arXiv subjects

Wei Zhu

Publications and source records attributed to Wei Zhu.

At least 181 records · Page 10Linked to original sources

arXiv2023

Candidate Soups: Fusing Candidate Results Improves Translation Quality for Non-Autoregressive Translation

Non-autoregressive translation (NAT) model achieves a much faster inference speed than the autoregressive translation (AT) model because it can simultaneously predict all tokens during inference. However, its translation quality suffers from degradation compared to AT. And existing NAT methods only focus on improving the NAT model's performance but do not fully utilize it. In this paper, we propose a simple but effective method called "Candidate Soups," which can obtain high-quality translations while maintaining the inference speed of NAT models. Unlike previous approaches that pick the individual result and discard the remainders, Candidate Soups (CDS) can fully use the valuable information in the different candidate translations through model uncertainty. Extensive experiments on two benchmarks (WMT'14 EN-DE and WMT'16 EN-RO) demonstrate the effectiveness and generality of our proposed method, which can significantly improve the translation quality of various base models. More notably, our best variant outperforms the AT model on three translation tasks with 7.6 times speedup.

cs.CL↗

arXiv2023

Looking for the Signals of the Missing Baryons in the Extragalactic Background Light

The missing baryons in the universe are assumed to be hidden in the whole space as a warm-hot intergalactic medium (WHIM). Finding them is one of the important subjects in modern cosmology. In this paper, we point out that the very high energy electron/positron rays may light up the WHIM due to the anomalous bremsstrahlung according to an improved Bethe-Heitler formula. The resulting excess of the extragalactic background light (EBL) can be observed by the direct measurement method. A possible explanation on the difference between the direct and indirect measurements of EBL is also proposed. Thus, we find a new window to probe the WHIM properties via the EBL.

astro-ph.CO↗

arXiv2023

Systematic KMTNet Planetary Anomaly Search, Paper VII: Complete Sample of $q < 10^{-4}$ Planets from the First Four-Year Survey

We present the analysis of seven microlensing planetary events with planet/host mass ratios $q < 10^{-4}$: KMT-2017-BLG-1194, KMT-2017-BLG-0428, KMT-2019-BLG-1806, KMT-2017-BLG-1003, KMT-2019-BLG-1367, OGLE-2017-BLG-1806, and KMT-2016-BLG-1105. They were identified by applying the Korea Microlensing Telescope Network (KMTNet) AnomalyFinder algorithm to 2016--2019 KMTNet events. A Bayesian analysis indicates that all the lens systems consist of a cold super-Earth orbiting an M or K dwarf. Together with 17 previously published and three that will be published elsewhere, AnomalyFinder has found a total of 27 planets that have solutions with $q < 10^{-4}$ from 2016--2019 KMTNet events, which lays the foundation for the first statistical analysis of the planetary mass-ratio function based on KMTNet data. By reviewing the 27 planets, we find that the missing planetary caustics problem in the KMTNet planetary sample has been solved by AnomalyFinder. We also find a desert of high-magnification planetary signals ($A \gtrsim 65$), and a follow-up project for KMTNet high-magnification events could detect at least two more $q < 10^{-4}$ planets per year and form an independent statistical sample.

astro-ph.EP↗

arXiv2022

One Dimensional Ferroelectric Nanothreads with Axial and Radial Polarization

Long-range ferroelectric crystalline order usually fades away as the spatial dimension decreases, hence there are few two-dimensional (2D) ferroelectrics and far fewer one-dimensional (1D) ferroelectrics. Due to the depolarization field, low-dimensional ferroelectrics rarely possess the polarization along the direction of reduced dimensionality. Here, using first-principles density functional theory, we explore the structural evolution of nanoribbons of varying widths constructed by cutting the 2D sheet of ferroelectric $α$-III$_2$IV$_3$ (III=Al, Ga, In; VI=S, Se, Te). We discover a one-dimensional ferroelectric nanothread (1DFENT) of ultrasmall diameter with both axial and radial polarization, potentially enabling ultra-dense data storage with a 1D domain of just three unit cells being the functional unit. The polarization in 1DFENT of Ga$_2$Se$_3$ exhibits an unusual piezoelectric response: a stretching stress along the axial direction will increase both axial and radial polarization, referred to as auxetic piezoelectric effect. Utilizing the intrinsically flat electronic bands, we demonstrate the coexistence of ferroelectricity and ferromagnetism in 1DFENT and a counterintuitive charge-doping-induced metal-to-insulator transition. The 1DFENT with both axial and radial polarization offers a counterexample to the Mermin--Wagner theorem in 1D and suggests a new platform for the design of ultrahigh-density memory and the exploration of exotic states of matter.

