This paper presents DISRPT, a multilingual, multi-domain, and cross-framework benchmark dataset for discourse processing, covering the tasks of discourse unit segmentation, connective identification, and relation classification. DISRPT includes 13 languages, with data from 24 corpora covering about 4 millions tokens and around 250,000 discourse relation instances from 4 discourse frameworks: RST, SDRT, PDTB, and Discourse Dependencies. We present an overview of the data, its development across three NLP shared tasks on discourse processing carried out in the past five years, and the latest modifications and added extensions. We also carry out an evaluation of state-of-the-art multilingual systems trained on the data for each task, showing plateau performance on segmentation, but important room for improvement for connective identification and relation classification. The DISRPT benchmark employs a unified format that we make available on GitHub and HuggingFace in order to encourage future work on discourse processing across languages, domains, and frameworks.

Classifying discourse relations is known as a hard task, relying on complex indices. On the other hand, discourse-annotated data is scarce, especially for languages other than English: many corpora, of limited size, exist for several languages but the domain is split between different theoretical frameworks that have a huge impact on the nature of the textual spans to be linked, and the label set used. Moreover, each annotation project implements modifications compared to the theoretical background and other projects. These discrepancies hinder the development of systems taking advantage of all the available data to tackle data sparsity and work on transfer between languages is very limited, almost nonexistent between frameworks, while it could improve our understanding of some theoretical aspects and enhance many applications. In this paper, we propose the first experiments on zero-shot learning for discourse relation classification and investigate several paths in the way source data can be combined, either based on languages, frameworks, or similarity measures. We demonstrate how difficult transfer is for the task at hand, and that the most impactful factor is label set divergence, where the notion of underlying framework possibly conceals crucial disagreements.

Depression is a serious mental illness that impacts the way people communicate, especially through their emotions, and, allegedly, the way they interact with others.This work examines depression signals in dialogs, a less studied setting that suffers from data sparsity.We hypothesize that depression and emotion can inform each other, and we propose to explore the influence of dialog structure through topic and dialog act prediction.We investigate a Multi-Task Learning (MTL) approach, where all tasks mentioned above are learned jointly with dialog-tailored hierarchical modeling.We experiment on the DAIC and DailyDialog corpora -both contain dialogs in English -and show important improvements over state-ofthe-art on depression detection (at best 70.6% F 1 ), which demonstrates the correlation of depression with emotion and dialog organization and the power of MTL to leverage information from different sources.

Various machine learning tasks can benefit from access to external information of different modalities, such as text and images. Recent work has focused on learning architectures with large memories capable of storing this knowledge. We propose augmenting generative Transformer neural networks with KNN-based Information Fetching (KIF) modules. Each KIF module learns a read operation to access fixed external knowledge. We apply these modules to generative dialog modeling, a challenging task where information must be flexibly retrieved and incorporated to maintain the topic and flow of conversation. We demonstrate the effectiveness of our approach by identifying relevant knowledge required for knowledgeable but engaging dialog from Wikipedia, images, and human-written dialog utterances, and show that leveraging this retrieved information improves model performance, measured by automatic and human evaluation.

Angela Fan, Claire Gardent, Chloé Braud, Antoine Bordes. Proceedings of the 2019 Conference on Empirical Methods in Natural Language Processing and the 9th International Joint Conference on Natural Language Processing (EMNLP-IJCNLP). 2019.

Discourse relation classification has proven to be a hard task, with rather low performance on several corpora that notably differ on the relation set they use.We propose to decompose the task into smaller, mostly binary tasks corresponding to various primitive concepts encoded into the discourse relation definitions.More precisely, we translate the discourse relations into a set of values for attributes based on distinctions used in the mappings between discourse frameworks proposed by Sanders et al. (2018).This arguably allows for a more robust representation of discourse relations, and enables us to address usually ignored aspects of discourse relation prediction, namely multiple labels and underspecified annotations.We study experimentally which of the conceptual primitives are harder to learn from the Penn Discourse Treebank English corpus, and propose a correspondence to predict the original labels, with preliminary empirical comparisons with a direct model.

Discourse parsing is an integral part of understanding information flow and argumentative structure in documents.Most previous research has focused on inducing and evaluating models from the English RST Discourse Treebank.However, discourse treebanks for other languages exist, including Spanish, German, Basque, Dutch and Brazilian Portuguese.The treebanks share the same underlying linguistic theory, but differ slightly in the way documents are annotated.In this paper, we present (a) a new discourse parser which is simpler, yet competitive (significantly better on 2/3 metrics) to state of the art for English, (b) a harmonization of discourse treebanks across languages, enabling us to present (c) what to the best of our knowledge are the first experiments on crosslingual discourse parsing.

Discourse segmentation is the first step in building discourse parsers.Most work on discourse segmentation does not scale to real-world discourse parsing across languages, for two reasons: (i) models rely on constituent trees, and (ii) experiments have relied on gold standard identification of sentence and token boundaries.We therefore investigate to what extent constituents can be replaced with universal dependencies, or left out completely, as well as how state-of-the-art segmenters fare in the absence of sentence boundaries.Our results show that dependency information is less useful than expected, but we provide a fully scalable, robust model that only relies on part-of-speech information, and show that it performs well across languages in the absence of any gold-standard annotation.

We experiment with different ways of training LSTM networks to predict RST discourse trees. The main challenge for RST discourse parsing is the limited amounts of training data. We combat this by regularizing our models using task supervision from related tasks as well as alternative views on discourse structures. We show that a simple LSTM sequential discourse parser takes advantage of this multi-view and multi-task framework with 12-15% error reductions over our baseline (depending on the metric) and results that rival more complex state-of-the-art parsers.

We introduce a simple semi-supervised approach to improve implicit discourse relation identification.This approach harnesses large amounts of automatically extracted discourse connectives along with their arguments to construct new distributional word representations.Specifically, we represent words in the space of discourse connectives as a way to directly encode their rhetorical function.Experiments on the Penn Discourse Treebank demonstrate the effectiveness of these task-tailored representations in predicting implicit discourse relations.Our results indeed show that, despite their simplicity, these connective-based representations outperform various off-the-shelf word embeddings, and achieve state-of-the-art performance on this problem.

This paper presents a detailed comparative framework for assessing the usefulness of unsupervised word representations for identifying so-called implicit discourse relations.Specifically, we compare standard one-hot word pair representations against low-dimensional ones based on Brown clusters and word embeddings.We also consider various word vector combination schemes for deriving discourse segment representations from word vectors, and compare representations based either on all words or limited to head words.Our main finding is that denser representations systematically outperform sparser ones and give state-of-the-art performance or above without the need for additional hand-crafted features.