Search

Neutron star (NS) has many extreme physical conditions, and one may obtain some important informations about NS via accreting neutron star binary (ANSB) systems. The upcoming Chinese Space Station Telescope (CSST) provides an opportunity to search for a large sample of ANSB candidates. Our goal is to check the completeness of the potential ANSB samples from CSST data. In this paper, we generate some ANSBs and normal binaries under CSST photometric system by binary evolution and binary population synthesis method and use a machine learning method to train a classification model. Although the Precision ($94.56~ \%$) of our machine learning model is as high as before study, the Recall is only about $63.29~ \%$. The Precision/Recall is mainly determined by the mass transfer rate between the NSs and their companions. In addition, we also find that the completeness of ANSB samples from CSST photometric data by the machine learning method also depends on the companion mass and the age of the system. ANSB candidates with low initial mass companion star ($0.1~ {\rm M}_\odot$ to $1~ {\rm M}_\odot$) have a relatively high Precision ($94.94~ \%$) and high Recall ($86.32~ \%$), whereas ANSB candidates with higher initial mass companion star ($1.1~ {\rm M}_\odot$ to $3~ {\rm M}_\odot$) have similar Precision ($93.88~ \%$) and quite low Recall ($42.67~ \%$). Our results indicate that although the machine learning method may obtain a relative pure sample of ANSBs, a completeness correction is necessary for one to obtain a complete sample., Comment: 12 pages, 5 figures

We consider the effects of relaxing the assumption that gravitational waves composing the stochastic gravitational wave background (SGWB) are uncorrelated between frequencies in analyses of the data from Pulsar Timing Arrays (PTAs). While individual monochromatic plane waves are often a good approximation, a background composed of astrophysical sources cannot be monochromatic since an infinite plane wave carries no signal. We consider how relaxing this assumption allows us to extract potential information about modified dispersion relations and other fundamental physics questions, as both the group and phase velocity of waves become relevant. After developing the formalism we carry out simple Gaussian wavepacket examples and then consider more realistic waveforms, such as that from binary inspirals. When the frequency evolves only slowly across the PTA temporal baseline, the monochromatic assumption at an effective mean frequency remains a good approximation and we provide scaling relations that characterize its accuracy., Comment: 30 pages, 6 figures

Pre-trained language models are increasingly being used in multi-document summarization tasks. However, these models need large-scale corpora for pre-training and are domain-dependent. Other non-neural unsupervised summarization approaches mostly rely on key sentence extraction, which can lead to information loss. To address these challenges, we propose a lightweight yet effective unsupervised approach called GLIMMER: a Graph and LexIcal features based unsupervised Multi-docuMEnt summaRization approach. It first constructs a sentence graph from the source documents, then automatically identifies semantic clusters by mining low-level features from raw texts, thereby improving intra-cluster correlation and the fluency of generated sentences. Finally, it summarizes clusters into natural sentences. Experiments conducted on Multi-News, Multi-XScience and DUC-2004 demonstrate that our approach outperforms existing unsupervised approaches. Furthermore, it surpasses state-of-the-art pre-trained multi-document summarization models (e.g. PEGASUS and PRIMERA) under zero-shot settings in terms of ROUGE scores. Additionally, human evaluations indicate that summaries generated by GLIMMER achieve high readability and informativeness scores. Our code is available at https://github.com/Oswald1997/GLIMMER., Comment: 19 pages, 7 figures. Accepted by ECAI 2024

Recent breakthroughs have ushered the quantum network into a new era, where quantum information can be stored, transferred, and processed across multiple nodes on a metropolitan scale. A key challenge in this new era is enhancing the capabilities of individual nodes, providing precise and robust control over multiple qubits and advanced functionality for scalable quantum networks. Here, we report on precise and complex control in a hybrid quantum node based on a diamond color center. We demonstrate hybrid coherent control by entangling three types of qubits: an electron spin as an interface qubit, a nuclear spin with long memory time, and a flying photonic qubit, with their qubit frequencies spanning three distinct regimes from the optical domain to the rf domain. By incorporating two additional memory qubits, we encode three memory qubits into a logical state using the three-qubit repetition code and entangle this logical qubit with a photonic qubit. Leveraging hybrid qubits and precise control, we repeatedly read out the error syndromes of memory qubits through the electron spin, serving as an auxiliary qubit, then apply a real-time feedback operation to correct bit-flip errors. We execute and verify active error correction for up to twelve rounds and demonstrate the improvement over the uncorrected counterpart. Our results demonstrate the feasibility of several key functionalities for next-generation quantum repeaters, paving the way towards full-fledged metropolitan-scale quantum networks for a wide range of practical applications., Comment: 8 pages, 3 figures

