Inter-sentence and Implicit Causality Extraction from Chinese Corpus

Jin, Xianxian; Wang, Xinzhi; Luo, Xiangfeng; Huang, Shucai; Gu, Shengwei

doi:10.1007/978-3-030-47426-3_57

Cited by 19 publications

(13 citation statements)

References 9 publications

(16 reference statements)

Supporting

Mentioning

Contrasting

Order By: Relevance

“…On analyzing the strength of the association between the observed disease burden and obesity-related coverage, we found that cardiovascular diseases were the most frequent disease type. In comparison, metabolic dysfunction and inflammation caused by obesity are ignored in the news, despite scientific evidence indicating that metabolic dysfunction caused by obesity contributes to a wide variety of disorders and effects on the nervous system [ 33 ].…”

Section: Discussionmentioning

confidence: 99%

See 1 more Smart Citation

Obesity-Related Communication in Digital Chinese News From Mainland China, Hong Kong, and Taiwan: Automated Content Analysis

Chang¹,

Schulz²,

Jiao³

et al. 2021

JMIR Public Health Surveill

View full text Add to dashboard Cite

Background The fact that the number of individuals with obesity has increased worldwide calls into question media efforts for informing the public. This study attempts to determine the ways in which the mainstream digital news covers the etiology of obesity and diseases associated with the burden of obesity. Objective The dual objectives of this study are to obtain an understanding of what the news reports on obesity and to explore meaning in data by extending the preconceived grounded theory. Methods The 10 years of news text from 2010 to 2019 compared the development of obesity-related coverage and its potential impact on its perception in Mainland China, Hong Kong, and Taiwan. Digital news stories on obesity along with affliction and inferences in 9 Chinese mainstream newspapers were sampled. An automatic content analysis tool, DiVoMiner was proposed. This computer-aided platform is designed to organize and filter large sets of data on the basis of the patterns of word occurrence and term discovery. Another programming language, Python 3, was used to explore connections and patterns created by the aggregated interactions. Results A total of 30,968 news stories were identified with increasing attention since 2016. The highest intensity of newspaper coverage of obesity communication was observed in Taiwan. Overall, a stronger focus on 2 shared causative attributes of obesity is on stress (n=4483, 33.0%) and tobacco use (n=3148, 23.2%). The burdens of obesity and cardiovascular diseases are implied to be the most, despite the aggregated interaction of edge centrality showing the highest link between the “cancer” and obesity. This study goes beyond traditional journalism studies by extending the framework of computational and customizable web-based text analysis. This could set a norm for researchers and practitioners who work on data projects largely for an innovative attempt. Conclusions Similar to previous studies, the discourse between the obesity epidemic and personal afflictions is the most emphasized approach. Our study also indicates that the inclination of blaming personal attributes for health afflictions potentially limits social and governmental responsibility for addressing this issue.

show abstract

Section: Discussionmentioning

confidence: 99%

“…Causality is one of the discourse analyses performed by extracting semantic relationships between cause and consequence for learning and predicting the structure of a sentence [ 33 ]. Analyzing causality is a common practice for studying the data corpus of machine learning [ 6 , 26 ].…”

Section: Methodsmentioning

confidence: 99%

Obesity-Related Communication in Digital Chinese News From Mainland China, Hong Kong, and Taiwan: Automated Content Analysis

Chang¹,

Schulz²,

Jiao³

et al. 2021

JMIR Public Health Surveill

View full text Add to dashboard Cite

show abstract

“…• CSNN 15 uses CNN and self-attention to capture features relationship, and the higher-level phrase representations are feed into BiLSTM and CRF layer for CEE.…”

Section: Baseline Methodsmentioning

confidence: 99%

“…Besides, Dasgupta and Dunietz et al 12,13 apply long short-term memory (LSTM) to CEE to capture the long-term dependence of sentences. Furthermore, Ma and Jin et al 14,15 combine CNN and LSTM to take advantage of both architectures. Though neural networks have achieved state-of-the-art performance, CEE is still remains a hard problem due to its complexity and ambiguity.…”

Section: Implicit Causal Sentencementioning

confidence: 99%

Chinese causal event extraction using causality‐associated graph neural network

Gao

Luo

Wang

2021

Concurrency and Computation

Self Cite

View full text Add to dashboard Cite

Causal event extraction (CEE) aims to identify and extract cause-effect event pairs from texts, which is a fundamental task in natural language processing. Recent research treat CEE as a sequence labeling problem. However, the linguistic complexity and ambiguity of textual description results in the low accuracy of extractors. To address the above issues, considering the prior knowledge like the causal network constructed based on the causal indicators, which can represent information transition between cause and effect, may helpful for CEE. In this article, we propose causality-associated graph neural network to incorporate in-domain knowledge by taking important causal words into account. External causal knowledge is modeled as causal associated graph (CAG). Then we use graph neural networks (GNN) to capture the complex relationship of intraevent mentions and interevent causality in a sentence based on the relationship obtained from CAG. Finally, sentence sequence and prior causal knowledge of GNN embedding are fed into multiscaled convolution and bidirectional long short-term memory networks. Experimental results on two datasets show that our method outperforms the state-of-the-art baseline.

show abstract

“…Recently, neural-based extraction methods have become a majority of related tasks in NLP, relation classification [18], relation extraction [14] and sequence tagging [10]. Among these methods, pre-trained language models [4] (e.g., BERT) dominate the state-of-the-art results on a wide range of NLP tasks.…”

Section: Introductionmentioning

confidence: 99%

Back to Prior Knowledge: Joint Event Causality Extraction via Convolutional Semantic Infusion

Wang

Luo

et al. 2021

Preprint

Self Cite

View full text Add to dashboard Cite

Joint event and causality extraction is a challenging yet essential task in information retrieval and data mining. Recently, pretrained language models (e.g., BERT) yield state-of-the-art results and dominate in a variety of NLP tasks. However, these models are incapable of imposing external knowledge in domain-specific extraction. Considering the prior knowledge of frequent n-grams that represent cause/effect events may benefit both event and causality extraction, in this paper, we propose convolutional knowledge infusion for frequent n-grams with different windows of length within a joint extraction framework. Knowledge infusion during convolutional filter initialization not only helps the model capture both intra-event (i.e., features in an event cluster) and interevent (i.e., associations across event clusters) features but also boosts training convergence. Experimental results on the benchmark datasets show that our model significantly outperforms the strong BERT+CSNN baseline.

show abstract

Inter-sentence and Implicit Causality Extraction from Chinese Corpus

Cited by 19 publications

References 9 publications

Obesity-Related Communication in Digital Chinese News From Mainland China, Hong Kong, and Taiwan: Automated Content Analysis

Obesity-Related Communication in Digital Chinese News From Mainland China, Hong Kong, and Taiwan: Automated Content Analysis

Chinese causal event extraction using causality‐associated graph neural network

Back to Prior Knowledge: Joint Event Causality Extraction via Convolutional Semantic Infusion

Contact Info

Product

Resources

About