D-STAG: A Formalism for Discourse Analysis Based on SDRT and Using Synchronous TAG

Danlos, Laurence

doi:10.1007/978-3-642-20169-1_5

Cited by 11 publications

(9 citation statements)

References 14 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…This shows for instance in the compositional treatment of the v adjunction model of the ensuite (afterwards) adverb. The type-theoretical foundations also offers an interesting link to more evolved theories of discourse relations using, in particular higher-order types [4]. The clarification of g-derivation tree status also suggests to study in a more principled way the relations between G-TAG and other approaches of generation using TAG such as [12,7].…”

Section: Resultsmentioning

confidence: 99%

An ACG View on G-TAG and Its g-Derivation

Danlos

Maskharashvili

Pogodalla

2014

Logical Aspects of Computational Linguistics

Self Cite

View full text Add to dashboard Cite

Abstract. G-TAG is a Tree Adjoining Grammar (TAG) based formalism which was specifically designed for the task of text generation. Contrary to TAG, the derivation structure becomes primary, as pivot between the conceptual representation and the surface form. This is a shared feature with the encoding of TAG into Abstract Categorial Grammars. This paper propose to study G-TAG from an ACG perspective. We rely on the reversibility property of ACG that makes both parsing and generation fall within a common morphism inversion process. Doing so, we show how to overcome some limitations of G-TAG regarding predicative adjunction and how to propose alternative approaches to some phenomena. OverviewTree-adjoining grammar (TAG) [8], [9] is one of the most extensively studied grammatical formalism as it considered to be a formalism that can encode the most of natural languages. The point is that TAGs produce string languages that are not in general context free, but mildly context-sensitive. The latter fact means that the tree language produced by a TAG, called derived tree language, is not a regular tree language and thus it might be problematic to deal with it in certain tasks. However, in TAG, together with the notion of derived tree, there is defined the notion of derivation tree that encodes the history of the process of the derivation of a derived tree. It turns out that while a TAG derived tree language generally is not regular, TAG derivation tree language is regular. This fact makes possible to explore the useful properties of regular tree languages on TAG derivation trees [7]. Furthermore, a type-logical account was given to TAG in [5], where TAG was represented as second order ACGs.G-TAG [3], [13] is a formalism designed for text generation, which makes use of the notions defined within TAG. However, in G-TAG the central notion of derivation tree (called g-derivation trees) is different from one that is defined in TAG. The goal of ⋆ This work has been supported by the French agency Agence Nationale de la Recherche (ANR-12-CORD-0004).2 the current paper is to show that g-derivation and g-derived trees can be encoded in the ACG framework in the same spirit as TAG is encoded as ACGs, and also to simulate G-TAG text generation process in the confines of the ACG framework. Tree Adjoining GrammarsTree adjoining grammars have been widely used in computational linguistics in a number of tasks including parsing and generation (e.g. [12], [7]).A TAG has two kinds of entries: initial and auxiliary trees, which are called elementary trees. An initial tree has interior nodes labeled by non-terminal symbol and frontier nodes, i.e. leaves, labeled by the symbols which can be either non-terminal or terminal ones. The non-terminal symbols appearing on the leaves of an initial tree are marked for substitution. For instance, γ reward , γ John and γ Mary are initial trees of TAG. Similarly to initial trees, an auxiliary tree has its interior nodes labeled by the non-terminal symbols and frontier nodes labeled by either non-terminal ...

show abstract

Section: Resultsmentioning

confidence: 99%

An ACG View on G-TAG and Its g-Derivation

Danlos

Maskharashvili

Pogodalla

2014

Logical Aspects of Computational Linguistics

Self Cite

View full text Add to dashboard Cite

show abstract

“…Martin [15] proposed discourse relations based on discourse connectives (e.g., because, but) expressed in the text. Danlos [2] extended sentence grammar and formalize discourse structure. Rhetorical Structure Theory or RST -used in our experiments -was proposed by Mann and Thompson [13].…”

Section: Rhetorical Analysismentioning

confidence: 99%

Method for Aspect-Based Sentiment Annotation Using Rhetorical Analysis

Augustyniak

Krzysztof

Kajdanowicz

2017

Intelligent Information and Database Systems

View full text Add to dashboard Cite

Abstract. This paper fills a gap in aspect-based sentiment analysis and aims to present a new method for preparing and analysing texts concerning opinion and generating user-friendly descriptive reports in natural language. We present a comprehensive set of techniques derived from Rhetorical Structure Theory and sentiment analysis to extract aspects from textual opinions and then build an abstractive summary of a set of opinions. Moreover, we propose aspect-aspect graphs to evaluate the importance of aspects and to filter out unimportant ones from the summary. Additionally, the paper presents a prototype solution of data flow with interesting and valuable results. The proposed method's results proved the high accuracy of aspect detection when applied to the gold standard dataset.

show abstract

“…It annotates the DC together with its binary arguments, which are semantic representations of discourse parts (cf. (Danlos, 2009)). The PDTB-style annotation has been extended to various languages other than English, namely, Arabic (Al-Saif and Markert, 2010), Chinese (Zhou and Xue, 2015), Hindi (Kolachina et al, 2012), and Turkish (Demirşahin and Zeyrek, 2017) as well as a recent multilingual resource, TED-Multilingual Discourse Bank, or TED-MDB (Zeyrek et al, 2019).…”

Section: Introductionmentioning

confidence: 99%

TCL - a Lexicon of Turkish Discourse Connectives

Zeyrek¹,

Başıbüyük²

2019

Proceedings of the First International Workshop on Designing Meaning Representations

View full text Add to dashboard Cite

It is known that discourse connectives are the most salient indicators of discourse relations. State-of-the-art parsers being developed to predict explicit discourse connectives exploit annotated discourse corpora but a lexicon of discourse connectives is also needed to enable further research in discourse structure and support the development of language technologies that use these structures for text understanding. This paper presents a lexicon of Turkish discourse connectives built by automatic means. The lexicon has the format of the German connective lexicon, DiMLex, where for each discourse connective, information about the connective's orthographic variants, syntactic category and senses are provided along with sample relations. In this paper, we describe the data sources we used and the development steps of the lexicon.

show abstract

D-STAG: A Formalism for Discourse Analysis Based on SDRT and Using Synchronous TAG

Cited by 11 publications

References 14 publications

An ACG View on G-TAG and Its g-Derivation

An ACG View on G-TAG and Its g-Derivation

Method for Aspect-Based Sentiment Annotation Using Rhetorical Analysis

TCL - a Lexicon of Turkish Discourse Connectives

Contact Info

Product

Resources

About