Browsing by Author "Keklik, Onur"

Now showing 1 - 3 of 3

Automatic Question Generation Using Natural Language Processing Techniques
(Izmir Institute of Technology, 2018-07) Keklik, Onur; Tuğlular, Tuğkan; Tekir, Selma; 03.04. Department of Computer Engineering; 03. Faculty of Engineering; 01. Izmir Institute of Technology
This thesis proposes a new rule based approach to automatic question generation. The proposed approach focuses on analysis of both syntactic and semantic structure of a sentence. The design and implementation of the proposed approach are also explained in detail. Although the primary objective of the designed system is question generation from sentences, automatic evaluation results shows that, it also achieves great performance on reading comprehension datasets, which focus on question generation from paragraphs. With respect to human evaluations, the designed system significantly outperforms all other systems and generated the most natural (human-like) questions.
Local Citation Recommendation With Graph Convolutional Networks
(01. Izmir Institute of Technology, 2024-07) Keklik, Onur; Tuğlular, Tuğkan; 03.04. Department of Computer Engineering; 03. Faculty of Engineering; 01. Izmir Institute of Technology
Local Citation Recommendation is a task that finds the missing reference in the corresponding citation placeholder. It is mainly contextual since context identifies the citation. On the other hand, a context can be a descriptor for a set of papers. In other words, there can be more than one candidate citation for a context. Thus, a further matching of a context with candidate papers is beneficial. Titles and abstracts of candidate papers serve as a global context to match with the local one. This work proposes a state-of-the-art approach for the Local Citation Recommendation task that exploits the similarities between global and local contexts to generate citation predictions. By utilizing a Graph Convolutional Network (GCN) with BERT embeddings, our proposed model demonstrates superior performance over previous methods. It not only outperforms all prior approaches on the benchmark datasets of ACL-200, FullTextPeerRead, RefSeer, and arXiv but also strikes a balance between speed, memory, and computational requirements. Once deployed as a production-level Local Citation Recommendation, it is fast enough to enable real-time recommendations for researchers.
Citation - WoS: 6
Citation - Scopus: 10
Rule-Based Automatic Question Generation Using Semantic Role Labeling
(Institute of Electronics, Information and Communication Engineers, 2019) Keklik, Onur; Tuğlular, Tuğkan; Tekir, Selma; 03.04. Department of Computer Engineering; 03. Faculty of Engineering; 01. Izmir Institute of Technology
This paper proposes a new rule-based approach to automatic question generation. The proposed approach focuses on analysis of both syntactic and semantic structure of a sentence. Although the primary objective of the designed system is question generation from sentences, automatic evaluation results shows that, it also achieves great performance on reading comprehension datasets, which focus on question generation from paragraphs. Especially, with respect to METEOR metric, the designed system significantly outperforms all other systems in automatic evaluation. As for human evaluation, the designed system exhibits similar performance by generating the most natural (human-like) questions.