Search

Posterior sampling is a task of central importance in Bayesian inference. For many applications in Bayesian meta-analysis and Bayesian transfer learning, the prior distribution is unknown and needs to be estimated from samples. In practice, the prior distribution can be high-dimensional, adding to the difficulty of efficient posterior inference. In this paper, we propose a novel Markov chain Monte Carlo algorithm, which we term graph-enabled MCMC, for posterior sampling with unknown and potentially high-dimensional prior distributions. The algorithm is based on constructing a geometric graph from prior samples and subsequently uses the graph structure to guide the transition of the Markov chain. Through extensive theoretical and numerical studies, we demonstrate that our graph-enabled MCMC algorithm provides reliable approximation to the posterior distribution and is highly computationally efficient., Comment: 45 pages, 11 figures

Compared to traditional sentiment analysis, which only considers text, multimodal sentiment analysis needs to consider emotional signals from multimodal sources simultaneously and is therefore more consistent with the way how humans process sentiment in real-world scenarios. It involves processing emotional information from various sources such as natural language, images, videos, audio, physiological signals, etc. However, although other modalities also contain diverse emotional cues, natural language usually contains richer contextual information and therefore always occupies a crucial position in multimodal sentiment analysis. The emergence of ChatGPT has opened up immense potential for applying large language models (LLMs) to text-centric multimodal tasks. However, it is still unclear how existing LLMs can adapt better to text-centric multimodal sentiment analysis tasks. This survey aims to (1) present a comprehensive review of recent research in text-centric multimodal sentiment analysis tasks, (2) examine the potential of LLMs for text-centric multimodal sentiment analysis, outlining their approaches, advantages, and limitations, (3) summarize the application scenarios of LLM-based multimodal sentiment analysis technology, and (4) explore the challenges and potential research directions for multimodal sentiment analysis in the future., Comment: arXiv admin note: text overlap with arXiv:2210.14556 by other authors

We present photometric and spectroscopic observations and analysis of the type IIb supernova (SN) SN 2019tua, which exhibits multiple bumps in its declining light curves between 40 and 65 days after discovery. SN 2019tua shows a time to peak of about 25 days similar to other type IIb SNe. Our observations indicate a decrease in its brightness of about 1 magnitude in the 60 days after the peak. At about days 50, and 60, its multiband light curves exhibit bumpy behavior. The complex luminosity evolution of SN 2019tua could not be well modeled with a single currently popular energy source model, e.g., radioactive decay of $^{56}$Ni, magnetar, interaction between the ejecta and a circumstellar shell. Even though the magnetar model has a smaller $\chi^2 / \text{dof}$ value, the complex changes in SN 2019tua's brightness suggest that more than one physical process might be involved. We propose a hybrid CSM interaction plus $^{56}$Ni model to explain the bolometric light curve (LC) of SN 2019tua. The fitting results show that the ejecta mass $M_{\rm ej} \approx 2.4~M_\odot$, the total CSM mass $M_{\rm CSM} \approx 1.0~M_\odot$, and the $^{56}$Ni mass $M_{\rm Ni} \approx 0.4~M_\odot$. The total kinetic energy of the ejecta is $E_k\approx 0.5 \times 10^{51}\rm~erg$. Pre-existing multiple shells suggest that the progenitor of SN 2019tua experienced mass ejections within approximately $\sim6 - 44$ years prior to the explosion., Comment: Accepted for publication in ApJ. 24 pages, 17 figures, 6 tables

A stellar-mass black hole (BH) or a millisecond magnetar is believed to be born as the central engine of Gamma-ray bursts (GRBs). The presence of plateaus in the X-ray extended emission or afterglow of GRBs is widely accepted as an indicator of magnetar central engine, particularly those with a sharp decay (faster than $t^{-3}$), so-called internal plateau. However, an alternative model, by taking the evolution of the magnetic flux at the BH horizon into account, suggests that an internal plateau can also arise from a Blandford-Znajek (BZ) mechanism powered jet (hereafter referred to as the BZ jet). In this study, we propose that a precessional BZ jet would manifest a Quasi-Periodic Oscillation (QPO) signature on the internal plateau and the subsequent sharp decay. Such lightcurves cannot be readily explained by the activity of a short-lived, supermassive magnetar, thus favoring a Kerr BH as the central engine. The X-ray afterglow of GRB 050904, comprising nine flares, is characterized by a QPO-modulated plateau and sharp decay, which can be well reproduced by a precessional BZ jet model. Therefore, one potential clue for distinguishing between these two engines lies in whether QPO signature throughout the entire plateau and the subsequent sharp decay, as the magnetar scenario suggests a collapse at the end of the plateau., Comment: Accepted for publication in ApJ

