Universal Dependencies

Abstract

Universal dependencies (UD) is a framework for morphosyntactic annotation of human language, which to date has been used to create treebanks for more than 100 languages. In this article, we outline the linguistic theory of the UD framework, which draws on a long tradition of typologically oriented grammatical theories. Grammatical relations between words are centrally used to explain how predicate–argument structures are encoded morphosyntactically in different languages while morphological features and part-of-speech classes give the properties of words. We argue that this theory is a good basis for cross-linguistically consistent annotation of typologically diverse languages in a way that supports computational natural language understanding as well as broader linguistic studies.

Syntax Vector Learning Using Correspondence for Natural Language Understanding

“…We used this method to find sentences having a core syntactic structure. For example, for dependency parsing, there are 48 syntactic dependency labels in the English corpora, in which the core dependents of the clausal predicates are the nominal dep (nsubj, nsubjpass, dobj), and predicate dep (csubj, csubjpass, ccomp, xcomp) [22]. We provide a weight when a sentence contains these core dependents.…”

Dependency Syntax in the Automatic Detection of Irony and Stance

“…The general philosophy is to provide a universal inventory of categories and guidelines to facilitate consistent annotation of similar constructions across languages, while allowing language-specific extensions when necessary. This framework is thus a perfectly good basis for crosslinguistically consistent annotation of typologically diverse languages in a way that supports computational natural language understanding as well as broader linguistic studies [De Marneffe et al, 2021]. The main principles underlying UD are those of offering a linguistic representation that could be useful for morphosyntactic research, semantic interpretation, and for practical natural language processing across different human languages.…”

“…The whole project started in 2016, with the first guidelines, involving treebanks for 33 different languages, and it is very much alive today with currently available treebanks for 104 languages. It draws on a long tradition of typologically oriented grammatical theories -in particular dependency grammarin which grammatical relations between words are centrally used to explain how predicate-argument structures are encoded morphosyntactically in different languages while morphological features and part-of-speech classes give the properties of words [De Marneffe et al, 2021].…”

“…In the UD framework, a dependency grammar perspective is adopted, meaning that a phrase has a head and all the other components that are contained are dependents of that head [De Marneffe et al, 2021]. A "dependency relation" is represented in diagrams Artificial Intelligence (XAI) https://sites.google.com/view/xai2020/home.…”

“…Then, words are connected to the head via a grammatical relation drawn from a universal typology of 37 grammatical relations that can be read here: https:// universaldependencies.org/u/dep/index.html - [De Marneffe et al, 2021]. The grammatical relations are organized around whether the head is the head of a clause or nominal, and whether the dependent is a clause, nominal, or modifier.…”