Bilgisayar Mühendisliği Bölümü Koleksiyonu
Permanent URI for this collectionhttps://hdl.handle.net/20.500.11779/1940
Browse
6 results
Search Results
Conference Object Citation - Scopus: 1Does Prompt Engineering Help Turkish Named Entity Recognition?(Institute of Electrical and Electronics Engineers Inc., 2024) Pektezol, A.S.; Ulugergerli, A.B.; Demir, Şeniz; Demir, Şeniz; 02.02. Department of Computer Engineering; 02. Faculty of Engineering; 01. MEF UniversityThe extraction of entity mentions in a text (named entity recognition) has been traditionally formulated as a sequence labeling problem. In recent years, this approach has evolved from recognizing entities to answering formulated questions related to entity types. The questions, constructed as prompts, are used to elicit desired entity mentions and their types from large language models. In this work, we investigated prompt engineering in Turkish named entity recognition and studied two prompting strategies to guide pretrained language models toward correctly identifying mentions. In particular, we examined the impact of zero-shot and few-shot prompting on the recognition of Turkish named entities by conducting experiments on two large language models. Our evaluations using different prompt templates revealed promising results and demonstrated that carefully constructed prompts can achieve high accuracy on entity recognition, even in languages with complex morphology. © 2024 IEEE.Article Mention Detection in Turkish Coreference Resolution(Tubitak Scientific & Technological Research Council Turkey, 2024) Demir, Seniz; Akdag, Hanifi Ibrahim; Demir, Şeniz; 02.02. Department of Computer Engineering; 02. Faculty of Engineering; 01. MEF UniversityA crucial step in understanding natural language is detecting mentions that refer to real-world entities in a text and correctly identifying their boundaries. Mention detection is commonly considered a preprocessing step in coreference resolution which is shown to be helpful in several language processing applications such as machine translation and text summarization. Despite recent efforts on Turkish coreference resolution, no standalone neural solution to mention detection has been proposed yet. In this article, we present two models designed for detecting Turkish mentions by using feed-forward neural networks. Both models extract all spans up to a fixed length from input text as candidates and classify them as mentions or not mentions. The models differ in terms of how candidate text spans are represented. The first model represents a span by focusing on its first and last words, whereas the representation also covers the preceding and proceeding words of a span in the second model. Mention span representations are formed by using contextual embeddings, part-of-speech embeddings, and named-entity embeddings of words in interest where contextual embeddings are obtained from pretrained Turkish language models. In our evaluation studies, we not only assess the impact of mention representation strategies on system performance but also demonstrate the usability of different pretrained language models in resolution task. We argue that our work provides useful insights to the existing literature and the first step in understanding the effectiveness of neural architectures in Turkish mention detection.Article Citation - WoS: 3Citation - Scopus: 3A Benchmark Dataset for Turkish Data-To Generation(Elsevier, 2023) Demir, Şeniz; Öktem, Seza; Demir, Şeniz; 02.02. Department of Computer Engineering; 02. Faculty of Engineering; 01. MEF UniversityIn the last decades, data-to-text (D2T) systems that directly learn from data have gained a lot of attention in natural language generation. These systems need data with high quality and large volume, but unfortunately some natural languages suffer from the lack of readily available generation datasets. This article describes our efforts to create a new Turkish dataset (Tr-D2T) that consists of meaning representation and reference sentence pairs without fine-grained word alignments. We utilize Turkish web resources and existing datasets in other languages for producing meaning representations and collect reference sentences by crowdsourcing native speakers. We particularly focus on the generation of single-sentence biographies and dining venue descriptions. In order to motivate future Turkish D2T studies, we present detailed benchmarking results of different sequence-to-sequence neural models trained on this dataset. To the best of our knowledge, this work is the first of its kind that provides preliminary findings and lessons learned from the creation of a new Turkish D2T dataset. Moreover, our work is the first extensive study that presents generation performances of transformer and recurrent neural network models from meaning representations in this morphologically-rich language.Article Citation - WoS: 7Citation - Scopus: 12Graph-Based Turkish Text Normalization and Its Impact