cond-mat.mtrl-sci↗

arXiv2022

Discovering Governing Equations in Discrete Systems Using PINNs

Sparse identification of nonlinear dynamical systems is a topic of continuously increasing significance in the dynamical systems community. Here we explore it at the level of lattice nonlinear dynamical systems of many degrees of freedom. We illustrate the ability of a suitable adaptation of Physics-Informed Neural Networks (PINNs) to solve the inverse problem of parameter identification in such discrete, high-dimensional systems inspired by physical applications. The methodology is illustrated in a diverse array of examples including real-field ones ($ϕ^4$ and sine-Gordon), as well as complex-field (discrete nonlinear Schr{ö}dinger equation) and going beyond Hamiltonian to dissipative cases (the discrete complex Ginzburg-Landau equation). Both the successes, as well as some limitations of the method are discussed along the way.

nlin.PS↗

arXiv2022

Occurrence rate of hot Jupiters around early-type M dwarfs based on TESS data

We present an estimate of the occurrence rate of hot Jupiters ($7\ R_{\oplus}\leq R_{p}\leq 2\ R_{J}$, $0.8 \leq P_{b}\leq 10$ days) around early-type M dwarfs based on stars observed by TESS during its Primary Mission. We adopt stellar parameters from the TESS Input Catalog, and construct a sample of 60,819 M dwarfs with $10.5 \leq T_{\rm mag}\leq 13.5$, effective temperature $2900 \leq T_{\rm eff}\leq 4000\ K$ and stellar mass $0.45\leq M_{\ast}\leq 0.65\ M_{\odot}$. We conduct a uninformed transit search using a detection pipeline based on the box least square search and characterize the searching completeness through an injection and recovery experiment. We combine a series of vetting steps including light centroid measurement, odd/even and secondary eclipse analysis, rotation and transit period synchronization tests as well as inspecting the ground-based photometric, spectroscopic and imaging observations. Finally, we find a total of nine planet candidates, all of which are known TESS objects of interest. We obtain an occurrence rate of $0.27\pm0.09\%$ for hot Jupiters around early-type M dwarfs that satisfy our selection criteria. Compared with previous studies, the occurrence rate of hot Jupiters around early-type M dwarfs is smaller than all measurements for FGK stars, although they are consistent within 1--2$σ$. Combining results from transit, radial velocity and microlensing surveys, we find that hot Jupiters around early-type M dwarfs possibly show a steeper decrease in occurrence rate per logarithmic semi-major axis bin (${{\rm d}N}/{\rm d}\log_{10} a$) when compared with FGK stars.

astro-ph.EP↗

arXiv2022

MAGIC: Microlensing Analysis Guided by Intelligent Computation

The modeling of binary microlensing light curves via the standard sampling-based method can be challenging, because of the time-consuming light-curve computation and the pathological likelihood landscape in the high-dimensional parameter space. In this work, we present MAGIC, which is a machine-learning framework to efficiently and accurately infer the microlensing parameters of binary events with realistic data quality. In MAGIC, binary microlensing parameters are divided into two groups and inferred separately with different neural networks. The key feature of MAGIC is the introduction of a neural controlled differential equation, which provides the capability to handle light curves with irregular sampling and large data gaps. Based on simulated light curves, we show that MAGIC can achieve fractional uncertainties of a few percent on the binary mass ratio and separation. We also test MAGIC on a real microlensing event. MAGIC is able to locate degenerate solutions even when large data gaps are introduced. As irregular samplings are common in astronomical surveys, our method also has implications for other studies that involve time series.

