Discourse Understanding and Factual Consistency in Abstractive Summarization

Hits-based attentional neural model for abstractive summarization

Knowledge-Based Systems ◽

10.1016/j.knosys.2021.106996 ◽

2021 ◽

pp. 106996

Author(s):

Xiaoyan Cai ◽

Kaile Shi ◽

Yuehan Jiang ◽

Libin Yang ◽

Sen Liu

Keyword(s):

Neural Model ◽

Abstractive Summarization

Download Full-text

Improving Abstractive Summarization Based on Dynamic Residual Network with Reinforce Dependency

Neurocomputing ◽

10.1016/j.neucom.2021.02.028 ◽

2021 ◽

Author(s):

Weizhi Liao ◽

Yaheng Ma ◽

Yanchao Yin ◽

Guanglei Ye ◽

Dongzhou Zuo

Keyword(s):

Residual Network ◽

Abstractive Summarization

Download Full-text

Discursive work in resisting stereotypic representations of the Chinese among Chinese students

Applied Linguistics Review ◽

10.1515/applirev-2019-0109 ◽

2020 ◽

Vol 0 (0) ◽

Author(s):

Hui Zanne Seng ◽

Mei Yuit Chan ◽

Ngee Thai Yap

Keyword(s):

Social Group ◽

Chinese Students ◽

Identity Work ◽

Negative Effects ◽

Fragmented Self ◽

Group Identities ◽

Discourse Strategies ◽

Discourse Understanding ◽

Insight Into ◽

Discursive Resources

AbstractThe negative effects of stereotyping arising from a victim’s acceptance and internalisation of stereotype identities are well-known. As stereotypes are created and maintained in discourse, understanding how targets of stereotyping employ discursive resources to resist the constraining structures of stereotypic identities imposed upon them can provide insight into the process of stereotyping and contribute to efforts to reduce the threat of stereotyping. We examined the strategies used by targets of stereotyping in contesting stereotypic representations of their social group through the mobilisation of a range of discourse strategies when presented with stereotyping attacks on the group. The findings revealed that stereotypes are subtle in nature and may not be easily recognised and hence, difficult to resist. Participants employed a number of discourse strategies to repair their fragmented self and group identities. However, in their attempt to maintain identity coherence, they ended up using stereotyping discourses themselves to devalue the perceived outgroup as well as subgroups they created within their own social group. The study highlights the complexity of stereotyping and its self-perpetuating character, and sheds light on the difficulty faced by targets of stereotyping discourse in reconciling their identities through intense discursive and identity work.

Download Full-text

ASoVS: Abstractive Summarization of Video Sequences

IEEE Access ◽

10.1109/access.2019.2902507 ◽

2019 ◽

Vol 7 ◽

pp. 29253-29263 ◽

Cited By ~ 1

Author(s):

Aniqa Dilawari ◽

Muhammad Usman Ghani Khan

Keyword(s):

Video Sequences ◽

Abstractive Summarization

Download Full-text

Towards a Broader Value Discourse: Understanding Sustainable and Public Value Potential

Journal of Creating Value ◽

10.1177/2394964317725449 ◽

2017 ◽

Vol 3 (2) ◽

pp. 193-199 ◽

Cited By ~ 3

Author(s):

Joona Keränen

Keyword(s):

Economic Benefits ◽

Business And Society ◽

Public Value ◽

Current Value ◽

Discourse Understanding ◽

Benefits And Costs

The current value discourse tends to be characterized by economic benefits and costs. This may resonate with business actors, but customers and society are increasingly interested in environmental, social and public value as well. This article discusses why and how practitioners and scholars should take sustainable and public value potential into account, and move towards a broader value discourse that would resonate with both business and society.

Download Full-text

SGAN4AbSum: A Semantic-Enhanced Generative Adversarial Network for Abstractive Text Summarization

10.21203/rs.3.rs-648146/v1 ◽

2021 ◽

Author(s):

Tham Vo

Keyword(s):

Ground Truth ◽

Text Summarization ◽

Generative Adversarial Network ◽

Convolutional Network ◽

Training Strategy ◽

Adversarial Network ◽

Deep Recurrent Neural Network ◽

Benchmark Datasets ◽

Latent Representations ◽

Abstractive Summarization

Abstract In abstractive summarization task, most of proposed models adopt the deep recurrent neural network (RNN)-based encoder-decoder architecture to learn and generate meaningful summary for a given input document. However, most of recent RNN-based models always suffer the challenges related to the involvement of much capturing high-frequency/reparative phrases in long documents during the training process which leads to the outcome of trivial and generic summaries are generated. Moreover, the lack of thorough analysis on the sequential and long-range dependency relationships between words within different contexts while learning the textual representation also make the generated summaries unnatural and incoherent. To deal with these challenges, in this paper we proposed a novel semantic-enhanced generative adversarial network (GAN)-based approach for abstractive text summarization task, called as: SGAN4AbSum. We use an adversarial training strategy for our text summarization model in which train the generator and discriminator to simultaneously handle the summary generation and distinguishing the generated summary with the ground-truth one. The input of generator is the jointed rich-semantic and global structural latent representations of training documents which are achieved by applying a combined BERT and graph convolutional network (GCN) textual embedding mechanism. Extensive experiments in benchmark datasets demonstrate the effectiveness of our proposed SGAN4AbSum which achieve the competitive ROUGE-based scores in comparing with state-of-the-art abstractive text summarization baselines.

Download Full-text