which we interpret as likely due to driving by large-scale coronal shear flows.