Self-Attentive Sentimental Sentence Embedding for Sentiment Analysis

Lin, Sheng-Chieh; Su, Wen-Yuh; Chien, Po-Chuan; Tsai, Ming-Feng; Wang, Chuan‐Ju

doi:10.1109/icassp40776.2020.9054274

Cited by 50 publications

(55 citation statements)

References 9 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…In our work, we aim to learn such influence changes (dynamic‐scale) at the sub‐graph level, that is, we assume that the sub‐cascade graph's influence decays as the interval index increases. Inspired by self‐attention mechanism, 53 we employ a neural function to learn the influence attention. First, we represent the time‐interval as a one‐hot vector

t^{j} \in R^{l}

, and then map

t^{j}

λ_{j}

through a fully‐connected layer with sigmoid function.…”

Section: Methodsmentioning

confidence: 99%

Multi‐scale graph capsule with influence attention for information cascades prediction

Chen

Zhang

et al. 2021

Int J of Intelligent Sys

View full text Add to dashboard Cite

Information cascade size prediction is one of the primary challenges for understanding the diffusion of information. Traditional feature‐based methods heavily rely on the quality of handcrafted features, requiring extensive domain knowledge and hard to generalize to new domains. Recently, inspired by the success of deep learning in computer vision and natural language processing, researchers have developed neural network‐based approaches for tackling this problem. However, existing deep learning‐based methods either focused on modeling the temporal characteristics of cascades but ignored the structural information or failed to take the order‐scale and position‐scale into consideration in modeling structures of information propagation. This paper proposed a novel graph neural network‐based model, called MUCas, to learn the latent representations of cascade graphs from a multi‐scale perspective, which can make full use of the direction‐scale, high‐order‐scale, position‐scale, and dynamic‐scale of cascades via a newly designed MUlti‐scale Graph Capsule Network (MUG‐Caps) and the influence‐attention mechanism. Extensive experiments conducted on two real‐world data sets demonstrate that our MUCas significantly outperforms the state‐of‐the‐art approaches.

show abstract

t^{j} \in R^{l}

, and then map

t^{j}

λ_{j}

through a fully‐connected layer with sigmoid function.…”

Section: Methodsmentioning

confidence: 99%

Multi‐scale graph capsule with influence attention for information cascades prediction

Chen

Zhang

et al. 2021

Int J of Intelligent Sys

View full text Add to dashboard Cite

show abstract

“…The comparison between the general sentiment and the market sentiment will also be discussed (Loughran and McDonald, 2011;Chen et al, 2020b). The lexicons for the sentiment analysis (Bodnaruk et al, 2015;Li and Shah, 2017;Sedinkina et al, 2019) in financial documents and the applications of adopting sentiment analysis results (Bollen et al, 2011;Du et al, 2019;Lin et al, 2020) will be included. This session also covers the sentiment analysis of financial narratives from different resources, including formal documents such as financial statements and professional analyst's reports and informal documents such as blogs and social media platforms.…”

Section: Coarse-grained Financial Opinion Miningmentioning

confidence: 99%

Financial Opinion Mining

Chen¹,

Huang²,

Chen³

2021

Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing: Tutorial Abstracts

View full text Add to dashboard Cite

show abstract

“…Self-attention is an attention mechanism that the attention weights are generated by a single sequence itself to compute the sequence representation. It is also known as intra-attention which is generally adopted in RNN structures [5,12,13,17]. Recently, a new self-attention [26] structure has been proposed and has achieved remarkable results in a wide range of tasks [33,35,36].…”

Section: Self-attentionmentioning

confidence: 99%

“…For distinguishing these two self-attention methods in this paper, selfattention refers to the self-attention method with multi-head attention and residual structure, which is proposed in [26]. While intra-attention refers to the self-attention method which is adopted in RNN structures [5,12,13,17].…”

Section: Self-attentionmentioning

confidence: 99%

“…The interaction mechanism calculates question-aware residuals which are then added to the representation of words in answers, which can make the words in answers to take question information into account. 1 The self-attention methods used in Transformer [26] and RNN based models [5,12,13,17] are different. In the context of this paper, self-attention refers to the self-attention mechanism with multi-head attention and residual structure, which is proposed in [26].…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Gated Group Self-Attention for Answer Selection

Xu,

Ji,

Huang

et al. 2019

Preprint

View full text Add to dashboard Cite

Answer selection (answer ranking) is one of the key steps in many kinds of question answering (QA) applications, where deep models have achieved state-of-the-art performance. Among these deep models, recurrent neural network (RNN) based models are most popular, typically with better performance than convolutional neural network (CNN) based models. Nevertheless, it is difficult for RNN based models to capture the information about long-range dependency among words in the sentences of questions and answers. In this paper, we propose a new deep model, called gated group selfattention (GGSA), for answer selection. GGSA is inspired by global self-attention which is originally proposed for machine translation and has not been explored in answer selection. GGSA tackles the problem of global self-attention that local and global information cannot be well distinguished. Furthermore, an interaction mechanism between questions and answers is also proposed to enhance GGSA by a residual structure. Experimental results on two popular QA datasets show that GGSA can outperform existing answer selection models to achieve state-of-the-art performance. Furthermore, GGSA can also achieve higher accuracy than global self-attention for the answer selection task, with a lower computation cost.

show abstract

Self-Attentive Sentimental Sentence Embedding for Sentiment Analysis

Cited by 50 publications

References 9 publications

Multi‐scale graph capsule with influence attention for information cascades prediction

Multi‐scale graph capsule with influence attention for information cascades prediction

Financial Opinion Mining

Gated Group Self-Attention for Answer Selection

Contact Info

Product

Resources

About