GSGüncel Sarıman

Güncel Sarıman

*This is an algorithmically generated author record based on publicly available information.

Current Organizations / Institutes

The organizations/institutes from the scholar's latest article


Güncel Sarıman has a total of 44 co-authors. Here are the top co-authors by number of shared publications:


Latest Publications

  • Veriden Karara: Bilgisayar Bilimlerinde Yapay Zekâ Temelli Yaklaşımlar From Data to Decision: Artificial Intelligence-Based Approaches in Computer Science

    by Cem Kalyoncu, Yusuf Kartal, Erkan Yaşar, Ferdi Güler, Halil İbRahim Yılmaz, Hacer Kübra Sevinç, Şüheda Semih Açmalı, Hilmi Özden, Abdullah Ortadeveci, İsHak Metehan Sis, et al.

    Veri çağında, geleceği anlamak ve kararları yeniden şekillendirmek için yapay zekânın gücünü keşfedin! Ham verinin karmaşasından doğru ve güvenilir kararlara ulaşmak, günümüz bilgi çağının en temel arayışıdır. Bu arayışın merkezinde duran yapay zekâ; artık yalnızca geleceğin vaadi değil, bilimin, sağlığın, üretimin ve yönetim mekanizmalarının bizzat bugünüdür. Büyük dil modellerinden otonom ajanlara, derin öğrenmeden optimizasyona uzanan gelişmeler, düşünme ve problem çözme alışkanlıklarımızı temelden dönüştürmektedir. Veriden Karara, bu dinamik dönüşümün sunduğu kuramsal derinliği ve pratik imkânları bir araya getiren kapsamlı bir rehber niteliğindedir. Kitap; veri analitiği, özellik mühendisliği ve metasezgisel optimizasyondan bilgisayarlı görü, Türkçe doğal dil işleme ve sağlık uygulamalarına uzanan zengin içeriğiyle okuru çok boyutlu bir bilimsel yolculuğa çıkarıyor. Ancak bir yapay zekâ modelinin gücü, yalnızca ürettiği çıktının başarısıyla ölçülemez. Modelin anlaşılabilir olması, mahremiyeti gözetmesi, etik sınırlara bağlı kalması ve her şeyden önemlisi güven vermesi gerekir. Bu eser, söz konusu sorumlulukları arka plana itmeden, teknolojiyi bütüncül ve eleştirel bir bakış açısıyla ele alıyor. Akademisyenlerden lisansüstü araştırmacılara, veri mühendislerinden teknoloji meraklılarına kadar herkes için ufuk açıcı bir başvuru kaynağı… Veriden bilgiye, bilgiden öngörüye, öngörüden güvenilir kararlara. Yapay zekâyı anlamakla yetinmeyin; onunla geleceği şekillendirin!

  • Open Access

    Hate Speech Detection In Social Media with Deep Learning And Language Models

    by Beste Akdik, Güncel Sarıman

    Nowadays, hate speech has started to spread rapidly with the increasing use of social media. Such abusive discourse can cause reputation damage and adversely affect psychological health. Large social media companies are trying to prevent this situation and increase their service quality with the increasing number of users every day. In this context, our study proposes a system that detects hate speech in texts and warns the user against hate speech. The project was implemented using machine learning, deep learning and language modeling techniques with a labeled hate speech dataset collected from various sources. The results show that BERTweet and DistilBERT language models achieved 90% accuracy. On the other hand, although the success of the classical models was lower, they were more effective temporally.

  • Open Access

    COMPARISON OF QUANTUM DEEP LEARNING METHODS FOR IMAGE CLASSIFICATION

    by Bekir Eray Kati, Ecir Uğur Küçüksille, Güncel Sarıman

    Nowadays, with the discovery of the power and potential of quantum computers, developing and understanding quantum-based deep learning models has become an important research area. This study investigates Quantum Transfer Learning and Quantum Hybrid Learning models that involve feature extraction and training processes using Convolutional Neural Networks (CNN) and Vision Transformer (ViT). The study aims to explore the potential advantages and differences of quantum deep learning techniques. It is envisioned that quantum computing can provide significant advantages in terms of computational speed and efficiency, especially in complex and large-scale data sets. Therefore, this study will contribute to a better understanding of the practical applications and potential impacts of quantum deep learning techniques. In this study, we evaluate the performance of four different quantum deep learning architectures using two different datasets. The classifiers used are the pre-trained ResNet-50 with a kernel size of 5x5 and the state-of-the-art CaiT-24-XXS-224 (CaiT) transducers. Optimization was performed with Adam optimizer using the cross entropy loss function. A total of eight models were trained, each with ten iterations. Accuracy (Acc), balanced accuracy (BA), overall F𝛽 (F_beta) macro score F1 and F2, Matthew's Correlation Coefficient (MCC), sensitivity (Sens) and specificity (Spec) were used as performance measures.


