A Minimalist Approach to Shallow Discourse Parsing and Implicit Relation Recognition

Chiarcos, Christian; Schenk, Niko

doi:10.18653/v1/k15-2006

Cited by 9 publications

(9 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…For discourse connective and argument extraction, token level features extracted from a fixed window centered on the target word token are generally used, and so are features extracted from syntactic parses. Distributional representations such as Brown clusters have generally been used to determine the senses (Chiarcos and Schenk, 2015;Devi et al, 2015;Kong et al, 2015;Song et al, 2015;Stepanov et al, 2015;Wang and Lan, 2015;Yoshida et al, 2015), although one team also used them in the sequence labeling task for argument extraction (Nguyen et al, 2015). Additional resources used by some systems for sense determination include word embeddings (Chiarcos and Schenk, 2015;, Verb-Net classes (Devi et al, 2015;Kong et al, 2015), and the MPQA polarity lexicon (Devi et al, 2015;Kong et al, 2015;Wang and Lan, 2015).…”

Section: Approachesmentioning

confidence: 99%

The CoNLL-2015 Shared Task on Shallow Discourse Parsing

Xue

Pradhan

et al. 2015

Proceedings of the Nineteenth Conference on Computational Natural Language Learning - Shared Task

109

View full text Add to dashboard Cite

The CoNLL-2015 Shared Task is on Shallow Discourse Parsing, a task focusing on identifying individual discourse relations that are present in a natural language text. A discourse relation can be expressed explicitly or implicitly, and takes two arguments realized as sentences, clauses, or in some rare cases, phrases. Sixteen teams from three continents participated in this task. For the first time in the history of the CoNLL shared tasks, participating teams, instead of running their systems on the test set and submitting the output, were asked to deploy their systems on a remote virtual machine and use a web-based evaluation platform to run their systems on the test set. This meant they were unable to actually see the data set, thus preserving its integrity and ensuring its replicability. In this paper, we present the task definition, the training and test sets, and the evaluation protocol and metric used during this shared task. We also summarize the different approaches adopted by the participating teams, and present the evaluation results. The evaluation data sets and the scorer will serve as a benchmark for future research on shallow discourse parsing.

show abstract

Section: Approachesmentioning

confidence: 99%

The CoNLL-2015 Shared Task on Shallow Discourse Parsing

Xue

Pradhan

et al. 2015

Proceedings of the Nineteenth Conference on Computational Natural Language Learning - Shared Task

109

View full text Add to dashboard Cite

show abstract

“…The so-obtained explicit argument pairs are sense labeled by a (linear-kernel) SVM classifier 12 with the connector word as the only feature, following the minimalist setting in Chiarcos and Schenk (2015).…”

Section: Chinese Full Task Pipeline (Cftp)mentioning

confidence: 99%

“…A challenging task is to detect the sense of NonExplicit discourse relations, as they usually don't have a connective that can help to determine their sense. In last year's task Non-Explicit relations have been tackled with features based on Brown clusters (Chiarcos and Schenk, 2015;, VerbNet classes and MPQA polarity lexicon . Earlier work (Rutherford and Xue, 2014) employed Brown cluster and coreference patterns to identify senses of implicit discourse relations in naturally occurring text.…”

Section: Related Workmentioning

confidence: 99%

“…In CoNLL-2015, various approaches were explored to conquer the sense classification problem, which is a straightforward multi-category classification task (Okita et al, 2015;Chiarcos and Schenk, 2015;Song et al, 2015;Yoshida et al, 2015;Sun et al, 2015;Laali et al, 2015). Typical data-driven machine learning methods, like Maximum Entropy and Support Vector Machine, were adopted.…”

