arXiv subjects

Yang Bai

Publications and source records attributed to Yang Bai.

At least 199 records · Page 11Linked to original sources

arXiv2021

Cosmological Constraints on First-Order Phase Transitions

First-order phase transitions exist in many models beyond the Standard Model and can generate detectable stochastic gravitational waves for a strong one. Using the cosmological observables in big bang nucleosynthesis and cosmic microwave background, we derive constraints on the phase transition temperature and strength parameter in a model-independent way. For a strong phase transition, we find that the phase transition temperature should be above around 2 MeV for both reheating photon and neutrino cases. For a weak one with the temperature below 1 MeV, the phase transition strength parameter is constrained to be smaller than around 0.1. Implications for using a first-order phase transition to explain the NANOGrav observed gravitational wave signal are also discussed.

hep-ph↗

arXiv2021

Diluted Axion Star Collisions with Neutron Stars

Diluted axion star, a self-gravitating object with the quantum pressure balancing gravity, has been predicted in many models with a QCD axion or axion-like particle. It can be formed in the early universe and composes a sizable fraction of dark matter. One could detect the transient radio signals when it passes by a magnetar with the axion particle converted into photon in the magnetic field. Using both numerical and semi-analytic approaches, we simulate the axion star's dynamic evolution and estimate the fraction of axion particles that can have a resonance conversion during such a collision event. We have found that both self-gravity and quantum pressure are not important after the diluted axion star enters the Roche radius. A free-fall approximate can capture individual particle trajectories very well. With some optimistic cosmological and astrophysical assumptions, the QCD axion parameter space can be probed from detecting such a collision event by radio telescopes.

astro-ph.CO↗

arXiv2021

Distributed Soft Bayesian Additive Regression Trees

Bayesian Additive Regression Trees(BART) is a Bayesian nonparametric approach which has been shown to be competitive with the best modern predictive methods such as random forest and Gradient Boosting Decision Tree.The sum of trees structure combined with a Bayesian inferential framework provide a accurate and robust statistic method.BART variant named SBART using randomized decision trees has been developed and show practical benefits compared to BART. The primary bottleneck of SBART is the speed to compute the sufficient statistics and the publicly avaiable implementation of the SBART algorithm in the R package is very slow.In this paper we show how the SBART algorithm can be modified and computed using single program,multiple data(SPMD) distributed computation with the Message Passing Interface(MPI) library.This approach scales nearly linearly in the number of processor cores, enabling the practitioner to perform statistical inference on massive datasets. Our approach can also handle datasets too massive to fit on any single data repository.We have made modification to this algorithm to make it capable to handle classfication problem which can not be done with the original R package.With data experiments we show the advantage of distributed SBART for classfication problem compared to BART.

stat.AP↗

arXiv2021

On Soft Bayesian Additive Regression Trees and asynchronous longitudinal regression analysis

In many longitudinal studies, the covariate and response are often intermittently observed at irregular, mismatched and subject-specific times. How to deal with such data when covariate and response are observed asynchronously is an often raised problem. Bayesian Additive Regression Trees(BART) is a Bayesian non-Parametric approach which has been shown to be competitive with the best modern predictive methods such as random forest and boosted decision trees. The sum of trees structure combined with a Bayesian inferential framework provide a accurate and robust statistic method. BART variant soft Bayesian Additive Regression Trees(SBART) constructed using randomized decision trees was developed and substantial theoretical and practical benefits were shown. In this paper, we propose a weighted SBART model solution for asynchronous longitudinal data. In comparison to other methods, the current methods are valid under with little assumptions on the covariate process. Extensive simulation studies provide numerical support for this solution. And data from an HIV study is used to illustrate our methodology

stat.ME↗

arXiv2021

Partially Fixed Bayes Additive Regression Trees for spatial-temporal related model

Bayes additive regression trees(BART) is a nonparametric regression model which has gained wide -spread popularity in recent years due to its flexibility and high accuracy of estimation .In spatio-temporal related model,the spatio or temporal variables are playing an important role in the model.The BART models select variables with uniform prior distribution that means treat every variable equally.Applying the BART model directly without properly using these prior information is not appropriate.This paper is aimed at a modification to the BART by fixing part of the tree's structure.We call this model partially fixed BART.By this new model we can improve efficiency of estimation.When we don't know the prior information,we can still use the new model to get more accurate estimation and more structure information for future use.Data experiments and real data examples show the improvement comparing to the original Bart model.

