Existing work on probing of pretrained language models (LMs) has predominantly focused
on sentence-level syntactic tasks. In this paper, we introduce document-level discourse …

Discourse processing suffers from data sparsity, especially for dialogues. As a result, we
explore approaches to build discourse structures for dialogues, based on attention matrices …

With a growing number of BERTology work analyzing different components of pre-trained
language models, we extend this line of research through an in-depth analysis of discourse …

Discourse segmentation and sentence-level discourse parsing play important roles for
various NLP tasks to consider textual coherence. Despite recent achievements in both tasks …

Hierarchical discourse structures benefit Natural Language Understanding tasks, such as
text summarization and sentiment analysis. Rhetorical Structure Theory (RST) is particularly …

A document is more than a random combination of sentences. It is, instead, a cohesive entity
where sentences interact with each other to create a coherent structure and convey specific …

As large and powerful neural language models are developed, researchers have been
increasingly interested in developing diagnostic tools to probe them. There are many papers …

Recently, neural natural language models have attained state-of-the-art performance on a
wide variety of tasks, but the high performance can result from superficial, surface-level cues …

Topic segmentation, as a fundamental NLP task, has been proposed and systematically
studied since the 1980s and received increased attention in recent years due to the surge in …

In recent years, the impressive abilities shown by deep neural network (DNN)-based
systems have led to the curiosity towards the intrinsic mechanisms. The query towards these …