The impressive performance of Large Language Models (LLMs) across various natural language processing tasks comes at the cost of vast computational resources and storage requirements. One-shot pruning techniques offer a way to alleviate these burdens by removing redundant weights without the need for retraining. Yet, the massive scale of LLMs often forces current pruning approaches to rely on heuristics instead of optimization-based techniques, potentially resulting in suboptimal compression. In this paper, we introduce ALPS, an optimization-based framework that tackles the pruning problem using the operator splitting technique and a preconditioned conjugate gradient-based post-processing step. Our approach incorporates novel techniques to accelerate and theoretically guarantee convergence while leveraging vectorization and GPU parallelism for efficiency. ALPS substantially outperforms state-of-the-art methods in terms of the pruning objective and perplexity reduction, particularly for highly sparse models. On the OPT-30B model with 70% sparsity, ALPS achieves a 13% reduction in test perplexity on the WikiText dataset and a 19% improvement in zero-shot benchmark performance compared to existing methods.

The origin of intermediate helium (He)-rich hot subdwarfs are still unclear. Previous studies have suggested that some surviving Type Ia supernovae (SNe Ia) companions from the white dwarf~+~main-sequence (WD+MS) channel may contribute to the intermediate He-rich hot subdwarfs. However, previous studies ignored the impact of atomic diffusion on the post-explosion evolution of surviving companion stars of SNe Ia, leading to that they could not explain the observed surface He abundance of intermediate He-rich hot subdwarfs. In this work, by taking the atomic diffusion and stellar wind into account, we trace the surviving companions of SNe Ia from the WD+MS channel using the one-dimensional stellar evolution code \textsc{MESA} until they evolve into hot subdwarfs. We find that the surface He-abundances of our surviving companion models during their core He-burning phases are in a range of $-1 \lesssim {\rm log}(N_{\rm He}/N_{\rm H}) \lesssim 0$, which are consistent with those observed in intermediate He-rich hot subdwarfs. This seems to further support that surviving companions of SNe Ia in the WD+MS channel are possible to form some intermediate He-rich hot subdwarfs., Comment: 10 pages, 5 figures

The increasing computational demands of modern neural networks present deployment challenges on resource-constrained devices. Network pruning offers a solution to reduce model size and computational cost while maintaining performance. However, most current pruning methods focus primarily on improving sparsity by reducing the number of nonzero parameters, often neglecting other deployment costs such as inference time, which are closely related to the number of floating-point operations (FLOPs). In this paper, we propose FALCON, a novel combinatorial-optimization-based framework for network pruning that jointly takes into account model accuracy (fidelity), FLOPs, and sparsity constraints. A main building block of our approach is an integer linear program (ILP) that simultaneously handles FLOP and sparsity constraints. We present a novel algorithm to approximately solve the ILP. We propose a novel first-order method for our optimization framework which makes use of our ILP solver. Using problem structure (e.g., the low-rank structure of approx. Hessian), we can address instances with millions of parameters. Our experiments demonstrate that FALCON achieves superior accuracy compared to other pruning approaches within a fixed FLOP budget. For instance, for ResNet50 with 20% of the total FLOPs retained, our approach improves the accuracy by 48% relative to state-of-the-art. Furthermore, in gradual pruning settings with re-training between pruning steps, our framework outperforms existing pruning methods, emphasizing the significance of incorporating both FLOP and sparsity constraints for effective network pruning.

