This article proposes a document-level prompt learning approach using LLMs to extract the timeline-based storyline. Through verification tests on datasets such as ESCv1.2 and Timeline17, the results show that the prom...This article proposes a document-level prompt learning approach using LLMs to extract the timeline-based storyline. Through verification tests on datasets such as ESCv1.2 and Timeline17, the results show that the prompt + one-shot learning proposed in this article works well. Meanwhile, our research findings indicate that although timeline-based storyline extraction has shown promising prospects in the practical applications of LLMs, it is still a complex natural language processing task that requires further research.展开更多
Social media like Twitter who serves as a novel news medium and has become increasingly popular since its establishment. Large scale first-hand user-generated tweets motivate automatic event detection on Twitter. Prev...Social media like Twitter who serves as a novel news medium and has become increasingly popular since its establishment. Large scale first-hand user-generated tweets motivate automatic event detection on Twitter. Previous unsupervised approaches detected events by clustering words. These methods detect events using burstiness,which measures surging frequencies of words at certain time windows. However,event clusters represented by a set of individual words are difficult to understand. This issue is addressed by building a document-level event detection model that directly calculates the burstiness of tweets,leveraging distributed word representations for modeling semantic information,thereby avoiding sparsity. Results show that the document-level model not only offers event summaries that are directly human-readable,but also gives significantly improved accuracies compared to previous methods on unsupervised tweet event detection,which are based on words/segments.展开更多
Evidential Document-level Event Factuality Identification(EvDEFI)aims to predict the factual nature of an event and extract evidential sentences from the document precisely.Previous work usually limited to only predic...Evidential Document-level Event Factuality Identification(EvDEFI)aims to predict the factual nature of an event and extract evidential sentences from the document precisely.Previous work usually limited to only predicting the factuality of an event with respect to a document,and neglected the interpretability of the task.As a more fine-grained and interpretable task,EvDEFI is still in the early stage.The existing model only used shallow similarity calculation to extract evidences,and employed simple attentions without lexical features,which is quite coarse-grained.Therefore,we propose a novel EvDEFI model named Heterogeneous and Extractive Graph Attention Network(HEGAT),which can update representations of events and sentences by multi-view graph attentions based on tokens and various lexical features from both local and global levels.Experiments on EB-DEF-v2 corpus demonstrate that HEGAT model is superior to several competitive baselines and can validate the interpretability of the task.展开更多
This paper focuses on document-level event factuality identification (DEFI), which predicts the factual nature of an event from the view of a document. As the document-level sub-task of event factuality identification...This paper focuses on document-level event factuality identification (DEFI), which predicts the factual nature of an event from the view of a document. As the document-level sub-task of event factuality identification (EFI), DEFI is a challenging and fundamental task in natural language processing (NLP). Currently, most existing studies focus on sentence-level event factuality identification (SEFI). However, DEFI is still in the early stage and related studies are quite limited. Previous work is heavily dependent on various NLP tools and annotated information, e.g., dependency trees, event triggers, speculative and negative cues, and does not consider filtering irrelevant and noisy texts that can lead to wrong results. To address these issues, this paper proposes a reinforced multi-granularity hierarchical network model: Reinforced Semantic Learning Network (RSLN), which means it can learn semantics from sentences and tokens at various levels of granularity and hierarchy. Since integrated with hierarchical reinforcement learning (HRL), the RSLN model is able to select relevant and meaningful sentences and tokens. Then, RSLN encodes the event and document according to these selected texts. To evaluate our model, based on the DLEF (Document-Level Event Factuality) corpus, we annotate the ExDLEF corpus as the benchmark dataset. Experimental results show that the RSLN model outperforms several state-of-the-arts.展开更多
