arXiv subjects

Yu Lu

Publications and source records attributed to Yu Lu.

At least 163 records · Page 9Linked to original sources

arXiv2017

The Nature of Massive Transition Galaxies in CANDELS, GAMA, and Cosmological Simulations

We explore observational and theoretical constraints on how galaxies might transition between the "star-forming main sequence" (SFMS) and varying "degrees of quiescence" out to $z=3$. Our analysis is focused on galaxies with stellar mass $M_*>10^{10}M_{\odot}$, and is enabled by GAMA and CANDELS observations, a semi-analytic model (SAM) of galaxy formation, and a cosmological hydrodynamical "zoom in" simulation with momentum-driven AGN feedback. In both the observations and the SAM, transition galaxies tend to have intermediate Sérsic indices, half-light radii, and surface stellar mass densities compared to star-forming and quiescent galaxies out to $z=3$. We place an observational upper limit on the average population transition timescale as a function of redshift, finding that the average high-redshift galaxy is on a "fast track" for quenching whereas the average low-redshift galaxy is on a "slow track" for quenching. We qualitatively identify four physical origin scenarios for transition galaxies in the SAM: oscillations on the SFMS, slow quenching, fast quenching, and rejuvenation. Quenching timescales in both the SAM and the hydrodynamical simulation are not fast enough to reproduce the quiescent population that we observe at $z\sim3$. In the SAM, we do not find a clear-cut morphological dependence of quenching timescales, but we do predict that the mean stellar ages, cold gas fractions, SMBH masses, and halo masses of transition galaxies tend to be intermediate relative to those of star-forming and quiescent galaxies at $z<3$.

astro-ph.GA↗

arXiv2017

Structured Matrix Estimation and Completion

We study the problem of matrix estimation and matrix completion under a general framework. This framework includes several important models as special cases such as the gaussian mixture model, mixed membership model, bi-clustering model and dictionary learning. We consider the optimal convergence rates in a minimax sense for estimation of the signal matrix under the Frobenius norm and under the spectral norm. As a consequence of our general result we obtain minimax optimal rates of convergence for various special models.

math.ST↗

arXiv2017

Modeling Charmonium-$η$ Decays of $J^{PC}=1^{--}$ Higher Charmonia

We propose a new model to create a light meson in the heavy quarkonium transition, which is inspired by the Nambu$-$Jona-Lasinio (NJL) model. Hadronic transitions of $J^{PC}=1^{--}$ higher charmonia with the emission of an $η$ meson are studied in the framework of the proposed model. The model shows its potential to reproduce the observed decay widths and make predictions for the unobserved channels. We present our predictions for the decay width of $Ψ\to J/ψη$ and $Ψ\to h_{c}(1P)η$, where $Ψ$ are higher $S$ and $D$ wave vector charmonia, which provide useful references to search for higher charmonia and determine their properties in forthcoming experiments. The predicted branching fraction $\mathcal B(ψ(4415)\to h_{c}(1P)η)=4.62\times 10^{-4}$ is one order of magnitude smaller than the $J/ψη$ channel. Estimates of partial decay width $Γ(Y \to J/ψη)$ are given for $Y(4360)$, $Y(4390)$ and $Y(4660)$ by assuming them as $c\bar{c}$ bound states with quantum numbers $3 ^3D_1$, $3 ^3D_1$ and $5 ^3S_1$, respectively. Our results are in favor of these assignments for $Y(4360)$ and $Y(4660)$. The corresponding experimental data for these $Y$ states has large statistical errors which do not provide any constraint on the mixing angle if we introduce $S-D$ mixing. To identify $Y(4390)$, precise measurements on its hadronic branching fraction are required which are eagerly awaited from BESIII.

