Search

In recent years, world business in online discussions and opinion sharing on social media is booming. Re-entry prediction task is thus proposed to help people keep track of the discussions which they wish to continue. Nevertheless, existing works only focus on exploiting chatting history and context information, and ignore the potential useful learning signals underlying conversation data, such as conversation thread patterns and repeated engagement of target users, which help better understand the behavior of target users in conversations. In this paper, we propose three interesting and well-founded auxiliary tasks, namely, Spread Pattern, Repeated Target user, and Turn Authorship, as the self-supervised signals for re-entry prediction. These auxiliary tasks are trained together with the main task in a multi-task manner. Experimental results on two datasets newly collected from Twitter and Reddit show that our method outperforms the previous state-of-the-arts with fewer parameters and faster convergence. Extensive experiments and analysis show the effectiveness of our proposed models and also point out some key ideas in designing self-supervised tasks., Comment: Accepted by EMNLP 2021 Findings

With the increasing popularity of social media, online interpersonal communication now plays an essential role in people's everyday information exchange. Whether and how a newcomer can better engage in the community has attracted great interest due to its application in many scenarios. Although some prior works that explore early socialization have obtained salient achievements, they are focusing on sociological surveys based on the small group. To help individuals get through the early socialization period and engage well in online conversations, we study a novel task to foresee whether a newcomer's message will be responded to by other participants in a multi-party conversation (henceforth Successful New-entry Prediction). The task would be an important part of the research in online assistants and social media. To further investigate the key factors indicating such engagement success, we employ an unsupervised neural network, Variational Auto-Encoder (VAE), to examine the topic content and discourse behavior from newcomer's chatting history and conversation's ongoing context. Furthermore, two large-scale datasets, from Reddit and Twitter, are collected to support further research on new-entries. Extensive experiments on both Twitter and Reddit datasets show that our model significantly outperforms all the baselines and popular neural models. Additional explainable and visual analyses on new-entry behavior shed light on how to better join in others' discussions., Comment: Accepted by WWW 2022

Quotations are crucial for successful explanations and persuasions in interpersonal communications. However, finding what to quote in a conversation is challenging for both humans and machines. This work studies automatic quotation generation in an online conversation and explores how language consistency affects whether a quotation fits the given context. Here, we capture the contextual consistency of a quotation in terms of latent topics, interactions with the dialogue history, and coherence to the query turn's existing content. Further, an encoder-decoder neural framework is employed to continue the context with a quotation via language generation. Experiment results on two large-scale datasets in English and Chinese demonstrate that our quotation generation model outperforms the state-of-the-art models. Further analysis shows that topic, interaction, and query consistency are all helpful to learn how to quote in online conversations., Comment: Accepted by EMNLP 2020, updated with dataset link

To help individuals express themselves better, quotation recommendation is receiving growing attention. Nevertheless, most prior efforts focus on modeling quotations and queries separately and ignore the relationship between the quotations and the queries. In this work, we introduce a transformation matrix that directly maps the query representations to quotation representations. To better learn the mapping relationship, we employ a mapping loss that minimizes the distance of two semantic spaces (one for quotation and another for mapped-query). Furthermore, we explore using the words in history queries to interpret the figurative language of quotations, where quotation-aware attention is applied on top of history queries to highlight the indicator words. Experiments on two datasets in English and Chinese show that our model outperforms previous state-of-the-art models., Comment: ACL 2021

MAXI J1820+070 is a low-mass black hole X-ray binary system with high luminosity in both optical and X-ray bands during the outburst periods. We present extensive photometry in X-ray, ultraviolet, and optical bands, as well as densely-cadenced optical spectra, covering the phase from the beginning of optical outburst to $\sim$550 days. During the rebrightening process, the optical emission preceded the X-ray by 20.80 $\pm$ 2.85 days. The spectra are characterized by blue continua and emission features of Balmer series, He I, He II lines and broad Bowen blend. The pseudo equivalent width (pEW) of emission lines are found to show anticorrelations with the X-ray flux measured at comparable phases, which is due to the increased suppression by the optical continuum. At around the X-ray peak, the full width at half maximums (FWHMs) of H$_{\beta}$ and He II $\lambda$4686 tend to stabilize at 19.4 Angstrom and 21.8 Angstrom, which corresponds to the line forming region at a radius of 1.7 and 1.3 R_sun within the disk. We further analyzed the absolute fluxes of the lines and found that the fluxes of H$_{\beta}$ and He II $\lambda$4686 show positive correlations with the X-ray flux, favoring that the irradiation model is responsible for the optical emission. However, the fact that X-ray emission experiences a dramatic flux drop at t$\sim$200 days after the outburst, while the optical flux only shows little variations suggests that additional energy such as viscous energy may contribute to the optical radiation in addition to the X-ray irradiation., Comment: Accepted for publication in MNRAS