stat.CO↗

arXiv2021

Discriminative Latent Semantic Graph for Video Captioning

Video captioning aims to automatically generate natural language sentences that can describe the visual contents of a given video. Existing generative models like encoder-decoder frameworks cannot explicitly explore the object-level interactions and frame-level information from complex spatio-temporal data to generate semantic-rich captions. Our main contribution is to identify three key problems in a joint framework for future video summarization tasks. 1) Enhanced Object Proposal: we propose a novel Conditional Graph that can fuse spatio-temporal information into latent object proposal. 2) Visual Knowledge: Latent Proposal Aggregation is proposed to dynamically extract visual words with higher semantic levels. 3) Sentence Validation: A novel Discriminative Language Validator is proposed to verify generated captions so that key semantic concepts can be effectively preserved. Our experiments on two public datasets (MVSD and MSR-VTT) manifest significant improvements over state-of-the-art approaches on all metrics, especially for BLEU-4 and CIDEr. Our code is available at https://github.com/baiyang4/D-LSG-Video-Caption.

cs.CV↗

arXiv2021

Towards Automated Fatigue Assessment using Wearable Sensing and Mixed-Effects Models

Fatigue is a broad, multifactorial concept that includes the subjective perception of reduced physical and mental energy levels. It is also one of the key factors that strongly affect patients' health-related quality of life. To date, most fatigue assessment methods were based on self-reporting, which may suffer from many factors such as recall bias. To address this issue, in this work, we recorded multi-modal physiological data (including ECG, accelerometer, skin temperature and respiratory rate, as well as demographic information such as age, BMI) in free-living environments and developed automated fatigue assessment models. Specifically, we extracted features from each modality and employed the random forest-based mixed-effects models, which can take advantage of the demographic information for improved performance. We conducted experiments on our collected dataset, and very promising preliminary results were achieved. Our results suggested ECG played an important role in the fatigue assessment tasks.

cs.HC↗

arXiv2021

Master equation approach to the stochastic accumulation dynamics of bacterial cell cycle

The mechanism of bacterial cell size control has been a mystery for decades, which involves the well-coordinated growth and division in the cell cycle. The revolutionary modern techniques of microfluidics and the advanced live imaging analysis techniques allow long term observations and high-throughput analysis of bacterial growth on single cell level, promoting a new wave of quantitative investigations on this puzzle. Taking the opportunity, this theoretical study aims to clarify the stochastic nature of bacterial cell size control under the assumption of the accumulation mechanism, which is favoured by recent experiments on species of bacteria. Via the master equation approach with properly chosen boundary conditions, the distributions concerned in cell size control are estimated and are confirmed by experiments. In this analysis, the inter-generation Green's function is analytically evaluated as the key to bridge two kinds of statistics used in batch-culture and mother machine experiments. This framework allows us to quantify the noise level in growth and accumulation according to experimental data. As a consequence of non-Gaussian noises of the added sizes, the non-equilibrium nature of bacterial cell size homeostasis is predicted, of which the biological meaning requires further investigation.

cond-mat.stat-mech↗

arXiv2021

Catalyzed Baryogenesis

A novel mechanism, "catalyzed baryogenesis," is proposed to explain the observed baryon asymmetry in our universe. In this mechanism, the motion of a ball-like catalyst provides the necessary out-of-equilibrium condition, its outer wall has CP-violating interactions with the Standard Model particles, and its interior has baryon number violating interactions. We use the electroweak-symmetric ball model as an example of such a catalyst. In this model, electroweak sphalerons inside the ball are active and convert baryons into leptons. The observed baryon number asymmetry can be produced for a light ball mass and a large ball radius. Due to direct detection constraints on relic balls, we consider a scenario in which the balls evaporate, leading to dark radiation at testable levels.

hep-ph↗

arXiv2021

Muon g-2 in Lepton Portal Dark Matter

The Lepton Portal Dark Matter model, in which dark matter states only coupling to the charged leptons, can explain the excess of the muon anomalous magnetic moment measured by the Muon $g - 2$ experiment. In this paper, we demonstrate that real, charge-neutral scalar dark matter with a large number of internal degrees of freedom and a mass approximately degenerate with the charged fermionic mediator state can accommodate the $(g - 2)_μ$ excess. The model remains consistent with the dark matter relic abundance, direct detection, and indirect detection constraints. The dark matter and its charged fermion partner masses are constrained to be below around 200 GeV. The high-luminosity LHC and future lepton colliders, as well as indirect searches at CTA and GAMMA-400, can test this scenario.

