Search

We present a detailed characterization of the Lya properties for 93 Lya emitters (LAEs) at z~3.1 selected from the D1 field of the Canada-France-Hawaii-Telescope Legacy Survey, including 24 members of a massive protocluster. The median-stacked Lya image shows an extended Lya halo (LAH) surrounding the galaxy with the exponential scale length 4.9+/-0.7kpc, which accounts for roughly half of the total line flux. Accounting for the LAH contribution, the total Lya escape fraction, f_esc, is 40+/-26%. Combining the dataset with existing measurements, we find a dependence of f_esc on the galaxy's UV slope (beta) and UV luminosity (L_UV). The simultaneous use of both parameters allows prediction of f_esc within 0.18dex, a substantial improvement over 0.23dex when only beta is used. The correlation between f_esc and E(B-V) suggests that Lya photons undergo interstellar dust attenuation in a similar manner to continuum photons. Yet, Lya transmission is typically higher than that expected for continuum photons at similar wavelength by a factor, which depends on UV luminosity, up to 2 in the samples we studied. These results hint at complex geometries and physical conditions of the interstellar medium, which affect the Lya transmission or production. Alternatively, the dust law may change with luminosity leading to over-or under-estimation of f_esc. Finally, we report that protocluster member LAEs tend to be bluer and more UV-luminous than their field cousins, resulting in systematically higher f_esc values. We speculate that it may be due to the widespread formation of young low-mass galaxies in dense gas-rich environments., Comment: 27 pages, 12 figures, accepted by ApJ

Ly$\alpha$-emitting galaxies (LAEs) are easily detectable in the high-redshift Universe and are potentially efficient tracers of large scale structure at early epochs, as long as their observed properties do not strongly depend on environment. We investigate the luminosity and equivalent width functions of LAEs in the overdense field of a protocluster at redshift $z \simeq 3.78$. Using a large sample of LAEs (many spectroscopically confirmed), we find that the Ly$\alpha$ luminosity distribution is well-represented by a Schechter (1976) function with $\log(L^{\ast}/{\rm erg s^{-1}}) = 43.26^{+0.20}_{-0.22}$ and $\log(\phi^{\ast}/{\rm Mpc^{-3}})=-3.40^{+0.03}_{-0.04}$ with $\alpha=-1.5$. Fitting the equivalent width distribution as an exponential, we find a scale factor of $\omega=79^{+15}_{-15}$ Angstroms. We also measured the Ly$\alpha$ luminosity and equivalent width functions using the subset of LAEs lying within the densest cores of the protocluster, finding similar values for $L^*$ and $\omega$. Hence, despite having a mean overdensity more than 2$\times$ that of the general field, the shape of the Ly$\alpha$ luminosity function and equivalent width distributions in the protocluster region are comparable to those measured in the field LAE population by other studies at similar redshift. While the observed Ly$\alpha$ luminosities and equivalent widths show correlations with the UV continuum luminosity in this LAE sample, we find that these are likely due to selection biases and are consistent with no intrinsic correlations within the sample. This protocluster sample supports the strong evolutionary trend observed in the Ly$\alpha$ escape fraction and suggest that lower redshift LAEs can be on average significantly more dusty that their counterparts at higher redshift., Comment: Updated version to match the one accepted by ApJ

We present a multiwavelength study of galaxies around D4UD01, a spectroscopically confirmed protocluster at z = 3.24 to investigate environmental trends. 450 galaxies are selected based on Ks band detection with photometric redshifts (photo-z) at 3.0 < z < 3.4, among which ~ 12% are classified as quiescent galaxies. The quiescent galaxies are among the most massive and reddest ones in the entire sample. We identify a large photo-z galaxy overdensity in the field, which lies close to the previously spectroscopically confirmed sources of the protocluster. We find that the quiescent galaxies are largely concentrated in the overdense protocluster region with a higher quiescent fraction, showing a sign of environmental quenching. Galaxies in the protocluster are forming faster than the field counterparts as seen in the stellar mass function, suggesting early and accelerated mass assembly in the overdense regions. Although weak evidence of suppressed star-formation is found in the protocluster, the statistics are not significant enough to draw a definite conclusion. Our work shed light on how the formation of massive galaxies is affected in the dense region of a protocluster when the Universe was only 2 Gyr old., Comment: 16 pages, 7 figures, accepted to ApJ