Structured pruning is a promising approach for reducing the inference costs of large vision and language models. By removing carefully chosen structures, e.g., neurons or attention heads, the improvements from this approach can be realized on standard deep learning hardware. In this work, we focus on structured pruning in the one-shot (post-training) setting, which does not require model retraining after pruning. We propose a novel combinatorial optimization framework for this problem, based on a layer-wise reconstruction objective and a careful reformulation that allows for scalable optimization. Moreover, we design a new local combinatorial optimization algorithm, which exploits low-rank updates for efficient local search. Our framework is time and memory-efficient and considerably improves upon state-of-the-art one-shot methods on vision models (e.g., ResNet50, MobileNet) and language models (e.g., OPT-1.3B -- OPT-30B). For language models, e.g., OPT-2.7B, OSSCAR can lead to $125\times$ lower test perplexity on WikiText with $2\times$ inference time speedup in comparison to the state-of-the-art ZipLM approach. Our framework is also $6\times$ -- $8\times$ faster. Notably, our work considers models with tens of billions of parameters, which is up to $100\times$ larger than what has been previously considered in the structured pruning literature.

Human pose analysis has garnered significant attention within both the research community and practical applications, owing to its expanding array of uses, including gaming, video surveillance, sports performance analysis, and human-computer interactions, among others. The advent of deep learning has significantly improved the accuracy of pose capture, making pose-based applications increasingly practical. This paper presents a comprehensive survey of pose-based applications utilizing deep learning, encompassing pose estimation, pose tracking, and action recognition.Pose estimation involves the determination of human joint positions from images or image sequences. Pose tracking is an emerging research direction aimed at generating consistent human pose trajectories over time. Action recognition, on the other hand, targets the identification of action types using pose estimation or tracking data. These three tasks are intricately interconnected, with the latter often reliant on the former. In this survey, we comprehensively review related works, spanning from single-person pose estimation to multi-person pose estimation, from 2D pose estimation to 3D pose estimation, from single image to video, from mining temporal context gradually to pose tracking, and lastly from tracking to pose-based action recognition. As a survey centered on the application of deep learning to pose analysis, we explicitly discuss both the strengths and limitations of existing techniques. Notably, we emphasize methodologies for integrating these three tasks into a unified framework within video sequences. Additionally, we explore the challenges involved and outline potential directions for future research., Comment: 47 pages

Astrometry holds the potential for testing fundamental physics through the effects of the Stochastic Gravitational Wave Background (SGWB) in the $\sim 1-100$ nHz frequency band on precision measurements of stellar positions. Such measurements are complementary to tests made possible by the detection of the SGWB using Pulsar Timing Arrays. Here, the feasibility of using astrometry for the identification of parity-violating signals within the SGWB is investigated. This is achieved by defining and quantifying a non-vanishing $EB$ correlation function within astrometric correlation functions, and investigating how one might estimate the detectability of such signals., Comment: Match the published version, adding references. 8 pages, 2 figures

This paper proposes a new task in the field of Answering Subjective Induction Question on Products (SUBJPQA). The answer to this kind of question is non-unique, but can be interpreted from many perspectives. For example, the answer to 'whether the phone is heavy' has a variety of different viewpoints. A satisfied answer should be able to summarize these subjective opinions from multiple sources and provide objective knowledge, such as the weight of a phone. That is quite different from the traditional QA task, in which the answer to a factoid question is unique and can be found from a single data source. To address this new task, we propose a three-steps method. We first retrieve all answer-related clues from multiple knowledge sources on facts and opinions. The implicit commonsense facts are also collected to supplement the necessary but missing contexts. We then capture their relevance with the questions by interactive attention. Next, we design a reinforcement-based summarizer to aggregate all these knowledgeable clues. Based on a template-controlled decoder, we can output a comprehensive and multi-perspective answer. Due to the lack of a relevant evaluated benchmark set for the new task, we construct a large-scale dataset, named SupQA, consisting of 48,352 samples across 15 product domains. Evaluation results show the effectiveness of our approach., Comment: 1. There are some errors in the data analysis table in the dataset SupQA, which needs to be corrected. 2. There is something wrong with the partial expression of the formula. 3. It will be resubmitted after modification

The $S_8$ tension between low-redshift galaxy surveys and the primary CMB signals a possible breakdown of the $\Lambda$CDM model. Recently differing results have been obtained using low-redshift galaxy surveys and the higher redshifts probed by CMB lensing, motivating a possible time-dependent modification to the growth of structure. We investigate a simple phenomenological model in which the growth of structure deviates from the $\Lambda$CDM prediction at late times, in particular as a simple function of the dark energy density. Fitting to galaxy lensing, CMB lensing, BAO, and Supernovae datasets, we find significant evidence - 2.5 - 3$\sigma$, depending on analysis choices - for a non-zero value of the parameter quantifying a deviation from $\Lambda$CDM. The preferred model, which has a slower growth of structure below $z\sim 1$, improves the joint fit to the data over $\Lambda$CDM. While the overall fit is improved, there is weak evidence for galaxy and CMB lensing favoring different changes in the growth of structure., Comment: Match the published version. Add more references. 6 pages, 5 figures, 1 table