hep-ph↗

arXiv2021

Dynamical Inflection Point Inflation

We provide a mechanism to obtain multiple inflection points for the inflaton potential based on the Coleman-Weinberg potential. The key ingredient is the existence of zeros of the inflaton quartic coupling beta function, which can be simply realized by a sign change of a gauge coupling beta function at a massive threshold scale in gauge-Yukawa models. A universal inflaton potential emerges for a wide range of models and can accommodate the experimental data in the small-field inflation scenario. While the ratio of tensor over scalar amplitudes is predicted to be small, the running of the scalar spectral index, $α_s$, has branches with both positive and negative values. The positive branch with $α_s \gtrsim 0.002$ will be tested by the coming cosmic microwave background experiments.

hep-ph↗

arXiv2021

Searching for Magnetic Monopoles with the Earth's Magnetic Field

Magnetic monopoles have long been predicted in theory and could exist as a stable object in our universe. As they move around in galaxies, magnetic monopoles could be captured by astrophysical objects like stars and planets. Here, we provide a novel method to search for magnetic monopoles by detecting the monopole moment of the Earth's magnetic field. Using over six years of public geomagnetic field data obtained by the Swarm satellites, we apply Gauss's law to measure the total magnetic flux, which is proportional to the total magnetic charge inside the Earth. To account for the secular variation of satellite altitudes, we define an altitude-rescaled magnetic flux to reduce the dominant magnetic dipole contribution. The measured magnetic flux is consistent with the existing magnetic field model that does not contain a monopole moment term. We therefore set an upper limit on the magnetic field strength at Earth's surface from magnetic monopoles to be $|B_{\rm m}| < 0.13$ nT at 95% confidence level, which is less than $2\times 10^{-6}$ of Earth's magnetic field strength. This constrains the abundance of magnetically-charged objects, including magnetic black holes with large magnetic charges.

hep-ph↗

arXiv2021

Phenomenology of Magnetic Black Holes with Electroweak-Symmetric Coronas

Magnetically charged black holes (MBHs) are interesting solutions of the Standard Model and general relativity. They may possess a "hairy" electroweak-symmetric corona outside the event horizon, which speeds up their Hawking radiation and leads them to become nearly extremal on short timescales. Their masses could range from the Planck scale up to the Earth mass. We study various methods to search for primordially produced MBHs and estimate the upper limits on their abundance. We revisit the Parker bound on magnetic monopoles and show that it can be extended by several orders of magnitude using the large-scale coherent magnetic fields in Andromeda. This sets a mass-independent constraint that MBHs have an abundance less than $6 \times 10^{-3}$ times that of dark matter. MBHs can also be captured in astrophysical systems like the Sun, the Earth, or neutron stars. There, they can become non-extremal either from merging with an oppositely charged MBH or absorbing nucleons. The resulting Hawking radiation can be detected as neutrinos, photons, or heat. High-energy neutrino searches in particular can set a stronger bound than the Parker bound for some MBH masses, down to an abundance $10^{-7}$ of dark matter.

hep-ph↗

arXiv2021

Impact of chemical structure on the dynamics of mass transfer of water in conjugated microporous polymers: A neutron spectroscopy study

Hydrogen fuel can contribute as a masterpiece in conceiving a robust carbon-free economic puzzle if cleaner methods to produce hydrogen become technically efficient and economically viable. Organic photocatalytic materials such as conjugated microporous materials (CMPs) are potential attractive candidates for water splitting as their energy levels and optical bandgap as well as porosity are tunable through chemical synthesis. The performances of CMPs depend also on the mass transfer of reactants, intermediates and products. Here, we study the mass transfer of water (H2O and D2O), and of triethylamine used as a hole scavenger for hydrogen evolution, by means of neutron spectroscopy. We find that the stiffness of the nodes of the CMPs is correlated with an increase in trapped water, reflected by motions too slow to be quantified by quasi-elastic neutron scattering (QENS). Our study highlights that the addition of the polar sulfone group results in additional interactions between water and the CMP, as evidenced by inelastic neutron scattering (INS), leading to changes in the translational diffusion of water, as determined from the QENS measurements. No changes in triethylamine motions could be observed within CMPs from the present investigations.