Document-level discourse parsing, in accordance with the Rhetorical Structure Theory (RST), remains notoriously challenging. Challenges include the deep structure of document-level discourse trees, the requirement of subtle semantic judgments, and the lack of large-scale training corpora. To address such challenges, we propose to exploit robust representations derived from multiple levels of granularity across syntax and semantics, and in turn incorporate such representations in an end-to-end encoder-decoder neural architecture for more resourceful discourse processing. In particular, we first use a pre-trained contextual language model that embodies high-order and long-range dependency to enable finer-grain semantic, syntactic, and organizational representations. We further encode such representations with boundary and hierarchical information to obtain more refined modeling for document-level discourse processing. Experimental results show that our parser achieves the state-of-the-art performance, approaching human-level performance on the benchmarked RST dataset., Comment: 11 pages, 3 figures, 7 tables

Text discourse parsing plays an important role in understanding information flow and argumentative structure in natural language. Previous research under the Rhetorical Structure Theory (RST) has mostly focused on inducing and evaluating models from the English treebank. However, the parsing tasks for other languages such as German, Dutch, and Portuguese are still challenging due to the shortage of annotated data. In this work, we investigate two approaches to establish a neural, cross-lingual discourse parser via: (1) utilizing multilingual vector representations; and (2) adopting segment-level translation of the source content. Experiment results show that both methods are effective even with limited training data, and achieve state-of-the-art performance on cross-lingual, document-level discourse parsing on all sub-tasks., Comment: Published in COLING2020

We present a detailed study of Near-IR selected galaxies in a protocluster field at z = 3.13. Protocluster galaxies are selected using the available mutliwavelength data with the photometric redshift (photo-z) at 2.9 < z < 3.3, reaching a mass completeness of ~10^10 M_sun. Diverse types of galaxies have been found in the field including normal star-forming galaxies, quiescent galaxies and dusty star-forming galaxies. The photo-z galaxies form two large overdense structures in the field, largely overlapping with the previously identified galaxy overdensities traced by Ly{\alpha} emitters (LAEs) and Lyman break galaxies (LBGs) respectively. The northern overdensity consists of a large fraction of old and/or dusty galaxy populations, while the southern one is mainly composed of normal star-forming galaxies which are spatially correlated with the LAEs. This agrees with our previous study arguing the spatial offset of different galaxy overdensities may be due to halo assembly bias. Given the large end-to-end sizes of the two overdensities, one possibility is that they will form into a supercluster by the present day. We also find strong evidence that the star-formation activities of the galaxies in the overdense protocluster regions are enhanced in comparison to their field counterparts, which suggests an accelerated mass assembly in this protocluster., Comment: 19 pages, 8 figures, accepted for publication in ApJ

Much progress has been made in text summarization, fueled by neural architectures using large-scale training corpora. However, in the news domain, neural models easily overfit by leveraging position-related features due to the prevalence of the inverted pyramid writing style. In addition, there is an unmet need to generate a variety of summaries for different users. In this paper, we propose a neural framework that can flexibly control summary generation by introducing a set of sub-aspect functions (i.e. importance, diversity, position). These sub-aspect functions are regulated by a set of control codes to decide which sub-aspect to focus on during summary generation. We demonstrate that extracted summaries with minimal position bias is comparable with those generated by standard models that take advantage of position preference. We also show that news summaries generated with a focus on diversity can be more preferred by human raters. These results suggest that a more flexible neural summarization framework providing more control options could be desirable in tailoring to different user preferences, which is useful since it is often impractical to articulate such preferences for different applications a priori., Comment: Accepted to Findings of EMNLP 2020

We revise the structure-preserving finite element method in [K. Hu, Y. MA and J. Xu. (2017) Stable finite element methods preserving $\nabla \cdot \mathbf{B}=0$ exactly for MHD models. Numer. Math.,135, 371-396]. The revised method is semi-implicit in time-discretization. We prove the linearized scheme preserves the divergence free property for the magnetic field exactly at each time step. Further, we showed the linearized scheme is unconditionally stable and we obtain optimal convergence in the energy norm of the revised method even for solutions with low regularity.