In this paper, we evaluate different abilities of GPT-4V including visual understanding, language understanding, visual puzzle solving, and understanding of other modalities such as depth, thermal, video, and audio. To estimate GPT-4V's performance, we manually construct 656 test instances and carefully evaluate the results of GPT-4V. The highlights of our findings are as follows: (1) GPT-4V exhibits impressive performance on English visual-centric benchmarks but fails to recognize simple Chinese texts in the images; (2) GPT-4V shows inconsistent refusal behavior when answering questions related to sensitive traits such as gender, race, and age; (3) GPT-4V obtains worse results than GPT-4 (API) on language understanding tasks including general language understanding benchmarks and visual commonsense knowledge evaluation benchmarks; (4) Few-shot prompting can improve GPT-4V's performance on both visual understanding and language understanding; (5) GPT-4V struggles to find the nuances between two similar images and solve the easy math picture puzzles; (6) GPT-4V shows non-trivial performance on the tasks of similar modalities to image, such as video and thermal. Our experimental results reveal the ability and limitations of GPT-4V and we hope our paper can provide some insights into the application and research of GPT-4V., Comment: Technical Report. Data are available at https://github.com/albertwy/GPT-4V-Evaluation

We construct four infinite families of chiral $3$-polytopes of type $\{4, 8\}$, with $1024m^4$, $2048m^4$, $4096m^4$ and $8192m^4$ automorphisms for every positive integer $m$, respectively. The automorphism groups of these polytopes are solvable groups, and when $m$ is a power of $2$, they provide examples with automorphism groups of order $2^n$ where $n \geq 10$. (On the other hand, no chiral polytopes of type $\{4, 8\}$ exist for $n \leq 9$.) In particular, our families give a partial answer to a problem proposed by Schulte and Weiss in [Problems on polytopes, their groups, and realizations, {\em Period. Math. Hungar.} 53 (2006), 231-255] and a problem proposed by Pellicer in [Developments and open problems on chiral polytopes, {\em Ars Math. Contemp} 5 (2012), 333-354]., Comment: 11pges,1 figures. arXiv admin note: substantial text overlap with arXiv:1912.03398

Modern climate projections lack adequate spatial and temporal resolution due to computational constraints, leading to inaccuracies in representing critical processes like thunderstorms that occur on the sub-resolution scale. Hybrid methods combining physics with machine learning (ML) offer faster, higher fidelity climate simulations by outsourcing compute-hungry, high-resolution simulations to ML emulators. However, these hybrid ML-physics simulations require domain-specific data and workflows that have been inaccessible to many ML experts. As an extension of the ClimSim dataset (Yu et al., 2024), we present ClimSim-Online, which also includes an end-to-end workflow for developing hybrid ML-physics simulators. The ClimSim dataset includes 5.7 billion pairs of multivariate input/output vectors, capturing the influence of high-resolution, high-fidelity physics on a host climate simulator's macro-scale state. The dataset is global and spans ten years at a high sampling frequency. We provide a cross-platform, containerized pipeline to integrate ML models into operational climate simulators for hybrid testing. We also implement various ML baselines, alongside a hybrid baseline simulator, to highlight the ML challenges of building stable, skillful emulators. The data (https://huggingface.co/datasets/LEAP/ClimSim_high-res) and code (https://leap-stc.github.io/ClimSim and https://github.com/leap-stc/climsim-online) are publicly released to support the development of hybrid ML-physics and high-fidelity climate simulations., Comment: This manuscript is an expanded version of our paper that received the Outstanding Paper Award at the NeurIPS 2023 conference