  • Open Access

    Enhancing Deepfake Detection Through Quantum Transfer Learning and Class-Attention Vision Transformer Architecture

    by Bekir Eray Katı, Ecir Uğur Küçüksille, Güncel Sarıman

    The widespread use of the internet, coupled with the increasing production of digital content, has caused significant challenges in information security and manipulation. Deepfake detection has become a critical research topic in both academic and practical domains, as it involves identifying forged elements in artificially generated videos using various deep learning and artificial intelligence techniques. In this dissertation, an innovative model was developed for detecting deepfake videos by combining the Quantum Transfer Learning (QTL) and Class-Attention Vision Transformer (CaiT) architectures. The Deepfake Detection Challenge (DFDC) dataset was used for training, and a system capable of detecting spatiotemporal inconsistencies was constructed by integrating QTL and CaiT technologies. In addition to existing preprocessing methods in the literature, a novel preprocessing function tailored to the requirements of deep learning models was developed for the dataset. The advantages of quantum computing offered by QTL were merged with the global feature extraction capabilities of the CaiT. The results demonstrated that the proposed method achieved a remarkable performance in detecting deepfake videos, with an accuracy of 90% and ROC AUC score of 0.94 achieved. The model’s performance was compared with other methods evaluated on the DFDC dataset, highlighting its efficiency in resource utilization and overall effectiveness. The findings reveal that the proposed QTL-CaiT-based system provides a strong foundation for deepfake detection and contributes significantly to the academic literature. Future research should focus on testing the model on real quantum devices and applying it to larger datasets to further enhance its applicability.


  • Open Access

    Akarsular üzerindeki antropojenik yapıların kıyı çizgisi üzerindeki etkisi: Dalaman Çayı

    by Ersin Ateş, Murat Gül, Güncel Sarıman, Iliya Bauchi Danladi

    Kıyı bölgeleri, deniz ticareti, turizm ve milyarlarca insanın evi olması bakımından hayati öneme sahiptir. Son zamanlarda kıyılar, küresel ısınma ve antropojenik etkilerden dolayı stres altında kalmıştır. Doğal ve antropojenik etkilerin kıyı bölgelerindeki etkilerini en iyi kısa nehirleri ve küçük drenaj alanları olan bölgelerden elde edilen verilerle açıklamak mümkündür (örneğin Dalaman Çayı). Dalaman Kıyısı, Güneybatı Türkiye'de popüler bir yaz turizmi destinasyonlarından biridir. Bu kıyıdaki kıyı çizgisi değişiklikleri, Weighted Linear Regression (WLR) analizi ve Digital Shoreline Analysis System (DSAS) kullanılarak belirlenmiştir. 30 yıl boyunca yağış, sıcaklık ve akım ile kıyı çizgisi değişiklikleri arasındaki ilişki, Pearson korelasyonu kullanılarak belirlenmiştir. 1989 ile 1999 doğal faktörlerin etkili olduğu bu dönemde tüm kıyı çizgisi ortalama -1.54 m kadar kabul edilebilir bir miktar gerilemiştir. 1999'da işletmeye alınan göletin ardından, tüm kıyı çizgisi 2000 ile 2010 yılları arasında ortalama -6.89 m, 2011 ile 2019 yılları arasında, tüm kıyı çizgisi ortalama -10.42 m gerilemiştir; bu, Akköprü Barajı'nın (2011'de işletmeye alınması) ve göletin sediment tutması sonucudur. Korelasyon analizinde, kıyı çizgisi doğal faktörlerin etkisi altındayken güçlü bir korelasyon gösterirken, Akköprü Barajı'nın devreye alınmasından sonra bu ilişki yaklaşık %50 oranında zayıflamıştır. Bölgede sürdürülebilir yaz turizmi ve tarım faaliyetlerini gerçekleştirmek için kıyı hattının, drenajın sürekli izlenmesi ve kıyı alanına dengeli müdahale yapılması gerekmektedir.


  • Next Word Prediction with Deep Learning Models

    by Abdullah Atçılı, Osman Özkaraca, Güncel Sarıman, Bogdan Patrut


  • A Novel Approach to Determine Software Security Level using Bayes Classifier via Static Code Metrics

    by Guncel Sarıman, Ecir Ugur Kucuksille

    Technological developments are increasing day by day and software products are growing in an uncontrolled way. This leads to the development of applications which do not comply with principles of design. Software which has not passed security testing may put the end user into danger. During the processes of error detection and verification of developed software, static and dynamic analysis may be used. Static code analysis provides analysis in different categories while coding without code compile. Source code metrics are also within these categories. Code metrics evaluate software quality, level of risk, and interchangeability by analysing software based on those metrics. In this study, we will describe our web-based application which is developed to determine the level of security in software. In this scope, software's metric calculation method will be explained. The scoring system we used to determine the security level calculation will be explained, taking into account metric thresholds that are acceptable in the literature. Bayes Classifier Method, distinguishing risks in the project files with the analysis of uploaded sample software files, will be described. Finally, objectives of this analysis method and planned activities will be explained.


  • Open Access

    Web Based Educational Tool for Metaheuristic Algorithms

    by Guncel Sariman, Ecir Ugur Kucuksille

    Metaheuristic optimization algorithms are nowadays being employed to solve a wide variety of optimization problems. These algorithms are not based on mathematical evidence and have mostly been developed by imitation of natural phenomenon. In this study, a web-based educational metaheuristics testing tool was developed. With this tool, the users are able to test Artificial Immune System and Artificial Bee Colony algorithms on Benchmark functions, observe the results of optimization by modifying the parameters for each algorithm, and at the same time perform optimization procedures by typing their own functions with their own constraints. In addition, information on the working steps of both algorithms are provided in the application