hep-ph↗

arXiv2017

Provably Optimal Algorithms for Generalized Linear Contextual Bandits

Contextual bandits are widely used in Internet services from news recommendation to advertising, and to Web search. Generalized linear models (logistical regression in particular) have demonstrated stronger performance than linear models in many applications where rewards are binary. However, most theoretical analyses on contextual bandits so far are on linear bandits. In this work, we propose an upper confidence bound based algorithm for generalized linear contextual bandits, which achieves an $\tilde{O}(\sqrt{dT})$ regret over $T$ rounds with $d$ dimensional feature vectors. This regret matches the minimax lower bound, up to logarithmic terms, and improves on the best previous result by a $\sqrt{d}$ factor, assuming the number of arms is fixed. A key component in our analysis is to establish a new, sharp finite-sample confidence bound for maximum-likelihood estimates in generalized linear models, which may be of independent interest. We also analyze a simpler upper confidence bound algorithm, which is useful in practice, and prove it to have optimal regret for certain cases.

cs.LG↗

arXiv2017

CANDELS Sheds Light on the Environmental Quenching of Low-mass Galaxies

We investigate the environmental quenching of galaxies, especially those with stellar masses (M*)$<10^{9.5} M_\odot$, beyond the local universe. Essentially all local low-mass quenched galaxies (QGs) are believed to live close to massive central galaxies, which is a demonstration of environmental quenching. We use CANDELS data to test {\it whether or not} such a dwarf QG--massive central galaxy connection exists beyond the local universe. To this purpose, we only need a statistically representative, rather than a complete, sample of low-mass galaxies, which enables our study to $z\gtrsim1.5$. For each low-mass galaxy, we measure the projected distance ($d_{proj}$) to its nearest massive neighbor (M*$>10^{10.5} M_\odot$) within a redshift range. At a given redshift and M*, the environmental quenching effect is considered to be observed if the $d_{proj}$ distribution of QGs ($d_{proj}^Q$) is significantly skewed toward lower values than that of star-forming galaxies ($d_{proj}^{SF}$). For galaxies with $10^{8} M_\odot < M* < 10^{10} M_\odot$, such a difference between $d_{proj}^Q$ and $d_{proj}^{SF}$ is detected up to $z\sim1$. Also, about 10\% of the quenched galaxies in our sample are located between two and four virial radii ($R_{Vir}$) of the massive halos. The median projected distance from low-mass QGs to their massive neighbors, $d_{proj}^Q / R_{Vir}$, decreases with satellite M* at $M* \lesssim 10^{9.5} M_\odot$, but increases with satellite M* at $M* \gtrsim 10^{9.5} M_\odot$. This trend suggests a smooth, if any, transition of the quenching timescale around $M* \sim 10^{9.5} M_\odot$ at $0.5<z<1.0$.

astro-ph.GA↗

arXiv2017

How Large is the Contribution of Excited Mesons in Coupled-Channel Effects?

We study the excited $B$ mesons' contributions to the coupled-channel effects under the framework of ${}^3P_0$ model for the bottomonium. Contrary to what has been widely accepted, the contributions of $P$ wave $B$ mesons are generally the largest and this result is independent of the potential parameters to some extent. We also push the calculation beyond $B(1P)$ and carefully analyze the contributions of $B(2S)$. A form factor is a key ingredient to suppress the contributions of $B(2S)$ for low lying bottomonia. However, this suppression mechanism is not efficient for highly excited bottomonia such as $Υ(5S)$ and $Υ(6S)$. We give explanations why this difficulty happens to ${}^3P_0$ model and suggest analyzing flux-tube breaking model for the full calculation of coupled-channel effects.

hep-ph↗

arXiv2016

Statistical and Computational Guarantees of Lloyd's Algorithm and its Variants

Clustering is a fundamental problem in statistics and machine learning. Lloyd's algorithm, proposed in 1957, is still possibly the most widely used clustering algorithm in practice due to its simplicity and empirical performance. However, there has been little theoretical investigation on the statistical and computational guarantees of Lloyd's algorithm. This paper is an attempt to bridge this gap between practice and theory. We investigate the performance of Lloyd's algorithm on clustering sub-Gaussian mixtures. Under an appropriate initialization for labels or centers, we show that Lloyd's algorithm converges to an exponentially small clustering error after an order of $\log n$ iterations, where $n$ is the sample size. The error rate is shown to be minimax optimal. For the two-mixture case, we only require the initializer to be slightly better than random guess. In addition, we extend the Lloyd's algorithm and its analysis to community detection and crowdsourcing, two problems that have received a lot of attention recently in statistics and machine learning. Two variants of Lloyd's algorithm are proposed respectively for community detection and crowdsourcing. On the theoretical side, we provide statistical and computational guarantees of the two algorithms, and the results improve upon some previous signal-to-noise ratio conditions in literature for both problems. Experimental results on simulated and real data sets demonstrate competitive performance of our algorithms to the state-of-the-art methods.