on Noisy Text Processing(Elsevier, 2022) Topçu, Berkay; Demir, Şeniz; 02.02. Department of Computer Engineering; 02. Faculty of Engineering; 01. MEF UniversityUser generated texts on the web are freely-available and lucrative sources of data for language technology researchers. Unfortunately, these texts are often dominated by informal writing styles and the language used in user generated content poses processing difficulties for natural language tools. Experienced performance drops and processing issues can be addressed either by adapting language tools to user generated content or by normalizing noisy texts before being processed. In this article, we propose a Turkish text normalizer that maps non-standard words to their appropriate standard forms using a graph-based methodology and a context-tailoring approach. Our normalizer benefits from both contextual and lexical similarities between normalization pairs as identified by a graph-based subnormalizer and a transformation-based subnormalizer. The performance of our normalizer is demonstrated on a tweet dataset in the most comprehensive intrinsic and extrinsic evaluations reported so far for Turkish. In this article, we present the first graph-based solution to Turkish text normalization with a novel context-tailoring approach, which advances the state-of-the-art results by outperforming other publicly available normalizers. For the first time in the literature, we measure the extent to which the accuracy of a Turkish language processing tool is affected by normalizing noisy texts before being processed. An analysis of these extrinsic evaluations that focus on more than one Turkish NLP task (i.e., part-of-speech tagger and dependency parser) reveals that Turkish language tools are not robust to noisy texts and a normalizer leads to remarkable performance improvements once used as a preprocessing tool in this morphologically-rich language.Article Citation - WoS: 1Citation - Scopus: 2Extracting, Computing, Coordination: What Does a Triphasic Erp Pattern Say About Language Processing?(Elsevier, 2021) Çakar, Tuna; Çakar, Tuna; Cedden, Gülay; 02.02. Department of Computer Engineering; 02. Faculty of Engineering; 01. MEF UniversityThe current study aims at contributing to the interpretation of the most prominent language-related ERP effects, N400 and P600, by investigating how neural responses to congruent and incongruent sentence endings vary, when the language processor processes the full array of the lexico-syntactic content in verbs with three affixes in canonical Turkish sentences. The ERP signals in response to three different violation conditions reveal a similar triphasic (P200/N400/P600) pattern resembling in topography and peak amplitude The P200 wave is interpreted as the extraction of meaning from written.form by generating a code which triggers the computation of neuronal ensembles in the distributed LTM (N400). The P600 potential reflects the widely distributed coordination process of activated neuronal patterns of semantic and morphosyntactic cues by connecting the generated subsets of these patterns and adapting them into the current context. It further can be deduced that these ERP components reflect cognitive rather than linguistic processes. © 2021 Informa UK Limited, trading as Taylor & Francis Group.Article Citation - WoS: 19Citation - Scopus: 28An Evaluation of Recent Neural Sequence Tagging Models in Turkish Named Entity Recognition(Elsevier, 2021) Makaroğlu, Didem; Demir, Şeniz; Demir, Şeniz; Aras, Gizem; Çakır, Altan; 02.02. Department of Computer Engineering; 02. Faculty of Engineering; 01. MEF UniversityNamed entity recognition (NER) is an extensively studied task that extracts and classifies named entities in a text. NER is crucial not only in downstream language processing applications such as relation extraction and question answering but also in large scale big data operations such as real-time analysis of online digital media content. Recent research efforts on Turkish, a less studied language with morphologically rich nature, have demonstrated the effectiveness of neural architectures on well-formed texts and yielded state-of-the art results by formulating the task as a sequence tagging problem. In this work, we empirically investigate the use of recent neural architectures (Bidirectional long short-term memory (BiLSTM) and Transformer-based networks) proposed for Turkish NER tagging in the same setting. Our results demonstrate that transformer-based networks which can model long-range context overcome the limitations of BiLSTM networks where different input features at the character, subword, and word levels are utilized. We also propose a transformer-based network with a conditional random field (CRF) layer that leads to the state-of-the-art result (95.95% f-measure) on a common dataset. Our study contributes to the literature that quantifies the impact of transfer learning on processing morphologically rich languages.