discourse
Coherence models are typically evaluated only on synthetic tasks, which may not be representative of their performance in downstream applications. To investigate how representative the synthetic tasks are of downstream use cases, we conduct experiments on benchmarking well-known traditional and neural coherence models on synthetic sentence ordering tasks, and contrast this with their performance on three downstream applications: coherence evaluation for MT and summarization, and next utterance prediction in retrieval-based dialog.
A unified coherence model that incorporates sentence grammar, inter-sentence coherence relations, and global coherence patterns into a common neural framework.
A hierarchical pointer network parsers applied to dependency and sentence-level discourse parsing tasks.
A neural approach for modeling coherence of asynchronus conversation
This repository contains the source code of our paper “A Unified Linear-Time Framework for Sentence-Level Discourse Parsing” in ACL 2019.
Getting Started These instructions will help you to run our unified discourse parser based on RST dataset.
Prerequisites * PyTorch 0.4 or higher * Python 3 * AllenNLP Dataset We train and evaluate the model with the standard RST Discourse Treebank (RST-DT) corpus. * Segmenter: we utilize all 7673 sentences for training and 991 sentences for testing.
About This package includes:
A discourse segmenter A discourse parser Evaluation metrics for discourse parsing Download Document-level Discourse Parser for English
Demo Link
Installation Required for the discourse segmenter:
Charniak’s reranking parser. Put it in Tools/CharniakParserRerank and install it. Taggers from UIUC. Download POS tagger and shallow chunker [LBJPOS.jar, LBJChunk.jar, LBJ2.jar, LBJ2Library.jar] and put these in Tools/UIUC_TOOLs/ Install scikit-learn and scipy (instructions) Install java if not installed (instructions for Ubuntu) Make sure the Tools/SPADE_UTILS/bin/edubreak is set to executable.
CON-S2V: A Generic Framework for Incorporating
Neural coherence for monologue
This resource addresses the problem of speech act recognition in written asynchronous conversations
This parser builds a discourse tree by applying an optimal parsing algorithm to probabilities inferred from two Conditional Random Fields: one for intra-sentential parsing and the other for multi-sentential parsing.