math.ST↗

arXiv2016

Evidence for Reduced Specific Star Formation Rates in the Centers of Massive Galaxies at z = 4

We perform the first spatially-resolved stellar population study of galaxies in the early universe (z = 3.5 - 6.5), utilizing the Hubble Space Telescope Cosmic Assembly Near-infrared Deep Extragalactic Legacy Survey (CANDELS) imaging dataset over the GOODS-S field. We select a sample of 418 bright and extended galaxies at z = 3.5 - 6.5 from a parent sample of ~ 8000 photometric-redshift selected galaxies from Finkelstein et al. (2015). We first examine galaxies at 3.5< z < 4.0 using additional deep K-band survey data from the HAWK-I UDS and GOODS Survey (HUGS) which covers the 4000A break at these redshifts. We measure the stellar mass, star formation rate, and dust extinction for galaxy inner and outer regions via spatially-resolved spectral energy distribution fitting based on a Markov Chain Monte Carlo algorithm. By comparing specific star formation rates (sSFRs) between inner and outer parts of the galaxies we find that the majority of galaxies with the high central mass densities show evidence for a preferentially lower sSFR in their centers than in their outer regions, indicative of reduced sSFRs in their central regions. We also study galaxies at z ~ 5 and 6 (here limited to high spatial resolution in the rest-frame ultraviolet only), finding that they show sSFRs which are generally independent of radial distance from the center of the galaxies. This indicates that stars are formed uniformly at all radii in massive galaxies at z ~ 5 - 6, contrary to massive galaxies at z < 4.

astro-ph.GA↗

arXiv2016

Coupled-Channel Effects for the Bottomonium with Realistic Wave Functions

With Gaussian expansion method (GEM), realistic wave functions are used to calculate coupled-channel effects for the bottomonium under the framework of ${}^3P_0$ model. The simplicity and accuracy of GEM are explained. We calculate the mass shifts, probabilities of the $B$ meson continuum, $S-D$ mixing angles, strong and dielectric decay widths. Our calculation shows that both $S-D$ mixing and the $B$ meson continuum can contribute to the suppression of the vector meson's dielectric decay width. We suggest more precise measurements on the radiative decays of $Υ(10580)$ and $Υ(11020)$ to distinguish these two effects. The above quantities are also calculated with simple harmonic oscillator (SHO) wave function approximation for comparison. The deviation between GEM and SHO indicates that it is essential to treat the wave functions accurately for near threshold states.

hep-ph↗

arXiv2016

The connection between the host halo and the satellite galaxies of the Milky Way

Many properties of the Milky Way's dark matter halo, including its mass assembly history, concentration, and subhalo population, remain poorly constrained. We explore the connection between these properties of the Milky Way and its satellite galaxy population, especially the implication of the presence of the Magellanic Clouds for the properties of the Milky Way halo. Using a suite of high-resolution $N$-body simulations of Milky Way-mass halos with a fixed final Mvir ~ 10^{12.1}Msun, we find that the presence of Magellanic Cloud-like satellites strongly correlates with the assembly history, concentration, and subhalo population of the host halo, such that Milky Way-mass systems with Magellanic Clouds have lower concentration, more rapid recent accretion, and more massive subhalos than typical halos of the same mass. Using a flexible semi-analytic galaxy formation model that is tuned to reproduce the stellar mass function of the classical dwarf galaxies of the Milky Way with Markov-Chain Monte-Carlo, we show that adopting host halos with different mass-assembly histories and concentrations can lead to different best-fit models for galaxy-formation physics, especially for the strength of feedback. These biases arise because the presence of the Magellanic Clouds boosts the overall population of high-mass subhalos, thus requiring a different stellar-mass-to-halo-mass ratio to match the data. These biases also lead to significant differences in the mass--metallicity relation, the kinematics of low-mass satellites, the number counts of small satellites associated with the Magellanic Clouds, and the stellar mass of Milky Way itself. Observations of these galaxy properties can thus provide useful constraints on the properties of the Milky Way halo.