A successful measurement of the Stochastic Gravitational Wave Background (SGWB) in Pulsar Timing Arrays (PTAs) would open up a new window through which to test the predictions of General Relativity (GR). We consider how these measurements might reveal deviations from GR by studying the overlap reduction function -- the quantity that in GR is approximated by the Hellings-Downs curve -- in some sample modifications of gravity, focusing on the generic prediction of a modified dispersion relation for gravitational waves. We find a distinct signature of such modifications to GR -- a shift in the minimum angle of the angular distribution -- and demonstrate that this shift is quantitatively sensitive to any change in the phase velocity. In a given modification of gravity, this result can be used, in some regions of parameter space, to distinguish the effect of a modified dispersion relation from that due to the presence of extra polarization modes., Comment: v2: add references and discussions. 23 pages, 4 figures

The sheer size of modern neural networks makes model serving a serious computational challenge. A popular class of compression techniques overcomes this challenge by pruning or sparsifying the weights of pretrained networks. While useful, these techniques often face serious tradeoffs between computational requirements and compression quality. In this work, we propose a novel optimization-based pruning framework that considers the combined effect of pruning (and updating) multiple weights subject to a sparsity constraint. Our approach, CHITA, extends the classical Optimal Brain Surgeon framework and results in significant improvements in speed, memory, and performance over existing optimization-based approaches for network pruning. CHITA's main workhorse performs combinatorial optimization updates on a memory-friendly representation of local quadratic approximation(s) of the loss function. On a standard benchmark of pretrained models and datasets, CHITA leads to significantly better sparsity-accuracy tradeoffs than competing methods. For example, for MLPNet with only 2% of the weights retained, our approach improves the accuracy by 63% relative to the state of the art. Furthermore, when used in conjunction with fine-tuning SGD steps, our method achieves significant accuracy gains over the state-of-the-art approaches.

Although type Ia supernovae are so important in many astrophysical field, e.g. in cosmology, their explosion mechanism and progenitor system are still unclear. In physics, the relative equivalent width (REW) of the Si II 635.5 nm absorption feature reflects the velocity interval of silicon in the supernova ejecta and then may provide constraints on the explosion mechanism of SNe Ia. In this paper, we divide the SNe Ia into broad line (BL) and normal line (NL) subsamples based on their REW of Si II 635.5 nm absorption lines around maximum light, and find that the BL SNe Ia have a dimmer mean brightness than NL ones, which possibly results from their different metallicities. However, based on the pixel statistics study on the environments of two subsamples, we do not find any significant potential difference on the environments between BL and NL SNe Ia, which implies that the explosion mechanism of SNe Ia could be independent of their progenitor populations., Comment: 17 pages, 13 figures, accepted for publication in ApJ

Early dark energy (EDE), whose cosmological role is localized in time around the epoch of matter-radiation equality in order to resolve the Hubble tension, introduces a new coincidence problem: why should the EDE dynamics occur near equality if EDE is decoupled from both matter and radiation? The resolution of this problem may lie in an {\it early dark sector} (EDS), wherein the dark matter mass is dependent on the EDE scalar field. Concretely, we consider a Planck-suppressed coupling of EDE to dark matter, as would naturally arise from breaking of the global $U(1)$ shift symmetry of the former by quantum gravity effects. With a sufficiently flat potential, the rise to dominance of dark matter at matter-radiation equality itself triggers the rolling and subsequent decay of the EDE. We show that this {\it trigger} EDS (tEDS) model can naturally resolve the EDE coincidence problem at the background level without any fine tuning of the coupling to dark matter or of the initial conditions. When fitting to current cosmological data, including that from the local distance ladder and the low-redshift amplitude of fluctuations, the tEDS maximum-likelihood model performs comparably to EDE for resolving the Hubble tension, achieving $H_0 =71.2$ km/s/Mpc. However, fitting the \emph{Planck} cosmic microwave background data requires a specific range of initial field positions to balance the scalar field fluctuations that drive acoustic oscillations, providing testable differences with other EDE models., Comment: 12 pages, 6 figures

