Frame Semantic-Enhanced Sentence Modeling for Sentence-level Extractive Text Summarization
31778Authors
nan
Abstract
AbstractSentence-level extractive text summarization aims to select important sentences from a given document. However, it is very challenging to model the importance of sentences. In this paper, we propose a novel Frame Semantic-Enhanced Sentence Modeling for Extractive Summarization, which leverages Frame semantics to model sentences from both intra-sentence level and inter-sentence level, facilitating the text summarization task. In particular, intra-sentence level semantics leverage Frames and Frame Elements to model internal semantic structure within a sentence, while inter-sentence level semantics leverage Frame-to-Frame relations to model relationships among sentences. Extensive experiments on two benchmark corpus CNN/DM and NYT demonstrate that our model outperforms six state-of-the-art methods significantly.
Keywords
Full Content
Proceedings of the 2021 Conference on Empirical Methods in Natural Language Processing, pages 4045–4052 November 7–11, 2021. c⃝2021 Association for Computational Linguistics 4045 Sentence 1: As a new parent , one of the most trying parts of caring for a baby is letting them to go to sleep. Sentence 2: You can say goodbye to sleepless nights thanks to one father who find a trick that can bring your baby to sleep in less than one minute. F:Preventing_or_letting F:Causation F:Becoming_aware F:Taking_time F:Statement F:Kinship F:Kinship F:Kinship F:Taking_time F:Causation (a) An example adapted from the CNN/Daily dataset. a trick that can bring your baby to sleep in less than one minute. F:Causation FE:Actor FE: Effect FE: Time (b) A Frame annotations. F-to-F Sentence 8: A father has made his baby fall asleep in 42 seconds. 4046 4047 SEP CLS CLS Graph Module Prediction Cd sen1 sen2 sen3 sen4 senn Cd Cg 1 0 0 1 0 Frame Element Node V fe Frame Node V f Sentence Node V sen Encoders BERT you can Document Encoding Frame Encoding Ecls Ecls Esep a cause causes SEP CLS Ecls Esep Document Input You can say goodbye to sleepless nights . A father has made his baby fall. a E1 E1 E1 E2 Definition of Frame Causat- ion: A cause causes an effect. Alternatively, an actor, a participant ef Intra Att Inter Att y1 y2 y3 y4 yn 4048 4049 4050 4051 4052
Labeled Text
AbstractSentence-level extractive text summarization aims to select important sentences from a given document. However, it is very challenging to model the importance of sentences. In this paper, we propose a novel Frame Semantic-Enhanced Sentence Modeling for Extractive Summarization, which leverages Frame semantics to model sentences from both intra-sentence level and inter-sentence level, facilitating the text summarization task. In particular, intra-sentence level semantics leverage Frames and Frame Elements to model internal semantic structure within a sentence, while inter-sentence level semantics leverage Frame-to-Frame relations to model relationships among sentences. Extensive experiments on two benchmark corpus CNN/DM and NYT demonstrate that our model outperforms six state-of-the-art methods significantly.
Publication Details
- Paper ID:
- 31778
- Year:
- 2021
- Publication:
- TIPSTER
- Publication Type:
- conf
- DOI:
- nan