We present multi-band photometry and spectroscopy of SN 2018cuf, a Type IIP ("P" for plateau) supernova (SN) discovered by the Distance Less Than 40 Mpc survey (DLT40) within 24 hours of explosion. SN 2018cuf appears to be a typical Type IIP SN, with an absolute $V$-band magnitude of $-$16.73 $\pm$ 0.32 at maximum and a decline rate of 0.21 $\pm$ 0.05 mag/50d during the plateau phase. The distance of the object is constrained to be 41.8 $\pm$ 5.7 Mpc by using the expanding photosphere method. We use spectroscopic and photometric observations from the first year after the explosion to constrain the progenitor of SN 2018cuf using both hydrodynamic light curve modelling and late-time spectroscopic modelling. The progenitor of SN 2018cuf was most likely a red supergiant of about 14.5 $\rm M_{\odot}$ that produced 0.04 $\pm$ 0.01 $\rm M_{\odot}$ $\rm ^{56}Ni$ during the explosion. We also found $\sim$ 0.07 $\rm M_{\odot}$ of circumstellar material (CSM) around the progenitor is needed to fit the early light curves, where the CSM may originate from pre-supernova outbursts. During the plateau phase, high velocity features at $\rm \sim 11000\ km~s^{-1}$ are detected both in the optical and near-infrared spectra, supporting the possibility that the ejecta were interacting with some CSM. A very shallow slope during the post-plateau phase is also observed and it is likely due to a low degree of nickel mixing or the relatively high nickel mass in the SN., Comment: 22 pages, 15 figures, submitted to ApJ

Supernova (SN) 2017cbv in NGC 5643 is one of a handful of type Ia supernovae (SNe~Ia) reported to have excess blue emission at early times. This paper presents extensive $BVRIYJHK_s$-band light curves of SN 2017cbv, covering the phase from $-16$ to $+125$ days relative to $B$-band maximum light. SN 2017cbv reached a $B$-band maximum of 11.710$\pm$0.006~mag, with a post-maximum magnitude decline $\Delta m_{15}(B)$=0.990$\pm$0.013 mag. The supernova suffered no host reddening based on Phillips intrinsic color, Lira-Phillips relation, and the CMAGIC diagram. By employing the CMAGIC distance modulus $\mu=30.58\pm0.05$~mag and assuming $H_0$=72~$\rm km \ s^{-1} \ Mpc^{-1}$, we found that 0.73~\msun $^{56}$Ni was synthesized during the explosion of SN 2017cbv, which is consistent with estimates using reddening-free and distance-free methods via the phases of the secondary maximum of the NIR-band light curves. We also present 14 near-infrared spectra from $-18$ to $+49$~days relative to the $B$-band maximum light, providing constraints on the amount of swept-up hydrogen from the companion star in the context of the single degenerate progenitor scenario. No $Pa{\beta}$ emission feature was detected from our post-maximum NIR spectra, placing a hydrogen mass upper limit of 0.1 $M_{\odot}$. The overall optical/NIR photometric and NIR spectral evolution of SN 2017cbv is similar to that of a normal SN~Ia, even though its early evolution is marked by a flux excess no seen in most other well-observed normal SNe~Ia. We also compare the exquisite light curves of SN 2017cbv with some $M_{ch}$ DDT models and sub-$M_{ch}$ double detonation models., Comment: 20 figures, Accepted for publication in The Astrophysical Journal

We present high-cadence ultraviolet (UV), optical, and near-infrared (NIR) data on the luminous Type II-P supernova SN 2017gmr from hours after discovery through the first 180 days. SN 2017gmr does not show signs of narrow, high-ionization emission lines in the early optical spectra, yet the optical lightcurve evolution suggests that an extra energy source from circumstellar medium (CSM) interaction must be present for at least 2 days after explosion. Modeling of the early lightcurve indicates a ~500R$_{\odot}$ progenitor radius, consistent with a rather compact red supergiant, and late-time luminosities indicate up to 0.130 $\pm$ 0.026 M$_{\odot}$ of $^{56}$Ni are present, if the lightcurve is solely powered by radioactive decay, although the $^{56}$Ni mass may be lower if CSM interaction contributes to the post-plateau luminosity. Prominent multi-peaked emission lines of H$\alpha$ and [O I] emerge after day 154, as a result of either an asymmetric explosion or asymmetries in the CSM. The lack of narrow lines within the first two days of explosion in the likely presence of CSM interaction may be an example of close, dense, asymmetric CSM that is quickly enveloped by the spherical supernova ejecta., Comment: Submitted to ApJ, 25 pages, plus Appendix