Document-level machine translation(MT)remains challenging due to its difficulty in efficiently using documentlevel global context for translation.In this paper,we propose a hierarchical model to learn the global conte...Document-level machine translation(MT)remains challenging due to its difficulty in efficiently using documentlevel global context for translation.In this paper,we propose a hierarchical model to learn the global context for documentlevel neural machine translation(NMT).This is done through a sentence encoder to capture intra-sentence dependencies and a document encoder to model document-level inter-sentence consistency and coherence.With this hierarchical architecture,we feedback the extracted document-level global context to each word in a top-down fashion to distinguish different translations of a word according to its specific surrounding context.Notably,we explore the effect of three popular attention functions during the information backward-distribution phase to take a deep look into the global context information distribution of our model.In addition,since large-scale in-domain document-level parallel corpora are usually unavailable,we use a two-step training strategy to take advantage of a large-scale corpus with out-of-domain parallel sentence pairs and a small-scale corpus with in-domain parallel document pairs to achieve the domain adaptability.Experimental results of our model on Chinese-English and English-German corpora significantly improve the Transformer baseline by 4.5 BLEU points on average which demonstrates the effectiveness of our proposed hierarchical model in document-level NMT.展开更多
Relation Extraction(RE)is to obtain a predefined relation type of two entities mentioned in a piece of text,e.g.,a sentence-level or a document-level text.Most existing studies suffer from the noise in the text,and ne...Relation Extraction(RE)is to obtain a predefined relation type of two entities mentioned in a piece of text,e.g.,a sentence-level or a document-level text.Most existing studies suffer from the noise in the text,and necessary pruning is of great importance.The conventional sentence-level RE task addresses this issue by a denoising method using the shortest dependency path to build a long-range semantic dependency between entity pairs.However,this kind of denoising method is scarce in document-level RE.In this work,we explicitly model a denoised document-level graph based on linguistic knowledge to capture various long-range semantic dependencies among entities.We first formalize a Syntactic Dependency Tree forest(SDT-forest)by introducing the syntax and discourse dependency relation.Then,the Steiner tree algorithm extracts a mention-level denoised graph,Steiner Graph(SG),removing linguistically irrelevant words from the SDT-forest.We then devise a slide residual attention to highlight word-level evidence on text and SG.Finally,the classification is established on the SG to infer the relations of entity pairs.We conduct extensive experiments on three public datasets.The results evidence that our method is beneficial to establish long-range semantic dependency and can improve the classification performance with longer texts.展开更多
文摘This article proposes a document-level prompt learning approach using LLMs to extract the timeline-based storyline. Through verification tests on datasets such as ESCv1.2 and Timeline17, the results show that the prompt + one-shot learning proposed in this article works well. Meanwhile, our research findings indicate that although timeline-based storyline extraction has shown promising prospects in the practical applications of LLMs, it is still a complex natural language processing task that requires further research.
基金Supported by the National High Technology Research and Development Programme of China(No.2015AA015405)
文摘Social media like Twitter who serves as a novel news medium and has become increasingly popular since its establishment. Large scale first-hand user-generated tweets motivate automatic event detection on Twitter. Previous unsupervised approaches detected events by clustering words. These methods detect events using burstiness,which measures surging frequencies of words at certain time windows. However,event clusters represented by a set of individual words are difficult to understand. This issue is addressed by building a document-level event detection model that directly calculates the burstiness of tweets,leveraging distributed word representations for modeling semantic information,thereby avoiding sparsity. Results show that the document-level model not only offers event summaries that are directly human-readable,but also gives significantly improved accuracies compared to previous methods on unsupervised tweet event detection,which are based on words/segments.
基金supported by the National Natural Science Foundation of China(NSFC)(Grant Nos.62006167 and 62276177)the Priority Academic Program Development of Jiangsu Higher Education Institutions(PAPD).
文摘Evidential Document-level Event Factuality Identification(EvDEFI)aims to predict the factual nature of an event and extract evidential sentences from the document precisely.Previous work usually limited to only predicting the factuality of an event with respect to a document,and neglected the interpretability of the task.As a more fine-grained and interpretable task,EvDEFI is still in the early stage.The existing model only used shallow similarity calculation to extract evidences,and employed simple attentions without lexical features,which is quite coarse-grained.Therefore,we propose a novel EvDEFI model named Heterogeneous and Extractive Graph Attention Network(HEGAT),which can update representations of events and sentences by multi-view graph attentions based on tokens and various lexical features from both local and global levels.Experiments on EB-DEF-v2 corpus demonstrate that HEGAT model is superior to several competitive baselines and can validate the interpretability of the task.