astro-ph.GA↗

arXiv2016

The Evolution of the Galaxy Stellar Mass Function at z= 4-8: A Steepening Low-mass-end Slope with Increasing Redshift

We present galaxy stellar mass functions (GSMFs) at $z=$ 4-8 from a rest-frame ultraviolet (UV) selected sample of $\sim$4500 galaxies, found via photometric redshifts over an area of $\sim$280 arcmin$^2$ in the CANDELS/GOODS fields and the Hubble Ultra Deep Field. The deepest Spitzer/IRAC data yet-to-date and the relatively large volume allow us to place a better constraint at both the low- and high-mass ends of the GSMFs compared to previous space-based studies from pre-CANDELS observations. Supplemented by a stacking analysis, we find a linear correlation between the rest-frame UV absolute magnitude at 1500 Å ($M_{\rm UV}$) and logarithmic stellar mass ($\log M_*$) that holds for galaxies with $\log(M_*/M_{\odot}) \lesssim 10$. We use simulations to validate our method of measuring the slope of the $\log M_*$-$M_{\rm UV}$ relation, finding that the bias is minimized with a hybrid technique combining photometry of individual bright galaxies with stacked photometry for faint galaxies. The resultant measured slopes do not significantly evolve over $z=$ 4-8, while the normalization of the trend exhibits a weak evolution toward lower masses at higher redshift. We combine the $\log M_*$-$M_{\rm UV}$ distribution with observed rest-frame UV luminosity functions at each redshift to derive the GSMFs, finding that the low-mass-end slope becomes steeper with increasing redshift from $α=-1.55^{+0.08}_{-0.07}$ at $z=4$ to $α=-2.25^{+0.72}_{-0.35}$ at $z=8$. The inferred stellar mass density, when integrated over $M_*=10^8$-$10^{13} M_{\odot}$, increases by a factor of $10^{+30}_{-2}$ between $z=7$ and $z=4$ and is in good agreement with the time integral of the cosmic star formation rate density.

astro-ph.GA↗

arXiv2016

Exact Exponent in Optimal Rates for Crowdsourcing

In many machine learning applications, crowdsourcing has become the primary means for label collection. In this paper, we study the optimal error rate for aggregating labels provided by a set of non-expert workers. Under the classic Dawid-Skene model, we establish matching upper and lower bounds with an exact exponent $mI(π)$ in which $m$ is the number of workers and $I(π)$ the average Chernoff information that characterizes the workers' collective ability. Such an exact characterization of the error exponent allows us to state a precise sample size requirement $m>\frac{1}{I(π)}\log\frac{1}ε$ in order to achieve an $ε$ misclassification error. In addition, our results imply the optimality of various EM algorithms for crowdsourcing initialized by consistent estimators.

stat.ML↗

arXiv2016

Stellar Mass--Gas-phase Metallicity Relation at $0.5\leq z\leq0.7$: A Power Law with Increasing Scatter toward the Low-mass Regime

We present the stellar mass ($M_{*}$)--gas-phase metallicity relation (MZR) and its scatter at intermediate redshifts ($0.5\leq z\leq0.7$) for 1381 field galaxies collected from deep spectroscopic surveys. The star formation rate (SFR) and color at a given $M_{*}$ of this magnitude-limited ($R\lesssim24$ AB) sample are representative of normal star-forming galaxies. For masses below $10^9 M_\odot$, our sample of 237 galaxies is $\sim$10 times larger than those in previous studies beyond the local universe. This huge gain in sample size enables superior constraints on the MZR and its scatter in the low-mass regime. We find a power-law MZR at $10^{8} M_\odot < M_{*} < 10^{11} M_\odot$: ${12+log(O/H) = (5.83\pm0.19) + (0.30\pm0.02)log(M_{*}/M_\odot)}$. Our MZR shows good agreement with others measured at similar redshifts in the literature in the intermediate and massive regimes, but is shallower than the extrapolation of the MZRs of others to masses below $10^{9} M_\odot$. The SFR dependence of the MZR in our sample is weaker than that found for local galaxies (known as the Fundamental Metallicity Relation). Compared to a variety of theoretical models, the slope of our MZR for low-mass galaxies agrees well with predictions incorporating supernova energy-driven winds. Being robust against currently uncertain metallicity calibrations, the scatter of the MZR serves as a powerful diagnostic of the stochastic history of gas accretion, gas recycling, and star formation of low-mass galaxies. Our major result is that the scatter of our MZR increases as $M_{*}$ decreases. Our result implies that either the scatter of the baryonic accretion rate or the scatter of the $M_{*}$--$M_{halo}$ relation increases as $M_{*}$ decreases. Moreover, our measures of scatter at $z=0.7$ appears consistent with that found for local galaxies.