astro-ph.IM↗

arXiv2022

Continuous Detection, Rapidly React: Unseen Rumors Detection based on Continual Prompt-Tuning

Since open social platforms allow for a large and continuous flow of unverified information, rumors can emerge unexpectedly and spread quickly. However, existing rumor detection (RD) models often assume the same training and testing distributions and can not cope with the continuously changing social network environment. This paper proposed a Continual Prompt-Tuning RD (CPT-RD) framework, which avoids catastrophic forgetting (CF) of upstream tasks during sequential task learning and enables bidirectional knowledge transfer between domain tasks. Specifically, we propose the following strategies: (a) Our design explicitly decouples shared and domain-specific knowledge, thus reducing the interference among different domains during optimization; (b) Several technologies aim to transfer knowledge of upstream tasks to deal with emergencies; (c) A task-conditioned prompt-wise hypernetwork (TPHNet) is used to consolidate past domains. In addition, CPT-RD avoids CF without the necessity of a rehearsal buffer.

cs.CL↗

arXiv2022

OGLE-2019-BLG-1470LABc: Another Microlensing Giant Planet in a Binary System?

We report the discovery and analysis of a candidate triple-lens single-source (3L1S) microlensing event, OGLE-2019-BLG-1470. This event was first classified as a normal binary-lens single-source (2L1S) event, but a careful 2L1S modelling showed that it needs an additional lens or source to fit the observed data. It is found that the 3L1S model provides the best fit, but the binary-lens binary-source (2L2S) model is only disfavoured by $Δχ^2 \simeq 18$. All of the feasible models include a planet with planet-to-host mass-ratios $10^{-3} \lesssim q \lesssim 10^{-2}$. A Bayesian analysis based on a Galactic model indicates that the planet is super-Jovian, and the projected host-planet separation is about 3 $\mathrm{au}$. Specifically, for the best-fit 3L1S model, the two stars have masses of $M_1=0.57^{+0.43}_{-0.32}M_{\odot}$, and $M_2=0.18^{+0.15}_{-0.10}M_{\odot}$, with projected separation of $1.3^{+0.5}_{-0.5}$ $\mathrm{au}$, and the planetary mass is $M_3=2.2^{+1.8}_{-1.3}M_{\rm{Jupiter}}$. For the 2L2S model, the masses of the host star and the planet are $0.55^{+0.44}_{-0.31}M_{\odot}$ and $4.6^{+3.7}_{-2.6}M_{\rm{Jupiter}}$, respectively. By investigating the properties of all known microlensing planets in binary systems, we find that all planets in binary systems published by the KMTNet survey are located inside the resonant caustics range with $q \gtrsim 2 \times 10^{-3}$, indicating the incompleteness of the KMTNet sample for planets in binary systems. Thus, planets in binary systems cannot be included in the current study of the KMTNet mass-ratio function, and a systematic search for planetary anomalies in KMTNet microlensing light curves of binary systems is needed.

astro-ph.EP↗

arXiv2022

Warning: The mini gamma-ray-bursts in planning hadron colliders beyond the LHC energies

Gluons may converge to a stable state at a critical momentum in nucleon. This gluon condensation will greatly increase the proton-proton cross section provided that the collision energies exceed the gluon condensation threshold. Based on the analyses of cosmic gamma-ray spectra, we find that the $p-Pb$ and $Pb-Pb$ collisions at the LHC are close to the energy region of the gluon condensation effect. We warn that for the next generation of hadron colliders increasing the collision energies, the extremely strong gamma-rays will be emitted in a narrow space of the accelerator due to the gluon condensation effect. Such artificial mini gamma-ray-bursts in the laboratory may damage the detectors.

hep-ph↗

arXiv2022

The gluon condensation in hadron collisions

Gluons may converge to a stable state at a critical momentum in hadrons. This gluon condensation is predicted by a nonlinear QCD evolution equation. We review the understanding of the gluon condensation and present a clear physical picture that produces the gluon condensation from the colour glass condensate. We summarize the applications of the GC effect in the $p-p(A)$ collisions and predict that the $p-Pb$ and $Pb-Pb$ collisions at the LHC are close to the energy region of the gluon condensation. We warn that for the next generation of hadron colliders with the increasing of the collision energy, the extremely strong gamma-rays will be emitted in a narrow space of the accelerator due to the gluon condensation effect. Such artificial mini gamma-ray bursts in the laboratory may damage the detectors.