Decision trees are one of the most useful and popular methods in the machine learning toolbox. In this paper, we consider the problem of learning optimal decision trees, a combinatorial optimization problem that is challenging to solve at scale. A common approach in the literature is to use greedy heuristics, which may not be optimal. Recently there has been significant interest in learning optimal decision trees using various approaches (e.g., based on integer programming, dynamic programming) -- to achieve computational scalability, most of these approaches focus on classification tasks with binary features. In this paper, we present a new discrete optimization method based on branch-and-bound (BnB) to obtain optimal decision trees. Different from existing customized approaches, we consider both regression and classification tasks with continuous features. The basic idea underlying our approach is to split the search space based on the quantiles of the feature distribution -- leading to upper and lower bounds for the underlying optimization problem along the BnB iterations. Our proposed algorithm Quant-BnB shows significant speedups compared to existing approaches for shallow optimal trees on various real datasets.

Low-energy alternatives to General Relativity (GR) generically modify the phase of gravitational waves (GWs) during their propagation. As detector sensitivities increase, it becomes key to understand how these modifications affect the GW higher modes and to disentangle possible degeneracies with astrophysical phenomena. We apply a general formalism -- the WKB approach -- for solving analytically wave propagation in the spatial domain with a modified dispersion relation (MDR). We compare this WKB approach to applying a stationary phase approximation (SPA) in the temporal domain with time delays associated to the group or particle velocity. To this end, we extend the SPA to generic signals with higher modes, keeping careful track of reference phases and arrival times. We find that the WKB approach coincides with the SPA using the group velocity, in agreement with the principles of wave propagation. We then explore the degeneracies between a GW propagation with an MDR and a strongly-lensed GW in GR, since the latter can introduce a frequency-independent phase shift which is not degenerate with source parameters in the presence of higher modes. We find that for a particular MDR there is an exact degeneracy for wave propagation, unlike with the SPA for particle propagation. For the other cases, we search for the values of the MDR parameters that minimize the $\chi^2$ and conclude that strongly-lensed GR GWs could be misinterpreted as GWs in modified gravity. Future MDR constraints with higher mode GWs should include the possibility of frequency-independent phase shifts, allowing for the identification of modified gravity and strong lensing distortions at the same time., Comment: 32 pages, 6 figures. Match the published version

We consider the interplay of the Early Dark Energy (EDE) model, the Swampland Distance Conjecture (SDC), and cosmological parameter tensions. EDE is a proposed resolution of the Hubble tension relying upon a near-Planckian scalar field excursion, while the SDC predicts an exponential sensitivity of masses of other fields to such an excursion, $m\propto e^{-c|\Delta \phi|/M_{\rm pl}}$ with $c\sim{\cal O}(1)$. Meanwhile, EDE is in tension with large-scale structure (LSS) data, due to shifts in the standard $\Lambda$CDM parameters necessary to fit the cosmic microwave background (CMB). One might hope that a proper treatment of the model, e.g., accounting for the SDC, may ameliorate the tension with LSS. Motivated by these considerations, we introduce the Early Dark Sector (EDS) model, wherein the mass of dark matter is exponentially sensitive to super-Planckian field excursions of the EDE scalar. The EDS model exhibits new phenomenology in both the early and late universe, the latter due to an EDE-mediated dark matter self-interaction. This dark matter-philic "fifth force", while constrained to be small, remains active in the late universe and is not screened in virialized halos. We find that the new interaction with dark matter partially resolves the LSS tension. However, the marginalized posteriors are nonetheless consistent with $f_{\rm EDE}=0$ at 95$\%$ CL once the Dark Energy Survey Year 3 measurement of $S_8$ is included. We study constraints on the model from Atacama Cosmology Telescope data, and find a factor of two improvement on the error bar on the SDC parameter $c$, along with an increased preference for the EDE component. We discuss the implications of these constraints for the SDC, and find the tightest observational constraints to date on a swampland parameter, suggesting that an EDE description of cosmological data is in tension with the SDC., Comment: v2: version published in Phys. Rev. D