astro-ph.GA↗

arXiv2015

Rate-optimal graphon estimation

Network analysis is becoming one of the most active research areas in statistics. Significant advances have been made recently on developing theories, methodologies and algorithms for analyzing networks. However, there has been little fundamental study on optimal estimation. In this paper, we establish optimal rate of convergence for graphon estimation. For the stochastic block model with $k$ clusters, we show that the optimal rate under the mean squared error is $n^{-1}\log k+k^2/n^2$. The minimax upper bound improves the existing results in literature through a technique of solving a quadratic equation. When $k\leq\sqrt{n\log n}$, as the number of the cluster $k$ grows, the minimax rate grows slowly with only a logarithmic order $n^{-1}\log k$. A key step to establish the lower bound is to construct a novel subset of the parameter space and then apply Fano's lemma, from which we see a clear distinction of the nonparametric graphon estimation problem from classical nonparametric regression, due to the lack of identifiability of the order of nodes in exchangeable random graph models. As an immediate application, we consider nonparametric graphon estimation in a Hölder class with smoothness $α$. When the smoothness $α\geq1$, the optimal rate of convergence is $n^{-1}\log n$, independent of $α$, while for $α\in(0,1)$, the rate is $n^{-2α/(α+1)}$, which is, to our surprise, identical to the classical nonparametric rate.

math.ST↗

arXiv2015

Satellite Quenching and Galactic Conformity at 0.3 < z < 2.5

We measure the evolution of the quiescent fraction and quenching efficiency of satellites around star-forming and quiescent central galaxies with stellar mass $\log(M_{\mathrm{cen}}/M_{\odot})>10.5$ at $0.3 9.3$. Satellites for both star-forming and quiescent central galaxies have higher quiescent fractions compared to field galaxies matched in stellar mass at all redshifts. We also observe "galactic conformity": satellites around quiescent centrals are more likely to be quenched compared to the satellites around star-forming centrals. In our sample, this conformity signal is significant at $\gtrsim3σ$ for $0.6<z<1.6$, whereas it is only weakly significant at $0.3<z<0.6$ and $1.6<z<2.5$. Therefore, conformity (and therefore satellite quenching) has been present for a significant fraction of the age of the universe. The satellite quenching efficiency increases with increasing stellar mass of the central, but does not appear to depend on the stellar mass of the satellite to the mass limit of our sample. When we compare the satellite quenching efficiency of star-forming centrals with stellar masses 0.2 dex higher than quiescent centrals (which should account for any difference in halo mass), the conformity signal decreases, but remains statistically significant at $0.6<z<0.9$. This is evidence that satellite quenching is connected to the star-formation properties of the central as well as to the mass of the halo. We discuss physical effects that may contribute to galactic conformity, and emphasize that they must allow for continued star-formation in the central galaxy even as the satellites are quenched.

astro-ph.GA↗

arXiv2015

Estimating stellar atmospheric parameters based on LASSO and support-vector regression