hep-ph↗

arXiv2022

Improving Micro-video Recommendation by Controlling Position Bias

As the micro-video apps become popular, the numbers of micro-videos and users increase rapidly, which highlights the importance of micro-video recommendation. Although the micro-video recommendation can be naturally treated as the sequential recommendation, the previous sequential recommendation models do not fully consider the characteristics of micro-video apps, and in their inductive biases, the role of positions is not in accord with the reality in the micro-video scenario. Therefore, in the paper, we present a model named PDMRec (Position Decoupled Micro-video Recommendation). PDMRec applies separate self-attention modules to model micro-video information and the positional information and then aggregate them together, avoid the noisy correlations between micro-video semantics and positional information being encoded into the sequence embeddings. Moreover, PDMRec proposes contrastive learning strategies which closely match with the characteristics of the micro-video scenario, thus reducing the interference from micro-video positions in sequences. We conduct the extensive experiments on two real-world datasets. The experimental results shows that PDMRec outperforms existing multiple state-of-the-art models and achieves significant performance improvements.

cs.IR↗

arXiv2022

Explored An Effective Methodology for Fine-Grained Snake Recognition

Fine-Grained Visual Classification (FGVC) is a longstanding and fundamental problem in computer vision and pattern recognition, and underpins a diverse set of real-world applications. This paper describes our contribution at SnakeCLEF2022 with FGVC. Firstly, we design a strong multimodal backbone to utilize various meta-information to assist in fine-grained identification. Secondly, we provide new loss functions to solve the long tail distribution with dataset. Then, in order to take full advantage of unlabeled datasets, we use self-supervised learning and supervised learning joint training to provide pre-trained model. Moreover, some effective data process tricks also are considered in our experiments. Last but not least, fine-tuned in downstream task with hard mining, ensambled kinds of model performance. Extensive experiments demonstrate that our method can effectively improve the performance of fine-grained recognition. Our method can achieve a macro f1 score 92.7% and 89.4% on private and public dataset, respectively, which is the 1st place among the participators on private leaderboard.

cs.CV↗

arXiv2022

OGLE-2018-BLG-0799Lb: a $q \sim 2.7 \times 10^{-3}$ Planet with Spitzer Parallax

We report the discovery and analysis of a planet in the microlensing event OGLE-2018-BLG-0799. The planetary signal was observed by several ground-based telescopes, and the planet-host mass ratio is $q = (2.65 \pm 0.16) \times 10^{-3}$. The ground-based observations yield a constraint on the angular Einstein radius $θ_{\rm E}$, and the microlensing parallax vector $\vecπ_{\rm E}$, is strongly constrained by the Spitzer data. However, the 2019 Spitzer baseline data reveal systematics in the Spitzer photometry, so there is ambiguity in the magnitude of the parallax. In our preferred interpretation, a full Bayesian analysis using a Galactic model indicates that the planetary system is composed of an $M_{\rm planet} = 0.26_{-0.11}^{+0.22}~M_{J}$ planet orbiting an $M_{\rm host} = 0.093_{-0.038}^{+0.082}~M_{\odot}$, at a distance of $D_{\rm L} = 3.71_{-1.70}^{+3.24}$ kpc. An alternate interpretation of the data shifts the localization of the minima along the arc-shaped microlens parallax constraints. This, in turn, yields a more massive host with median mass of $0.13 {M_{\odot}}$ at a distance of 6.3 kpc. This analysis demonstrates the robustness of the osculating circles formalism, but shows that further investigation is needed to assess how systematics affect the specific localization of the microlens parallax vector and, consequently, the inferred physical parameters.

astro-ph.EP↗

arXiv2022

Systematic KMTNet Planetary Anomaly Search. IV. Complete Sample of 2019 Prime-Field

