Cross-lingual RST Discourse Parsing

Braud, Chloé; Coavoux, Maximin; Søgaard, Anders

doi:10.18653/v1/e17-1028

Cited by 67 publications

(100 citation statements)

References 34 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…LLC16 is a CKY chart parser with a hierarchical neural network model (attention-based hierarchical bi-LSTM) (Li et al, 2016). BCS17 mono, BCS17 cross+dev are two variants of a transition-based parser that uses a feed-forward neural network model (Braud et al, 2017). JE14 DPLP is a shift-reduce parser that uses an SVM model (Ji and Eisenstein, 2014).…”

Section: A Sample Of Rst Discourse Parsersmentioning

confidence: 99%

How much progress have we made on RST discourse parsing? A replication study of recent results on the RST-DT

Morey¹,

Muller²,

Asher³

2017

Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

This article evaluates purported progress over the past years in RST discourse parsing. Several studies report a relative error reduction of 24 to 51% on all metrics that authors attribute to the introduction of distributed representations of discourse units. We replicate the standard evaluation of 9 parsers, 5 of which use distributed representations, from 8 studies published between 2013 and 2017, using their predictions on the test set of the RST-DT. Our main finding is that most recently reported increases in RST discourse parser performance are an artefact of differences in implementations of the evaluation procedure. We evaluate all these parsers with the standard Parseval procedure to provide a more accurate picture of the actual RST discourse parsers performance in standard evaluation settings. Under this more stringent procedure, the gains attributable to distributed representations represent at most a 16% relative error reduction on fully-labelled structures.

show abstract

Section: A Sample Of Rst Discourse Parsersmentioning

confidence: 99%

How much progress have we made on RST discourse parsing? A replication study of recent results on the RST-DT

Morey¹,

Muller²,

Asher³

2017

Proceedings of the 2017 Conference on Empirical Methods in Natural Language Processing

View full text Add to dashboard Cite

show abstract

“…PDTB focuses on shallow discourse relations but ignores the overall discourse structure (Yang and Li 2018), while in this paper we aim to parse discourse structures. As for RST, there have been many approaches including transition-based methods (Braud, Coavoux, and Søgaard 2017;Wang, Li, and Wang 2017;Yu, Zhang, and Fu 2018) and those involving CYK-like algorithms (Joty, Carenini, and Ng 2015;Li, Li, and Chang 2016;Liu and Lapata 2017) or greedy bottom-up algorithms (Feng and Hirst 2014). However, constituency-based RST does not allow non-adjacent relations, which makes it inapplicable for multi-party dialogues.…”

Section: Related Workmentioning

confidence: 99%

“…Moreover, state-of-the-art approaches for discourse dependency parsing as mentioned above still rely on handcrafted features or external parsers. Neural networks have recently been widely applied in various NLP tasks, including RST discourse parsing (Li, Li, and Chang 2016;Braud, Coavoux, and Søgaard 2017) and dialogue act recognition (Kumar et al 2018;Chen et al 2018). And (Jia et al 2018a;2018b) also applied neural networks in their transition-based dependency parsing models.…”

Section: Related Workmentioning

confidence: 99%

A Deep Sequential Model for Discourse Parsing on Multi-Party Dialogues

Shi

Huang

2019

AAAI

View full text Add to dashboard Cite

Discourse structures are beneficial for various NLP tasks such as dialogue understanding, question answering, sentiment analysis, and so on. This paper presents a deep sequential model for parsing discourse dependency structures of multi-party dialogues. The proposed model aims to construct a discourse dependency tree by predicting dependency relations and constructing the discourse structure jointly and alternately. It makes a sequential scan of the Elementary Discourse Units (EDUs) 1 in a dialogue. For each EDU, the model decides to which previous EDU the current one should link and what the corresponding relation type is. The predicted link and relation type are then used to build the discourse structure incrementally with a structured encoder. During link prediction and relation classification, the model utilizes not only local information that represents the concerned EDUs, but also global information that encodes the EDU sequence and the discourse structure that is already built at the current step. Experiments show that the proposed model outperforms all the state-of-the-art baselines.

show abstract

“…For Morey et al (2017)'s study, they submitted predicted discourse trees from an updated, unpublished version of their parser. 12 In the cross+dev setting, Braud et al (2017) train their parser on RST discourse treebanks for several languages. other parsers except for Feng and Hirst (2014a)'s graph CRF model.…”

Section: Training and Hyperparametersmentioning

confidence: 99%

Neural Generative Rhetorical Structure Parsing

Mabona¹,

Rimell²,

Clark³

et al. 2019

Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conferen

View full text Add to dashboard Cite

Rhetorical structure trees have been shown to be useful for several document-level tasks including summarization and document classification. Previous approaches to RST parsing have used discriminative models; however, these are less sample efficient than generative models, and RST parsing datasets are typically small. In this paper, we present the first generative model for RST parsing. Our model is a document-level RNN grammar (RNNG) with a bottom-up traversal order. We show that, for our parser's traversal order, previous beam search algorithms for RNNGs have a left-branching bias which is ill-suited for RST parsing. We develop a novel beam search algorithm that keeps track of both structureand word-generating actions without exhibiting this branching bias and results in absolute improvements of 6.8 and 2.9 on unlabelled and labelled F1 over previous algorithms. Overall, our generative model outperforms a discriminative model with the same features by 2.6 F1 points and achieves performance comparable to the state-of-the-art, outperforming all published parsers from a recent replication study that do not use additional training data. ✠ JUSTIFY Acme Inc. has closed several widget factories. | ATTRIBUTION The CEO told investors they were no longer profitable.

show abstract

Cross-lingual RST Discourse Parsing

Cited by 67 publications

References 34 publications

How much progress have we made on RST discourse parsing? A replication study of recent results on the RST-DT

How much progress have we made on RST discourse parsing? A replication study of recent results on the RST-DT

A Deep Sequential Model for Discourse Parsing on Multi-Party Dialogues

Neural Generative Rhetorical Structure Parsing

Contact Info

Product

Resources

About