GRB 230307A is one of the brightest gamma-ray bursts detected so far. With the excellent observation of GRB 230307A by Fermi-GBM, we can reveal the details of prompt emission evolution. As found in high-time-resolution spectral analysis, the early low-energy spectral indices ($\alpha$) of this burst exceed the limit of synchrotron radiation ($\alpha=-2/3$), and gradually decreases with the energy flux ($F$). A tight $E_{\rm p}\propto F^{0.54}$ correlation anyhow holds within the whole duration of the burst, where $E_{\rm p}$ is the spectral peak energy. Such evolution pattern of $\alpha$ and $E_{\rm p}$ with intensity is called ``double tracking". For the $\alpha-F$ relation, we find a log Bayes factor $\sim$ 210 in favor of a smoothly broken power-law function over a linear function in log-linear space. We call this particular $\alpha-F$ relation as broken ``$\alpha$-intensity", and interpret it as the evolution of the ratio of thermal and non-thermal components, which is also the evolution of the photosphere. GRB 230307A with a duration of $\sim 35~\rm s$, if indeed at a redshift of $z=0.065$, is likely a neutron star merger event (i.e., it is intrinsically ``short"). Intriguingly, different from GRB 060614 and GRB 211211A, this long event is not composed of a hard spike followed by a soft tail, suggesting that the properties of the prompt emission light curves are not a good tracer of the astrophysical origins of the bursts. The other possibility of $z=3.87$ would point toward very peculiar nature of both GRB 230307A and its late time thermal-like emission., Comment: 14 pages, 6 figures, 1 table, accepted for publication in ApJL

Predictive modeling uncovers knowledge and insights regarding a hypothesized data generating mechanism (DGM). Results from different studies on a complex DGM, derived from different data sets, and using complicated models and algorithms, are hard to quantitatively compare due to random noise and statistical uncertainty in model results. This has been one of the main contributors to the replication crisis in the behavioral sciences. The contribution of this paper is to apply prediction scoring to the problem of comparing two studies, such as can arise when evaluating replications or competing evidence. We examine the role of predictive models in quantitatively assessing agreement between two datasets that are assumed to come from two distinct DGMs. We formalize a distance between the DGMs that is estimated using cross validation. We argue that the resulting prediction scores depend on the predictive models created by cross validation. In this sense, the prediction scores measure the distance between DGMs, along the dimension of the particular predictive model. Using human behavior data from experimental economics, we demonstrate that prediction scores can be used to evaluate preregistered hypotheses and provide insights comparing data from different populations and settings. We examine the asymptotic behavior of the prediction scores using simulated experimental data and demonstrate that leveraging competing predictive models can reveal important differences between underlying DGMs. Our proposed cross-validated prediction scores are capable of quantifying differences between unobserved data generating mechanisms and allow for the validation and assessment of results from complex models.

Learning conditional densities and identifying factors that influence the entire distribution are vital tasks in data-driven applications. Conventional approaches work mostly with summary statistics, and are hence inadequate for a comprehensive investigation. Recently, there have been developments on functional regression methods to model density curves as functional outcomes. A major challenge for developing such models lies in the inherent constraint of non-negativity and unit integral for the functional space of density outcomes. To overcome this fundamental issue, we propose Wasserstein Distributional Learning (WDL), a flexible density-on-scalar regression modeling framework that starts with the Wasserstein distance $W_2$ as a proper metric for the space of density outcomes. We then introduce a heterogeneous and flexible class of Semi-parametric Conditional Gaussian Mixture Models (SCGMM) as the model class $\mathfrak{F} \otimes \mathcal{T}$. The resulting metric space $(\mathfrak{F} \otimes \mathcal{T}, W_2)$ satisfies the required constraints and offers a dense and closed functional subspace. For fitting the proposed model, we further develop an efficient algorithm based on Majorization-Minimization optimization with boosted trees. Compared with methods in the previous literature, WDL better characterizes and uncovers the nonlinear dependence of the conditional densities, and their derived summary statistics. We demonstrate the effectiveness of the WDL framework through simulations and real-world applications.