Section: Introductionmentioning

confidence: 99%

See 1 more Smart Citation

Proceedings of the CoNLL-16 shared task

Sang

2016

View full text Add to dashboard Cite

(Volume 2) ii Introduction This volume contains papers describing the CoNLL-2016 Shared Task and the participating systems. The 2016 shared task is on multilingual Shallow Discourse Parsing (SDP), and is a follow-on to the 2015 shared task. The languages covered in this shared task are English and Chinese. The SDP task involves identifying individual discourse relations that are present in a natural language text. A discourse relation can be expressed explicitly or implicitly, and takes two arguments realized as sentences, clauses, or in some rare cases, phrases. Shallow Discourse Parsing is a fundamental NLP task and can potentially benefit a range of natural language applications such as Information Extraction, Text Summarization, Question Answering, Machine Translation, and Sentiment Analysis.A total of 24 teams from three continents participated in this task, and 20 of them submitted system description papers. Many different approaches were adopted by the participants, and we hope that these approaches help to advance the state of the art in Shallow Discourse Parsing. The training, development, and test sets for English and Chinese were adapted from the Penn Discourse TreeBank (PDTB) and the Chinese Discourse TreeBank (CDTB) respectively. In addition, we also annotated a blind test set for each language following the PDTB and CDTB guidelines solely for the shared task. The results on the blind test sets were used to rank the participating systems. The evaluation scorer, also developed for this shared task, adopts an F1 based metric that takes into account the accuracy of identifying the senses and arguments of discourse relations as well as explicit discourse connectives. We hope that the data sets and the scorer, which are freely available upon the completion of the shared task, will be a useful resource for researchers interested in discourse parsing.As with the 2015 CoNLL shared task on SDP, participants did not each run their systems locally on the test set. Instead, they were asked to deploy their systems on a remote virtual machine and use a webbased evaluation platform called TIRA to run their systems on the test set. This kept them from seeing the data set, thus preserving its integrity and ensuring its replicability.We would like to thank all the participants of the 2016 Shared Task, as well as the program committee for helping us review the system description papers. Special thanks are due to the SIGNLL board members, Xavier Carreras and Julia Hockenmaier, for their support of the shared task over the last two years. We would also like to thank the PDTB team and CDTB team for providing annotated data for the shared task, and the Linguistic Data Consortium for their help with releasing the data to the participants. Special thanks go to Martin Potthast and the TIRA team, who provided their computing resources and more importantly their time in assisting teams to run their systems. AbstractThe CoNLL-2016 Shared Task is the second edition of the CoNLL-2015 Shared Task, now on Multilingual Shallow di...

show abstract

“…In CoNLL-2015, various approaches were explored to conquer the sense classification problem, which is a straightforward multi-category classification task (Okita et al, 2015;Wang and Lan, 2015;Chiarcos and Schenk, 2015;Song et al, 2015;Stepanov et al, 2015;Yoshida et al, 2015;Sun et al, 2015;Nguyen et al, 2015;Laali et al, 2015). Typical data-driven machine learning methods, like Maximum Entropy and Support Vector Machine, were adopted.…”

Section: Introductionmentioning

confidence: 99%

Shallow Discourse Parsing Using Convolutional Neural Network

Qin¹,

Zhang²,

Zhao³

2016

Proceedings of the CoNLL-16 Shared Task

View full text Add to dashboard Cite

This paper describes a discourse parsing system for our participation in the CoNLL 2016 Shared Task. We focus on the supplementary task: Sense Classification, especially the Non-Explicit one which is the bottleneck of discourse parsing system. To improve Non-Explicit sense classification, we propose a Convolutional Neural Network (CNN) model to determine the senses for both English and Chinese tasks. We also explore a traditional linear model with novel dependency features for Explicit sense classification. Compared with the best system in CoNLL-2015, our system achieves competitive performances. Moreover, as shown in the results, our system has higher F1 score on Non-Explicit sense classification.

show abstract

A Minimalist Approach to Shallow Discourse Parsing and Implicit Relation Recognition

Cited by 9 publications

References 20 publications

The CoNLL-2015 Shared Task on Shallow Discourse Parsing

The CoNLL-2015 Shared Task on Shallow Discourse Parsing

Proceedings of the CoNLL-16 shared task

Shallow Discourse Parsing Using Convolutional Neural Network

Contact Info

Product

Resources

About