cond-mat.mtrl-sci↗

arXiv2020

Hairy Magnetic and Dyonic Black Holes in the Standard Model

Spherically symmetric magnetic and dyonic black holes with a magnetic charge $Q=2$ are studied in the Standard Model and general relativity. A magnetically charged black hole with mass below $9.3\times 10^{35}$ GeV has a "hairy" cloud of electroweak gauge and Higgs fields outside the event horizon with $1/m_W$ in size. An extremal magnetic black hole has a hair mass of 3.6 TeV, while an extremal dyonic black hole has an additional mass of $q^2 \times 1.6$ GeV for a small electric charge $q \ll 2π/e^2$. A hairy dyonic black hole with an integer charge is not stable and can decay into a magnetic one plus charged fermions. On the other hand, a hairy magnetic black hole can evolve via Hawking radiation into a nearly extremal one that is cosmologically stable and an interesting object to be searched for.

hep-ph↗

arXiv2020

SparTerm: Learning Term-based Sparse Representation for Fast Text Retrieval

Term-based sparse representations dominate the first-stage text retrieval in industrial applications, due to its advantage in efficiency, interpretability, and exact term matching. In this paper, we study the problem of transferring the deep knowledge of the pre-trained language model (PLM) to Term-based Sparse representations, aiming to improve the representation capacity of bag-of-words(BoW) method for semantic-level matching, while still keeping its advantages. Specifically, we propose a novel framework SparTerm to directly learn sparse text representations in the full vocabulary space. The proposed SparTerm comprises an importance predictor to predict the importance for each term in the vocabulary, and a gating controller to control the term activation. These two modules cooperatively ensure the sparsity and flexibility of the final text representation, which unifies the term-weighting and expansion in the same framework. Evaluated on MSMARCO dataset, SparTerm significantly outperforms traditional sparse methods and achieves state of the art ranking performance among all the PLM-based sparse models.

cs.IR↗

arXiv2020

Tests of Dark MACHOs: Lensing, Accretion, and Glow

Dark matter could take the form of dark massive compact halo objects (dMACHOs); i.e., composite objects that are made up of dark-sector elementary particles, that could have a macroscopic mass from the Planck scale to above the solar mass scale, and that also admit a wide range of energy densities and sizes. Concentrating on the gravitational interaction of dMACHOs with visible matter, we map out the mass-radius parameter space that is consistent with gravitational lensing experiments, as well as anisotropies of the cosmic microwave background (CMB) based on the spherical accretion of matter onto a dMACHO in the hydrostatic approximation. For dMACHOs with a uniform-density mass profile and total mass in the range of $\sim 10^{-12} - 10\,M_\odot$, we find that a dMACHO could explain 100% of the dark matter if its radius is above $\approx 3$ times the Einstein radius of the lensing system. For a larger mass above $10\,M_\odot$, a dMACHO with radius above $\sim 1 \times 10^8 \mbox{cm} \times(M/100\,M_\odot)^{9/2}$ is consistent with CMB observables. For a lighter dMACHO with mass below $\sim 10^{-12}\,M_\odot$, there still is not a good experimental probe. Finally, we point out that heavier dMACHOs with masses $\sim 0.1\,M_\odot$ may be observed by X-ray and optical telescopes if they reside at rest in a large molecular cloud, nearby to our solar system, and accrete ordinary matter to emit photons.

astro-ph.CO↗

arXiv2020

Improving Query Efficiency of Black-box Adversarial Attack

Deep neural networks (DNNs) have demonstrated excellent performance on various tasks, however they are under the risk of adversarial examples that can be easily generated when the target model is accessible to an attacker (white-box setting). As plenty of machine learning models have been deployed via online services that only provide query outputs from inaccessible models (e.g. Google Cloud Vision API2), black-box adversarial attacks (inaccessible target model) are of critical security concerns in practice rather than white-box ones. However, existing query-based black-box adversarial attacks often require excessive model queries to maintain a high attack success rate. Therefore, in order to improve query efficiency, we explore the distribution of adversarial examples around benign inputs with the help of image structure information characterized by a Neural Process, and propose a Neural Process based black-box adversarial attack (NP-Attack) in this paper. Extensive experiments show that NP-Attack could greatly decrease the query counts under the black-box setting.

cs.LG↗