The physical composition of the ejecta of gamma-ray bursts (GRBs) remains an open question. The radiation mechanism of the prompt gamma rays is also in debate. This problem can be solved for the bursts hosting distinct thermal radiation. However, the events with dominant thermal spectral components are still rare. In this work, we focus on GRB 220426A, a recent event detected by Fermi-GBM. The time-resolved and time-integrated data analyses yield very hard low-energy spectral indices and rather soft high-energy spectral indices. This means that the spectra of GRB 220426A are narrowly distributed. And the Bayesian inference results are in favor of the multicolor blackbody (mBB) model. The physical properties of the relativistic outflow are calculated. Assuming a redshift $z= 1.4$, the bulk Lorentz factors $\Gamma$ of the shells are found to be between $274_{-18}^{+24}$ and $827_{-71}^{+100}$, and the corresponding photosphere radii $R_{\rm ph}$ are in the range of $1.83_{-0.50}^{+0.52} \times 10^{11}$ and $2.97_{-0.15}^{+0.14} \times 10^{12}$ cm. Similar to GRB 090902B, the time-resolved properties of GRB 220426A satisfy the observed $\Gamma-L$ and $E_p-L$ correlations, where $L$ is the luminosity of the prompt emission and $E_{p}$ is the spectral peak energy., Comment: 14 pages, 6 figures, 2 tables; Accepted for publication in ApJ

Spin systems are an attractive candidate for quantum-enhanced metrology. Here we develop a variational method to generate metrological states in small dipolar-interacting ensembles with limited qubit controls and unknown spin locations. The generated states enable sensing beyond the standard quantum limit (SQL) and approaching the Heisenberg limit (HL). Depending on the circuit depth and the level of readout noise, the resulting states resemble Greenberger-Horne-Zeilinger (GHZ) states or Spin Squeezed States (SSS). Sensing beyond the SQL holds in the presence of finite spin polarization and a non-Markovian noise environment., Comment: 6 pages, 4 figures, 1 table in main text. 21 pages, 8 figures, 3 tables in supplemental material

Protein structure prediction helps to understand gene translation and protein function, which is of growing interest and importance in structural biology. The AlphaFold model, which used transformer architecture to achieve atomic-level accuracy in protein structure prediction, was a significant breakthrough. However, training and inference of the AlphaFold model are challenging due to its high computation and memory cost. In this work, we present FastFold, an efficient implementation of AlphaFold for both training and inference. We propose Dynamic Axial Parallelism and Duality Async Operations to improve the scaling efficiency of model parallelism. Besides, AutoChunk is proposed to reduce memory cost by over 80% during inference by automatically determining the chunk strategy. Experimental results show that FastFold reduces overall training time from 11 days to 67 hours and achieves 7.5X - 9.5X speedup for long-sequence inference. Furthermore, we scale FastFold to 512 GPUs and achieve an aggregate throughput of 6.02 PetaFLOP/s with 90.1% parallel efficiency.

We propose an effective approach for generating significant amount of entanglement and asymmetric steering between photon and phonon in a cavity magnomechanical system which consists of a microwave cavity and a yttrium iron garnet sphere. By driving the magnon mode of the yttrium iron garnet sphere with blue-detuned microwave field, the magnon mode can be acted as an engineered resevoir cools the Bogoliubov modes of microwave cavity mode and mechanical mode via beam-splitter-like interaction. In this way, the microwave cavity mode and mechanical mode are driven to two-mode squeezed states in the stationary limit. In particular, strong two-way and one-way asymmetric quantum steering between the photon and phonon modes can be obtained with even equal dissipation. It is very different from the conventional proposal of asymmetric quantum steering, where additional unbalanced losses or noises on the two subsystems has been imposed. Our finding may be significant to expand our understanding of the essential physics of asymmetric steering and extend the potential application of the cavity spintronics to device-independent quantum key distribution., Comment: 15 pages, 3 subfigures

Probabilistic machine learning increasingly informs critical decisions in medicine, economics, politics, and beyond. We need evidence to support that the resulting decisions are well-founded. To aid development of trust in these decisions, we develop a taxonomy delineating where trust in an analysis can break down: (1) in the translation of real-world goals to goals on a particular set of available training data, (2) in the translation of abstract goals on the training data to a concrete mathematical problem, (3) in the use of an algorithm to solve the stated mathematical problem, and (4) in the use of a particular code implementation of the chosen algorithm. We detail how trust can fail at each step and illustrate our taxonomy with two case studies: an analysis of the efficacy of microcredit and The Economist's predictions of the 2020 US presidential election. Finally, we describe a wide variety of methods that can be used to increase trust at each step of our taxonomy. The use of our taxonomy highlights steps where existing research work on trust tends to concentrate and also steps where establishing trust is particularly challenging., Comment: 18 pages, 2 figures

