Neslihan Cesur

Özyeğin University

İstanbul, Turkey

I am a Master’s student in Artificial Intelligence in the department of Computer Engineering at Özyeğin University where I focus on how languages can be modeled and analyzed computationally. My main interests lie in the computational representation of lexical information, morphology, syntax, argument structure, and semantic roles. I’m currently working on building reliable guidelines and datasets for AI research, and automating annotation processes with rule-based and LLM-based methods for more reliable data. For my future research, I’m particularly interested in hybrid systems which combine symbolic representations and neural models to achieve data-efficient and cost-effective machine learning processes.

Before my current studies, I earned a Bachelor’s degree in Comparative Linguistics and French Literature from Galatasaray University and an M.A. in Linguistics from Boğaziçi University, where my research centered on linguistic typology and morphosyntax. For my master’s thesis, I investigated negation in Pazar Laz, a low-resource minority language spoken in northeastern Turkey. My thesis Negation in Pazar Laz can be accessed here.

news

Nov 28, 2025	Our paper, “A Semi-Automated Approach to the Annotation of Argument Structures in Turkish Datasets,” is now published in the proceedings of the 13th Global WordNet Conference. Read it on the ACL Anthology.
Jun 02, 2025	The Java repo and slides for CS201 Data Structures & Algorithms course are now on my Github
Jan 15, 2016	A simple inline announcement with Markdown emoji!

selected publications

A Semi-Automated Approach to the Annotation of Argument Structures in Turkish Datasets

Neslihan Cesur, Sabri Ince, Ali Hakkı Aydın, and 7 more authors

In Proceedings of the 13th Global Wordnet Conference, 2025

Abs

This paper presents a PropBank annotation project for Turkish, focusing on core arguments in matrix clauses. Our dataset comprises of five different corpora with 25,580 sentences labeled for verbal predicates and core arguments. Using a semi-automatic approach, we leveraged a dependency layer to pre-assign some ARG0s and ARG1s, followed by manual corrections. Our work will be used in the development of Abstract Meaning Representations (AMRs), enhancing Turkish NLP resources for semantic parsing and higher-level language tasks.
Building Annotated Parallel Corpora Using the ATIS Dataset: Two UD-style treebanks in English and Turkish

Neslihan Cesur, Aslı Kuzgun, Mehmet Kose, and 1 more author

In Proceedings of the 17th Workshop on Building and Using Comparable Corpora (BUCC) @ LREC-COLING 2024, May 2024
From Constituency to UD-Style Dependency: Building the First Conversion Tool of Turkish

Aslı Kuzgun, Oğuz Kerem Yıldız, Neslihan Cesur, and 6 more authors

In Proceedings of the International Conference on Recent Advances in Natural Language Processing (RANLP 2021), Sep 2021

Abs

This paper deliberates on the process of building the first constituency-to-dependency conversion tool of Turkish. The starting point of this work is a previous study in which 10,000 phrase structure trees were manually transformed into Turkish from the original PennTreebank corpus. Within the scope of this project, these Turkish phrase structure trees were automatically converted into UD-style dependency structures, using both a rule-based algorithm and a machine learning algorithm specific to the requirements of the Turkish language. The results of both algorithms were compared and the machine learning approach proved to be more accurate than the rule-based algorithm. The output was revised by a team of linguists. The refined versions were taken as gold standard annotations for the evaluation of the algorithms. In addition to its contribution to the UD Project with a large dataset of 10,000 Turkish dependency trees, this project also fulfills the important gap of a Turkish conversion tool, enabling the quick compilation of dependency corpora which can be used for the training of better dependency parsers.