基金supported by the National Natural Science Foundation of China under Grant Nos.62006167,62276177,62376181,and 62376178the Natural Science Foundation of the Jiangsu Higher Education Institutions of China under Grant No.24KJB520036the Project Funded by the Priority Academic Program Development(PAPD)of Jiangsu Higher Education Institutions.
文摘This paper focuses on document-level event factuality identification (DEFI), which predicts the factual nature of an event from the view of a document. As the document-level sub-task of event factuality identification (EFI), DEFI is a challenging and fundamental task in natural language processing (NLP). Currently, most existing studies focus on sentence-level event factuality identification (SEFI). However, DEFI is still in the early stage and related studies are quite limited. Previous work is heavily dependent on various NLP tools and annotated information, e.g., dependency trees, event triggers, speculative and negative cues, and does not consider filtering irrelevant and noisy texts that can lead to wrong results. To address these issues, this paper proposes a reinforced multi-granularity hierarchical network model: Reinforced Semantic Learning Network (RSLN), which means it can learn semantics from sentences and tokens at various levels of granularity and hierarchy. Since integrated with hierarchical reinforcement learning (HRL), the RSLN model is able to select relevant and meaningful sentences and tokens. Then, RSLN encodes the event and document according to these selected texts. To evaluate our model, based on the DLEF (Document-Level Event Factuality) corpus, we annotate the ExDLEF corpus as the benchmark dataset. Experimental results show that the RSLN model outperforms several state-of-the-arts.
基金supported by the National Natural Science Foundation of China under Grant Nos.61751206,61673290 and 61876118the Postgraduate Research&Practice Innovation Program of Jiangsu Province of China under Grant No.KYCX20_2669a project funded by the Priority Academic Program Development of Jiangsu Higher Education Institutions(PAPD).
文摘Document-level machine translation(MT)remains challenging due to its difficulty in efficiently using documentlevel global context for translation.In this paper,we propose a hierarchical model to learn the global context for documentlevel neural machine translation(NMT).This is done through a sentence encoder to capture intra-sentence dependencies and a document encoder to model document-level inter-sentence consistency and coherence.With this hierarchical architecture,we feedback the extracted document-level global context to each word in a top-down fashion to distinguish different translations of a word according to its specific surrounding context.Notably,we explore the effect of three popular attention functions during the information backward-distribution phase to take a deep look into the global context information distribution of our model.In addition,since large-scale in-domain document-level parallel corpora are usually unavailable,we use a two-step training strategy to take advantage of a large-scale corpus with out-of-domain parallel sentence pairs and a small-scale corpus with in-domain parallel document pairs to achieve the domain adaptability.Experimental results of our model on Chinese-English and English-German corpora significantly improve the Transformer baseline by 4.5 BLEU points on average which demonstrates the effectiveness of our proposed hierarchical model in document-level NMT.
基金supported by the National Natural Science Foundation of China(Nos.U19A2059&62176046).
文摘Relation Extraction(RE)is to obtain a predefined relation type of two entities mentioned in a piece of text,e.g.,a sentence-level or a document-level text.Most existing studies suffer from the noise in the text,and necessary pruning is of great importance.The conventional sentence-level RE task addresses this issue by a denoising method using the shortest dependency path to build a long-range semantic dependency between entity pairs.However,this kind of denoising method is scarce in document-level RE.In this work,we explicitly model a denoised document-level graph based on linguistic knowledge to capture various long-range semantic dependencies among entities.We first formalize a Syntactic Dependency Tree forest(SDT-forest)by introducing the syntax and discourse dependency relation.Then,the Steiner tree algorithm extracts a mention-level denoised graph,Steiner Graph(SG),removing linguistically irrelevant words from the SDT-forest.We then devise a slide residual attention to highlight word-level evidence on text and SG.Finally,the classification is established on the SG to infer the relations of entity pairs.We conduct extensive experiments on three public datasets.The results evidence that our method is beneficial to establish long-range semantic dependency and can improve the classification performance with longer texts.