Assuming that the shallow-decaying phase in the early X-ray lightcurves of gamma-ray bursts (GRBs) is attributed to the dipole radiations (DRs) of a newborn magnetar, we present a comparative analysis for the magnetars born in death of massive stars and merger of compact binaries with long and short GRB (lGRB and sGRB) data observed with the {\em Swift} mission. We show that the typical braking index ($n$) of the magnetars is $\sim 3$ in the sGRB sample, and it is $\sim 4$ for the magnetars in the lGRB sample. Selecting a sub-sample of the magnetars whose spin-down is dominated by DRs ($n\lesssim 3$) and adopting a universal radiation efficiency of $0.3$, we find that the typical magnetic field strength ($B_p$) is $10^{16}$ G {\em vs.} $10^{15}$ G and the typical initial period ($P_0$) is $\sim 20$ ms {\em vs.} $2$ ms for the magnetars in the sGRBs {\em vs.} lGRBs. They follow the same relation between $P_0$ and the isotropic GRB energy as $ P_0\propto E_{\rm jet}^{-0.4}$. We also extend our comparison analysis to superluminous supernovae (SLSNe) and stable pulsars. Our results show that a magnetar born in merger of compact stars tends to have a stronger $B_p$ and a longer $P_0$ by about one order of magnitude than that born in collapse of massive stars. Its spin-down is dominated by the magnetic DRs as old pulsars, being due to its strong magnetic field strength, whereas the early spin-down of magnetars born in massive star collapse is governed by both the DRs and gravitational wave (GW) emission. A magnetar with a faster rotation speed should power a more energetic jet, being independent of its formation approach., Comment: 10 pages, 2 tables, 7 figures. Published in MNRAS

The unusual multiwavelength lightcurves of GRB 101225A are revisited by assuming that it is from an off-axis GRB powered by a newborn magnetar. We show that its optical afterglow lightcurve is fitted with the forward shock model by parameterizing its jet structure as a Gaussian function with a half opening angle of the jet core as $1.67^{\rm o}$. The derived initial Lorentz factor ($\Gamma_0$) is 120, and the viewing angle to the jet axis is $\theta_v=3.7^{\rm o}$. Tentative QPO signatures of $P=488$ seconds and $P=250\sim 300$ seconds are found with a confidence level of 90\% by analysing its X-ray flares observed in the time interval of $[4900,\ 7500]$ seconds. Its global gamma-ray/X-ray lightcurve and the QPO signatures are represented with the magnetar dipole radiation (DR) model by considering the magnetar precession motion, assuming that the magnetar spindown is dominated by the GW emission. The bulk Lorentz factor of the DR ejecta is limited to 8, being much lower than $\Gamma_0$. Comparing GRB 101225A with the extremely off-axis GRB 170817A, we suspect that the nature of the two-component jet in GRB 170817A is a combination of a co-axial GRB jet and a DR ejecta. GRB 101225A would be among the brightest ones of the CDF-S XT2 like X-ray transient population driven by newborn magnetars. Discussion on detectability of its gravitational wave emission is also presented., Comment: 8 pages,4 figures. Accepted for publication in ApJL on 10/12/2021

Probabilistic machine learning increasingly informs critical decisions in medicine, economics, politics, and beyond. To aid the development of trust in these decisions, we develop a taxonomy delineating where trust in an analysis can break down: (1) in the translation of real-world goals to goals on a particular set of training data, (2) in the translation of abstract goals on the training data to a concrete mathematical problem, (3) in the use of an algorithm to solve the stated mathematical problem, and (4) in the use of a particular code implementation of the chosen algorithm. We detail how trust can fail at each step and illustrate our taxonomy with two case studies. Finally, we describe a wide variety of methods that can be used to increase trust at each step of our taxonomy. The use of our taxonomy highlights steps where existing research work on trust tends to concentrate and also steps where building trust is particularly challenging. [This paper was published in "Science Advances."]