Improving the Annotations in the Turkish Universal Dependency Treebank

Türk, Utku; Atmaca, Furkan; Özateş, Şaziye Betül; Başaran, Balkız Öztürk; Güngör, Tunga; Özgür, Arzucan

doi:10.18653/v1/w19-8013

Cited by 7 publications

(6 citation statements)

References 12 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…The most up-to-date version of this treebank is made available through Universal Dependencies (UD, Nivre et al, 2016;de Marneffe et al, 2021) repositories based on a semi-automatic conversion (Sulubacak et al, 2016) of a version from Istanbul Technical University (ITU) and hence, named UD-IMST (ITU-METU-Sabancı Treebank). Even the latest version is reported to have a large number of errors, carried over from earlier versions or introduced along the way by many automated conversion processes (see, e.g., Türk et al, 2019). Burga et al (2017) present a conversion of the same treebank into another related framework, namely Surface-Syntactic Universal Dependencies (SUD, Gerdes et al, 2018).…”

Section: Treebankmentioning

confidence: 99%

Resources for Turkish Natural Language Processing: A critical survey

Çöltekin¹,

Doğruöz²,

Çetinoğlu³

2022

Preprint

View full text Add to dashboard Cite

This paper presents a comprehensive survey of corpora and lexical resources available for Turkish. We review a broad range of resources, focusing on the ones that are publicly available. In addition to providing information about the available linguistic resources, we present a set of recommendations, and identify gaps in the data available for conducting research and building applications in Turkish Linguistics and Natural Language Processing.

show abstract

Section: Treebankmentioning

confidence: 99%

Resources for Turkish Natural Language Processing: A critical survey

Çöltekin¹,

Doğruöz²,

Çetinoğlu³

2022

Preprint

View full text Add to dashboard Cite

show abstract

“…Decisions regarding the annotation of the BOUN Treebank were made in line with the recent efforts for unifying the Turkish UD treebanks through manual reannotation [51]. In this context, the IMST-UD Treebank and Turkish PUD Treebank were also re-annotated manually [51], [59] and these re-annotated versions comply with the BOUN Treebank in terms of annotation decisions. Although the current version of the PUD Treebank (which is also the version used in our work) is this re-annotated version, the re-annotated version of the IMST-UD Treebank has not been FIGURE 10.…”

Section: The Effect Of Training Data Sizementioning

confidence: 99%

“…However, note that there are some major differences in the annotations of the current version of the IMST-UD Treebank and the BOUN Treebank. For more detailed information, see [59] and [52]. In each setup, the trained models were evaluated on four different test sets.…”

Section: Validated So We Use the Original Version Of The Imst-udmentioning

confidence: 99%

A Hybrid Deep Dependency Parsing Approach Enhanced With Rules and Morphology: A Case Study for Turkish

Özateş¹,

Özgür²,

Güngör³

et al. 2022

IEEE Access

View full text Add to dashboard Cite

Fully data-driven, deep learning-based models are usually designed as language-independent and have been shown to be successful for many natural language processing tasks. However, when the studied language is not high-resource and the amount of training data is insufficient, these models can benefit from the integration of natural language grammar-based information. We propose two approaches to dependency parsing especially for languages with restricted amount of training data. Our first approach combines a state-of-the-art deep learning-based parser with a rule-based approach and the second one incorporates morphological information into the parser. In the rule-based approach, the parsing decisions made by the rules are encoded and concatenated with the vector representations of the input words as additional information to the deep network. The morphology-based approach proposes different methods to include the morphological structure of words into the parser network. Experiments are conducted on three different Turkish treebanks and the results suggest that integration of explicit knowledge about the target language to a neural parser through a rule-based parsing system and morphological analysis leads to more accurate annotations and hence, increases the parsing performance in terms of attachment scores. The proposed methods are developed for Turkish, but can be adapted to other languages as well.INDEX TERMS Dependency parsing, computational linguistics, recurrent neural networks.

show abstract

“…In Sulubacak et al (2016b) they converted IMST into UD framework via a semi-automated process and mapping, naming it as IMST-UD which is one of the first Turkish UD treebanks. This treebank was later reannotated by Türk et al (2019b), improving the annotation of embedded structures and core versus non-core dependents.…”

Section: Related Workmentioning

confidence: 99%

Overcoming the challenges in morphological annotation of Turkish in universal dependencies framework

Bedir¹,

Şahin²,

Gungor³

et al. 2021

Proceedings of the Joint 15th Linguistic Annotation Workshop (LAW) and 3rd Designing Meaning Representations (DMR) Workshop

Self Cite

View full text Add to dashboard Cite

This paper presents several challenges faced when annotating Turkish treebanks in accordance with the Universal Dependencies (UD) guidelines and proposes solutions to address them. Most of these challenges stem from the lack of adequate support in the UD framework to accurately represent null morphemes and complex derivations, which results in significant loss of information for Turkish. This loss negatively impacts the tools that are developed based on these treebanks. We raised and discussed these issues within the community on the official UD portal. This paper presents these issues and our proposals to more accurately represent morphosyntactic information for Turkish while adhering to guidelines of UD. This work aims to contribute to the representation of Turkish and other agglutinative languages in UD-based treebanks, which in turn aids to develop more accurately annotated datasets for such languages.

show abstract

Improving the Annotations in the Turkish Universal Dependency Treebank

Cited by 7 publications

References 12 publications

Resources for Turkish Natural Language Processing: A critical survey

Resources for Turkish Natural Language Processing: A critical survey

A Hybrid Deep Dependency Parsing Approach Enhanced With Rules and Morphology: A Case Study for Turkish

Overcoming the challenges in morphological annotation of Turkish in universal dependencies framework

Contact Info

Product

Resources

About