We report the complete statistical planetary sample from the prime fields ($Γ\geq 2~{\rm hr}^{-1}$) of the 2019 Korea Microlensing Telescope Network (KMTNet) microlensing survey. We develop the optimized KMTNet AnomalyFinder algorithm and apply it to the 2019 KMTNet prime fields. We find a total of 14 homogeneously selected planets and report the analysis of three planetary events, KMT-2019-BLG-(1042,1552,2974). The planet-host mass ratios, $q$, for the three planetary events are $6.34 \times 10^{-4}, 4.89 \times 10^{-3}$ and $6.18 \times 10^{-4}$, respectively. A Bayesian analysis indicates the three planets are all cold giant planets beyond the snow line of their host stars. The 13 planets are basically uniform in $\log q$ over the range $-5.0 < \log q < -1.5$. This result suggests that the planets below $q_{\rm break} = 1.7 \times 10^{-4}$ proposed by the MOA-II survey may be more common than previously believed. This work is an early component of a large project to determine the KMTNet mass-ratio function, and the whole sample of 2016--2019 KMTNet events should contain about 120 planets.

astro-ph.EP↗

arXiv2022

Quantum Algorithm Implementations for Beginners

As quantum computers become available to the general public, the need has arisen to train a cohort of quantum programmers, many of whom have been developing classical computer programs for most of their careers. While currently available quantum computers have less than 100 qubits, quantum computing hardware is widely expected to grow in terms of qubit count, quality, and connectivity. This review aims to explain the principles of quantum programming, which are quite different from classical programming, with straightforward algebra that makes understanding of the underlying fascinating quantum mechanical principles optional. We give an introduction to quantum computing algorithms and their implementation on real quantum hardware. We survey 20 different quantum algorithms, attempting to describe each in a succinct and self-contained fashion. We show how these algorithms can be implemented on IBM's quantum computer, and in each case, we discuss the results of the implementation with respect to differences between the simulator and the actual hardware runs. This article introduces computer scientists, physicists, and engineers to quantum algorithms and provides a blueprint for their implementations.

cs.ET↗

arXiv2022

Two Candidate KH 15D-like Systems from the Zwicky Transient Facility

KH 15D contains a circumbinary disk that is tilted relative to the orbital plane of the central binary. The precession of the disk and the orbital motion of the binary together produce rich phenomena in the photometric light curve. In this work, we present the discovery and preliminary analysis of two objects that resemble the key features of KH 15D from the Zwicky Transient Facility. These new objects, Bernhard-1 and Bernhard-2, show large-amplitude ($>1.5\,$mag), long-duration (more than tens of days), and periodic dimming events. A one-sided screen model is developed to model the photometric behaviour of these objects, the physical interpretation of which is a tilted, warped circumbinary disk occulting the inner binary. Changes in the object light curves suggest potential precession periods over timescales longer than 10 years. Additional photometric and spectroscopic observations are encouraged to better understand the nature of these interesting systems.

astro-ph.SR↗

arXiv2022

Structure-preserving GANs

Generative adversarial networks (GANs), a class of distribution-learning methods based on a two-player game between a generator and a discriminator, can generally be formulated as a minmax problem based on the variational representation of a divergence between the unknown and the generated distributions. We introduce structure-preserving GANs as a data-efficient framework for learning distributions with additional structure such as group symmetry, by developing new variational representations for divergences. Our theory shows that we can reduce the discriminator space to its projection on the invariant discriminator space, using the conditional expectation with respect to the sigma-algebra associated to the underlying structure. In addition, we prove that the discriminator space reduction must be accompanied by a careful design of structured generators, as flawed designs may easily lead to a catastrophic "mode collapse" of the learned distribution. We contextualize our framework by building symmetry-preserving GANs for distributions with intrinsic group symmetry, and demonstrate that both players, namely the equivariant generator and invariant discriminator, play important but distinct roles in the learning process. Empirical experiments and ablation studies across a broad range of data sets, including real-world medical imaging, validate our theory, and show our proposed methods achieve significantly improved sample fidelity and diversity -- almost an order of magnitude measured in Fréchet Inception Distance -- especially in the small data regime.

cs.LG↗