Search

Strong lensing of gravitational wave (GW) sources allows the observer to see the GW source from different lines-of-sight (LOS) through the corresponding images, which provides a way for constraining the relative proper motion of the GW source. This is possible as the GW signals received from each image will have slightly different projected velocity components, from which one can `Doppler-Triangulate' for the GW source velocity vector. The difference in projected velocity between the different images can be observationally inferred through pairwise GW phase measurements that accumulate over the time-of-observation. In this paper we study lensed eccentric GW sources and explore how the observable GW phase shift between images evolve as a function of time, eccentricity, lens- and binary parameters. Next generation GW observatories, including the Einstein Telescope and Cosmic Explorer, will see $\sim $hundreds/year of lensed GW sources, where a significant fraction of these are expected to be eccentric. We discuss the expected unique observables for such eccentric lensed GW sources, and the relation to their observable relative linear motion, which otherwise is exceedingly difficult to constrain in general., Comment: 7 pages. Comments welcome

A computer's clock rate ultimately determines the minimum time between sequential operations or instructions. Despite exponential advances in electronic computer performance owing to Moore's Law and increasingly parallel system architectures, computer clock rates have remained stagnant at $\sim5~\mathrm{GHz}$ for almost two decades. This poses an intractable problem for applications requiring real-time processing or control of ultrafast information systems. Here we break this barrier by proposing and experimentally demonstrating computing based on an end-to-end and all-optical recurrent neural network harnessing the ultrafast nature of linear and nonlinear optical operations while avoiding electronic operations. The all-optical computer realizes linear operations, nonlinear functions, and memory entirely in the optical domain with $>100~\mathrm{GHz}$ clock rates. We experimentally demonstrate a prototypical task of noisy waveform classification as well as perform ultrafast in-situ analysis of the soliton states from integrated optical microresonators. We further illustrate the application of the architecture for generative artificial intelligence based on quantum fluctuations to generate images even in the absence of input optical signals. Our results highlight the potential of all-optical computing beyond what can be achieved with digital electronics by utilizing ultrafast linear, nonlinear, and memory functions and quantum fluctuations.

Cosmological solitons are widely predicted by scenarios of the early Universe. In this work, we investigate the isotropic background and anisotropies of gravitational waves (GWs) induced by soliton isocurvature perturbations, especially considering the effects of non-Gaussianity in these perturbations. Regardless of non-Gaussianity, the energy-density fraction spectrum of isocurvature-induced GWs approximately has a universal shape within the perturbative regime, thus serving as a distinctive signal of solitons. We derive the angular power spectrum of isocurvature-induced GWs to characterize their anisotropies. Non-Gaussianity plays a key role in generating anisotropies through the couplings between large- and small-scale isocurvature perturbations, making the angular power spectrum to be a powerful probe of non-Gaussianity. Moreover, the isocurvature-induced GWs have nearly no cross-correlations with the cosmic microwave background, providing a new observable to distinguish them from other GW sources, e.g., GWs induced by cosmological curvature perturbations enhanced at small scales. Therefore, detection of both the isotropic background and anisotropies of isocurvature-induced GWs could reveal important implications for the solitons as well as the early Universe., Comment: 20 pages, 5 figures

The availability of challenging simulation environments is pivotal for advancing the field of Multi-Agent Reinforcement Learning (MARL). In cooperative MARL settings, the StarCraft Multi-Agent Challenge (SMAC) has gained prominence as a benchmark for algorithms following centralized training with decentralized execution paradigm. However, with continual advancements in SMAC, many algorithms now exhibit near-optimal performance, complicating the evaluation of their true effectiveness. To alleviate this problem, in this work, we highlight a critical issue: the default opponent policy in these environments lacks sufficient diversity, leading MARL algorithms to overfit and exploit unintended vulnerabilities rather than learning robust strategies. To overcome these limitations, we propose SMAC-HARD, a novel benchmark designed to enhance training robustness and evaluation comprehensiveness. SMAC-HARD supports customizable opponent strategies, randomization of adversarial policies, and interfaces for MARL self-play, enabling agents to generalize to varying opponent behaviors and improve model stability. Furthermore, we introduce a black-box testing framework wherein agents are trained without exposure to the edited opponent scripts but are tested against these scripts to evaluate the policy coverage and adaptability of MARL algorithms. We conduct extensive evaluations of widely used and state-of-the-art algorithms on SMAC-HARD, revealing the substantial challenges posed by edited and mixed strategy opponents. Additionally, the black-box strategy tests illustrate the difficulty of transferring learned policies to unseen adversaries. We envision SMAC-HARD as a critical step toward benchmarking the next generation of MARL algorithms, fostering progress in self-play methods for multi-agent systems. Our code is available at https://github.com/devindeng94/smac-hard.

Observations of strongly gravitationally lensed gravitational wave (GW) sources provide a unique opportunity for constraining their transverse motion, which otherwise is exceedingly hard for GW mergers in general. Strong lensing makes this possible when two or more images of the lensed GW source are observed, as each image essentially allows the observer to see the GW source from different directional lines-of-sight. If the GW source is moving relative to the lens and observer, the observed GW signal from one image will therefore generally appear blue- or redshifted compared to GW signal from the other image. This velocity induced differential Doppler shift gives rise to an observable GW phase shift between the GW signals from the different images, which provides a rare glimpse into the relative motion of GW sources and their host environment across redshift. We illustrate that detecting such GW phase shifts is within reach of next-generation ground-based detectors such as Einstein Telescope, that is expected to detect $\sim$hundreds of lensed GW mergers per year. This opens up completely new ways of inferring the environment of GW sources, as well as studying cosmological velocity flows across redshift., Comment: comments welcome