A scheme for estimating atmospheric parameters T$_{eff}$, log$~g$, and [Fe/H] is proposed on the basis of Least Absolute Shrinkage and Selection Operator (LASSO) algorithm and Haar wavelet. The proposed scheme consists of three processes. A spectrum is decomposed using the Haar wavelet transform and low-frequency components at the fourth level are considered as candidate features. Then, spectral features from the candidate features are detected using the LASSO algorithm to estimate the atmospheric parameters. Finally, atmospheric parameters are estimated from the extracted spectral features using the support-vector regression (SVR) method. The proposed scheme was evaluated using three sets of stellar spectra respectively from Sloan Digital Sky Survey (SDSS), Large Sky Area Multi-object Fiber Spectroscopic Telescope (LAMOST), and Kurucz's model, respectively. The mean absolute errors are as follows: for 40~000 SDSS spectra, 0.0062 dex for log~T$_{eff}$ (85.83 K for T$_{eff}$), 0.2035 dex for log$~g$ and 0.1512 dex for [Fe/H]; for 23963 LAMOST spectra, 0.0074 dex for log~T$_{eff}$ (95.37 K for T$_{eff}$), 0.1528 dex for log~$g$, and 0.1146 dex for [Fe/H]; and for 10469 synthetic spectra, 0.0010 dex for log T$_{eff}$(14.42K for T$_{eff}$), 0.0123 dex for log~$g$, and 0.0125 dex for [Fe/H].

astro-ph.SR↗

arXiv2015

An analytical model for galaxy metallicity: What do metallicity relations tell us about star formation and outflow?

We develop a simple analytical model that tracks galactic metallicities governed by star formation and feedback to gain insight from the observed galaxy stellar mass-metallicity relations over a large range of stellar masses and redshifts. The model reveals the following implications of star formation and feedback processes in galaxy formation. First, the observed metallicity relations provide a stringent upper limit for the averaged outflow mass-loading factors of local galaxies, which is ~20 for M_*~10^9Msun galaxies and monotonically decreases to ~1 for M_*~10^{11}Msun galaxies. Second, the inferred upper-limit for the outflow mass-loading factor sensitively depends on whether the outflow is metal-enriched with respect to the ISM metallicity. If half of the metals ejected from SNe leave the galaxy in metal-enriched winds, the outflow mass-loading factor for galaxies at any mass can barely be higher than ~10, which puts strong constraints on galaxy formation models. Third, the relatively lower stellar-phase to gas-phase metallicity ratio for lower-mass galaxies indicate that low-mass galaxies are still rapidly enriching their metallicities in recent times, while high-mass galaxies are more settled, which seems to show a downsizing effect in the metallicity evolution of galaxies. The analysis presented in the paper demonstrates the importance of accurate measurements of galaxy metallicities and the cold gas fraction of galaxies at different redshifts for constraining star formation and feedback processes, and demonstrates the power of these relations in constraining the physics of galaxy formation.

astro-ph.GA↗

arXiv2015

Using Galaxy Pairs to Probe Star Formation During Major Halo Mergers

Currently-proposed galaxy quenching mechanisms predict very different behaviours during major halo mergers, ranging from significant quenching enhancement (e.g., clump-induced gravitational heating models) to significant star formation enhancement (e.g., gas starvation models). To test real galaxies' behaviour, we present an observational galaxy pair method for selecting galaxies whose host haloes are preferentially undergoing major mergers. Applying the method to central L* (10^10 Msun < M_* < 10^10.5 Msun) galaxies in the Sloan Digital Sky Survey (SDSS) at z<0.06, we find that major halo mergers can at most modestly reduce the star-forming fraction, from 59% to 47%. Consistent with past research, however, mergers accompany enhanced specific star formation rates for star-forming L* centrals: ~10% when a paired galaxy is within 200 kpc (approximately the host halo's virial radius), climbing to ~70% when a paired galaxy is within 30 kpc. No evidence is seen for even extremely close pairs (<30 kpc separation) rejuvenating star formation in quenched galaxies. For galaxy formation models, our results suggest: (1) quenching in L* galaxies likely begins due to decoupling of the galaxy from existing hot and cold gas reservoirs, rather than a lack of available gas or gravitational heating from infalling clumps, (2) state-of-the-art semi-analytic models currently over-predict the effect of major halo mergers on quenching, and (3) major halo mergers can trigger enhanced star formation in non-quenched central galaxies.

astro-ph.GA↗