arXiv subjects

Lanjun Zhou

Publications and source records attributed to Lanjun Zhou.

8 recordsLinked to original sources

arXiv2024

Reply with Sticker: New Dataset and Model for Sticker Retrieval

Using stickers in online chatting is very prevalent on social media platforms, where the stickers used in the conversation can express someone's intention/emotion/attitude in a vivid, tactful, and intuitive way. Existing sticker retrieval research typically retrieves stickers based on context and the current utterance delivered by the user. That is, the stickers serve as a supplement to the current utterance. However, in the real-world scenario, using stickers to express what we want to say rather than as a supplement to our words only is also important. Therefore, in this paper, we create a new dataset for sticker retrieval in conversation, called \textbf{StickerInt}, where stickers are used to reply to previous conversations or supplement our words. Based on the created dataset, we present a simple yet effective framework for sticker retrieval in conversation based on the learning of intention and the cross-modal relationships between conversation context and stickers, coined as \textbf{Int-RA}. Specifically, we first devise a knowledge-enhanced intention predictor to introduce the intention information into the conversation representations. Subsequently, a relation-aware sticker selector is devised to retrieve the response sticker via cross-modal relationships. Extensive experiments on two datasets show that the proposed model achieves state-of-the-art performance and generalization capability in sticker retrieval. The dataset and source code of this work are released at https://github.com/HITSZ-HLT/Int-RA.

cs.MM↗

arXiv2023

DiVa: An Iterative Framework to Harvest More Diverse and Valid Labels from User Comments for Music

Towards sufficient music searching, it is vital to form a complete set of labels for each song. However, current solutions fail to resolve it as they cannot produce diverse enough mappings to make up for the information missed by the gold labels. Based on the observation that such missing information may already be presented in user comments, we propose to study the automated music labeling in an essential but under-explored setting, where the model is required to harvest more diverse and valid labels from the users' comments given limited gold labels. To this end, we design an iterative framework (DiVa) to harvest more $\underline{\text{Di}}$verse and $\underline{\text{Va}}$lid labels from user comments for music. The framework makes a classifier able to form complete sets of labels for songs via pseudo-labels inferred from pre-trained classifiers and a novel joint score function. The experiment on a densely annotated testing set reveals the superiority of the Diva over state-of-the-art solutions in producing more diverse labels missed by the gold labels. We hope our work can inspire future research on automated music labeling.

cs.IR↗

arXiv2022

Generating Tips from Song Reviews: A New Dataset and Framework

Reviews of songs play an important role in online music service platforms. Prior research shows that users can make quicker and more informed decisions when presented with meaningful song reviews. However, reviews of music songs are generally long in length and most of them are non-informative for users. It is difficult for users to efficiently grasp meaningful messages for making decisions. To solve this problem, one practical strategy is to provide tips, i.e., short, concise, empathetic, and self-contained descriptions about songs. Tips are produced from song reviews and should express non-trivial insights about the songs. To the best of our knowledge, no prior studies have explored the tip generation task in music domain. In this paper, we create a dataset named MTips for the task and propose a framework named GENTMS for automatically generating tips from song reviews. The dataset involves 8,003 Chinese tips/non-tips from 128 songs which are distributed in five different song genres. Experimental results show that GENTMS achieves top-10 precision at 85.56%, outperforming the baseline models by at least 3.34%. Besides, to simulate the practical usage of our proposed framework, we also experiment with previously-unseen songs, during which GENTMS also achieves the best performance with top-10 precision at 78.89% on average. The results demonstrate the effectiveness of the proposed framework in tip generation of the music domain.

cs.IR↗

arXiv2022

Empathetic Response Generation with State Management

A good empathetic dialogue system should first track and understand a user's emotion and then reply with an appropriate emotion. However, current approaches to this task either focus on improving the understanding of users' emotion or on proposing better responding strategies, and very few works consider both at the same time. Our work attempts to fill this vacancy. Inspired by task-oriented dialogue systems, we propose a novel empathetic response generation model with emotion-aware dialogue management. The emotion-aware dialogue management contains two parts: (1) Emotion state tracking maintains the current emotion state of the user and (2) Empathetic dialogue policy selection predicts a target emotion and a user's intent based on the results of the emotion state tracking. The predicted information is then used to guide the generation of responses. Experimental results show that dynamically managing different information can help the model generate more empathetic responses compared with several baselines under both automatic and human evaluations.