We introduce DA-Code, a code generation benchmark specifically designed to assess LLMs on agent-based data science tasks. This benchmark features three core elements: First, the tasks within DA-Code are inherently challenging, setting them apart from traditional code generation tasks and demanding advanced coding skills in grounding and planning. Second, examples in DA-Code are all based on real and diverse data, covering a wide range of complex data wrangling and analytics tasks. Third, to solve the tasks, the models must utilize complex data science programming languages, to perform intricate data processing and derive the answers. We set up the benchmark in a controllable and executable environment that aligns with real-world data analysis scenarios and is scalable. The annotators meticulously design the evaluation suite to ensure the accuracy and robustness of the evaluation. We develop the DA-Agent baseline. Experiments show that although the baseline performs better than other existing frameworks, using the current best LLMs achieves only 30.5% accuracy, leaving ample room for improvement. We release our benchmark at https://da-code-bench.github.io., Comment: EMNLP 2024

Let \mu_{M,D} be the self-similar measure generated by the positive integer M=RN^q and the product-form digit set D=\{0,1,\dots,N-1\}\oplus N^{p_1}\{0,1,\dots,N-1\}\oplus \cdots \oplus N^{p_s}\{0,1,\dots,N-1\}, where R>1, N>1, q, p_i(1\leq i\leq s) are positive integers with gcd(R,N)=1 and p_1

Obesity is widely recognized as a critical and pervasive health concern. We strive to identify important genetic risk factors from hundreds of thousands of single nucleotide polymorphisms (SNPs) for obesity. We propose and apply a novel Quantile Regression with Insight Fusion (QRIF) approach that can integrate insights from established studies or domain knowledge to simultaneously select variables and modeling for ultra-high dimensional genetic data, focusing on high conditional quantiles of body mass index (BMI) that are of most interest. We discover interesting new SNPs and shed new light on a comprehensive view of the underlying genetic risk factors for different levels of BMI. This may potentially pave the way for more precise and targeted treatment strategies. The QRIF approach intends to balance the trade-off between the prior insights and the observed data while being robust to potential false information. We further establish the desirable asymptotic properties under the challenging non-differentiable check loss functions via Huber loss approximation and nonconvex SCAD penalty via local linear approximation. Finally, we develop an efficient algorithm for the QRIF approach. Our simulation studies further demonstrate its effectiveness.

Model-based methods in reinforcement learning offer a promising approach to enhance data efficiency by facilitating policy exploration within a dynamics model. However, accurately predicting sequential steps in the dynamics model remains a challenge due to the bootstrapping prediction, which attributes the next state to the prediction of the current state. This leads to accumulated errors during model roll-out. In this paper, we propose the Any-step Dynamics Model (ADM) to mitigate the compounding error by reducing bootstrapping prediction to direct prediction. ADM allows for the use of variable-length plans as inputs for predicting future states without frequent bootstrapping. We design two algorithms, ADMPO-ON and ADMPO-OFF, which apply ADM in online and offline model-based frameworks, respectively. In the online setting, ADMPO-ON demonstrates improved sample efficiency compared to previous state-of-the-art methods. In the offline setting, ADMPO-OFF not only demonstrates superior performance compared to recent state-of-the-art offline approaches but also offers better quantification of model uncertainty using only a single ADM.

Silk damping is well known in the study of cosmic microwave background (CMB) and accounts for suppression of the angular power spectrum of CMB on large angular multipoles. In this Letter, we study the effect of Silk damping on the scalar-induced gravitational waves (SIGWs). Resulting from the dissipation of cosmic fluid, the Silk damping notably suppresses the energy-density spectrum of SIGWs on scales comparable to a diffusion scale at the decoupling time of feebly-interacting particles. The effect offers a novel observable for probing the underlying particle interaction, especially for those mediated by heavy gauge bosons beyond the standard model of particles. We anticipate that pulsar timing arrays are sensitive to gauge bosons with mass $\sim10^{3}-10^{4}\,\mathrm{GeV}$, while space- and ground-based interferometers to those with mass $\sim10^7-10^{12}\,\mathrm{GeV}$, leading to essential complements to on-going and future experiments of high-energy physics., Comment: 7 pages, 2 figures, additional references added

Although supervised machine learning is popular for information extraction from clinical notes, creating large annotated datasets requires extensive domain expertise and is time-consuming. Meanwhile, large language models (LLMs) have demonstrated promising transfer learning capability. In this study, we explored whether recent LLMs can reduce the need for large-scale data annotations. We curated a manually-labeled dataset of 769 breast cancer pathology reports, labeled with 13 categories, to compare zero-shot classification capability of the GPT-4 model and the GPT-3.5 model with supervised classification performance of three model architectures: random forests classifier, long short-term memory networks with attention (LSTM-Att), and the UCSF-BERT model. Across all 13 tasks, the GPT-4 model performed either significantly better than or as well as the best supervised model, the LSTM-Att model (average macro F1 score of 0.83 vs. 0.75). On tasks with high imbalance between labels, the differences were more prominent. Frequent sources of GPT-4 errors included inferences from multiple samples and complex task design. On complex tasks where large annotated datasets cannot be easily collected, LLMs can reduce the burden of large-scale data labeling. However, if the use of LLMs is prohibitive, the use of simpler supervised models with large annotated datasets can provide comparable results. LLMs demonstrated the potential to speed up the execution of clinical NLP studies by reducing the need for curating large annotated datasets. This may result in an increase in the utilization of NLP-based variables and outcomes in observational clinical studies.