2nd Shared Task on Discourse Relation Parsing and Treebanking (DISRPT 2021) (2021)


up

bib (full) Proceedings of the 2nd Shared Task on Discourse Relation Parsing and Treebanking (DISRPT 2021)

pdf bib
Proceedings of the 2nd Shared Task on Discourse Relation Parsing and Treebanking (DISRPT 2021)
Amir Zeldes | Yang Janet Liu | Mikel Iruskieta | Philippe Muller | Chloé Braud | Sonia Badene

pdf bib
The DISRPT 2021 Shared Task on Elementary Discourse Unit Segmentation, Connective Detection, and Relation Classification
Amir Zeldes | Yang Janet Liu | Mikel Iruskieta | Philippe Muller | Chloé Braud | Sonia Badene

In 2021, we organized the second iteration of a shared task dedicated to the underlying units used in discourse parsing across formalisms: the DISRPT Shared Task (Discourse Relation Parsing and Treebanking). Adding to the 2019 tasks on Elementary Discourse Unit Segmentation and Connective Detection, this iteration of the Shared Task included for the first time a track on discourse relation classification across three formalisms: RST, SDRT, and PDTB. In this paper we review the data included in the Shared Task, which covers nearly 3 million manually annotated tokens from 16 datasets in 11 languages, survey and compare submitted systems and report on system performance on each task for both annotated and plain-tokenized versions of the data.

pdf bib
A Transformer Based Approach towards Identification of Discourse Unit Segments and Connectives
Sahil Bakshi | Dipti Sharma

Discourse parsing, which involves understanding the structure, information flow, and modeling the coherence of a given text, is an important task in natural language processing. It forms the basis of several natural language processing tasks such as question-answering, text summarization, and sentiment analysis. Discourse unit segmentation is one of the fundamental tasks in discourse parsing and refers to identifying the elementary units of text that combine to form a coherent text. In this paper, we present a transformer based approach towards the automated identification of discourse unit segments and connectives. Early approaches towards segmentation relied on rule-based systems using POS tags and other syntactic information to identify discourse segments. Recently, transformer based neural systems have shown promising results in this domain. Our system, SegFormers, employs this transformer based approach to perform multilingual discourse segmentation and connective identification across 16 datasets encompassing 11 languages and 3 different annotation frameworks. We evaluate the system based on F1 scores for both tasks, with the best system reporting the highest F1 score of 97.02% for the treebanked English RST-DT dataset.

pdf
Multi-lingual Discourse Segmentation and Connective Identification: MELODI at Disrpt2021
Morteza Kamaladdini Ezzabady | Philippe Muller | Chloé Braud

We present an approach for discourse segmentation and discourse connective identification, both at the sentence and document level, within the Disrpt 2021 shared task, a multi-lingual and multi-formalism evaluation campaign. Building on the most successful architecture from the 2019 similar shared task, we leverage datasets in the same or similar languages to augment training data and improve on the best systems from the previous campaign on 3 out of 4 subtasks, with a mean improvement on all 16 datasets of 0.85%. Within the Disrpt 21 campaign the system ranks 3rd overall, very close to the 2nd system, but with a significant gap with respect to the best system, which uses a rich set of additional features. The system is nonetheless the best on languages that benefited from crosslingual training on sentence internal segmentation (German and Spanish).

pdf
Delexicalised Multilingual Discourse Segmentation for DISRPT 2021 and Tense, Mood, Voice and Modality Tagging for 11 Languages
Tillmann Dönicke

This paper describes our participating system for the Shared Task on Discourse Segmentation and Connective Identification across Formalisms and Languages. Key features of the presented approach are the formulation as a clause-level classification task, a language-independent feature inventory based on Universal Dependencies grammar, and composite-verb-form analysis. The achieved F1 is 92% for German and English and lower for other languages. The paper also presents a clause-level tagger for grammatical tense, aspect, mood, voice and modality in 11 languages.

pdf
A Unified Approach to Discourse Relation Classification in nine Languages
Hanna Varachkina | Franziska Pannach

This paper presents efforts to solve the shared task on discourse relation classification (disrpt task 3). The intricate prediction task aims to predict a large number of classes from the Rhetorical Structure Theory (RST) framework for nine target languages. Labels include discourse relations such as background, condition, contrast and elaboration. We present an approach using euclidean distance between sentence embeddings that were extracted using multlingual sentence BERT (sBERT) and directionality as features. The data was combined into five classes which were used for initial prediction. The second classification step predicts the target classes. We observe a substantial difference in results depending on the number of occurrences of the target label in the training data. We achieve the best results on Chinese, where our system achieves 70 % accuracy on 20 labels.

pdf
DisCoDisCo at the DISRPT2021 Shared Task: A System for Discourse Segmentation, Classification, and Connective Detection
Luke Gessler | Shabnam Behzad | Yang Janet Liu | Siyao Peng | Yilun Zhu | Amir Zeldes

This paper describes our submission to the DISRPT2021 Shared Task on Discourse Unit Segmentation, Connective Detection, and Relation Classification. Our system, called DisCoDisCo, is a Transformer-based neural classifier which enhances contextualized word embeddings (CWEs) with hand-crafted features, relying on tokenwise sequence tagging for discourse segmentation and connective detection, and a feature-rich, encoder-less sentence pair classifier for relation classification. Our results for the first two tasks outperform SOTA scores from the previous 2019 shared task, and results on relation classification suggest strong performance on the new 2021 benchmark. Ablation tests show that including features beyond CWEs are helpful for both tasks, and a partial evaluation of multiple pretrained Transformer-based language models indicates that models pre-trained on the Next Sentence Prediction (NSP) task are optimal for relation classification.