cs.CL↗

arXiv2017

A closer look at the cosmological implications of the $Λ$HDE model

In a previous paper, we proposed a heterotic dark energy model, called $Λ$HDE, in which dark energy is composed of two components: cosmological constant (CC) and holographic dark energy (HDE). The aim of this work is to give a more comprehensive and systematic investigation on the cosmological implications of the $Λ$HDE model. Firstly, we make use of the current observations to constrain the $Λ$HDE model, and compare its cosmology-fit results with the results of the $Λ$CDM and the HDE model. Then, by combining a qualitative theoretical analysis with a quantitative numerical study, we discuss the impact of considering curvature on the cosmic evolutions of fractional HDE density $Ω_{hde}$ and fractional CC density $Ω_Λ$, as well as on the ultimate cosmic fate. Finally, we explore the effects of adopting different types of observational data. We find that: (1) the current observational data cannot distinguish the $Λ$HDE model from the $Λ$CDM and the HDE model; this indicates that DE may contain multiple components. (2) the asymptotic solution of $Ω_{hde}$ and the corresponding cosmic fate in a flat universe can be extended to the case of a non-flat universe; moreover, compared with the case of a flat universe, considering curvature will make HDE closer to a phantom dark energy. (3) compared with JLA dataset, SNLS3 data more favor a phantom type HDE; in contrast, using other types of observational data have no significant impact on the cosmic evolutions of the $Λ$HDE model.

astro-ph.CO↗

arXiv2016

More Evidence for the Redshift Dependence of Color from the JLA Supernova Sample Using Redshift Tomography

In this work, by applying the redshift tomography method to Joint Light-curve Analysis (JLA) supernova sample, we explore the possible redshift-dependence of stretch-luminosity parameter $α$ and color-luminosity parameter $β$. The basic idea is to divide the JLA sample into different redshift bins, assuming that $α$ and $β$ are piecewise constants. Then, by constraining the $Λ$CDM model, we check the consistency of cosmology-fit results given by the SN sample of each redshift bin. We also adopt the same technique to explore the possible evolution of $β$ in various subsamples of JLA. Using the full JLA data, we find that $α$ is always consistent with a constant. In contrast, at high redshift $β$ has a significant trend of decreasing, at $\sim 3.5σ$ confidence level (CL). Moreover, we find that low-$z$ subsample favors a constant $β$; in contrast, SDSS and SNLS subsamples favor a decreasing $β$ at 2$σ$ and $3.3σ$ CL, respectively. Besides, by using a binned parameterization of $β$, we study the impacts of $β$'s evolution on parameter estimation. We find that compared with a constant $β$, a varying $β$ yields a larger best-fit value of fractional matter density $Ω_{m0}$, which slightly deviates from the best-fit result given by other cosmological observations. However, for both the varying $β$ and the constant $β$ cases, the $1σ$ regions of $Ω_{m0}$ are still consistent with the result given by other observations.

astro-ph.CO↗

arXiv2016

Diagnosing $Λ$HDE model with statefinder hierarchy and fractional growth parameter

Recently, a new dark energy model called $Λ$HDE was proposed. In this model, dark energy consists of two parts: cosmological constant $Λ$ and holographic dark energy (HDE). Two key parameters of this model are the fractional density of cosmological constant $Ω_{\Lambda0}$, and the dimensionless HDE parameter $c$. Since these two parameters determine the dynamical properties of DE and the destiny of universe, it is important to study the impacts of different values of $Ω_{\Lambda0}$ and $c$ on the $Λ$HDE model. In this paper, we apply various DE diagnostic tools to diagnose $Λ$HDE models with different values of $Ω_{\Lambda0}$ and $c$; these tools include statefinder hierarchy \{$S_3^{(1)}, S_4^{(1)}$\}, fractional growth parameter $ε$, and composite null diagnostic (CND), which is a combination of \{$S_3^{(1)}, S_4^{(1)}$\} and $ε$. We find that: (1) adopting different values of $Ω_{\Lambda0}$ only has quantitative impacts on the evolution of the $Λ$HDE model, while adopting different $c$ has qualitative impacts; (2) compared with $S_3^{(1)}$, $S_4^{(1)}$ can give larger differences among the cosmic evolutions of the $Λ$HDE model associated with different $Ω_{\Lambda0}$ or different $c$; (3) compared with the case of using a single diagnostic, adopting a CND pair has much stronger ability to diagnose the $Λ$HDE model.

astro-ph.CO↗

arXiv2012

Replaying neutrino bremsstrahlung with general dispersion relations

It is generally held that neutrinos with superluminal velocity will lose their energy spontaneously by radiating electron-positron pairs, similar to bremsstrahlung process. Recently, this process was closely studied for neutrinos whose energy is roughly proportional to their momentum. Confronted with an increasing amount of superluminal neutrino models, it is urgent to calculate the same process for general dispersion relations. The calculation is performed in this paper, without resorting to any nontrivial frame such as the effective "rest frame".

hep-ph↗