13 results
Search Results
Now showing 1 - 10 of 13
Review Citation - WoS: 3Citation - Scopus: 6Machine Learning for Sustainable Reutilization of Waste Materials as Energy Sources - a Comprehensive Review(Taylor & Francis inc, 2024) Peng, Wei; Sadaghiani, Omid KarimiThis work reviews Machine Learning applications in the sustainable utilization of waste materials as energy source so that analysis of the past works exposed the lack of reviewing study. To solve it, the origin of waste biomass raw materials is explained, and the application of Machine Learning in this section is scrutinized. After analysis of numerous papers, it is concluded that Machine Learning and Deep Learning are widely utilized in waste biomass production areas to enhance the quality and quantity of production, improve the predictions, diminish the losses, as well as increase storage and transformation conditions. The positive effects and application with the utilized algorithms and other effective information are collected in this work for the first time. According to the statistical analysis, in 20% out of the studies conducted about the application of Machine Learning and Deep Learning in waste biomass raw materials, Artificial Neural Network (ANN) algorithm has been applied. Afterward, the Super Vector Machine (SVM) and Random Forest (RF) are the second and third most-utilized algorithms applied in 15% and 14% of studies. Meanwhile, 27% of studies focused on the applications of Machine Learning and Deep Learning in the Forest wastes.Conference Object Citation - Scopus: 2Enhancing Image Resolution With Generative Adversarial Networks(Institute of Electrical and Electronics Engineers Inc., 2022) Yildiz,B.Super-resolution is the process of generating high-resolution images from low-resolution images. There are a variety of practical applications used in real-world problems such as high-definition content creation, surveillance imaging, gaming, and medical imaging. Super-resolution has been the subject of many researches over the past few decades, as improving image resolution offers many advantages. Going beyond the previously presented methods, Generative Adversarial Networks offers a very promising solution. In this work, we will use the Generative Adversarial Networks-based approach to obtain 4x resolution images that are perceptually better than previous solutions. Our extensive experiments, including perceptual comparison, Peak Signal-to-Noise Ratio, and classification success metrics, show that our approach is quite promising for image super-resolution. © 2022 IEEE.Master Thesis Yapay Zeka Teknikleri Kullanılarak Bor Nitrür Kaplamalarının Modellenmesi(2025) Küçüköztaş, Korcan; Turhan, Çiğdem; Kaftanoğlu, BilginBor nitrür (BN), yüksek ısıl iletkenlik, düşük sürtünme katsayısı ve yüksek sertlik gibi mükemmel özelliklere sahip bir seramik malzemedir. Ancak, BN kaplamalarının Fiziksel Buhar Biriktirme (FBB) süreci ile Magnetron Saçtırma (MS) tekniği kullanılarak uygulanması, süreç parametreleri ile kaplama özellikleri arasındaki karmaşık etkileşimler nedeniyle zorludur. Bu tez, altı gelişmiş makine öğrenmesi tekniğinden yararlanan Yapay Zeka (YZ) tabanlı bir çerçeve kullanılarak BN kaplama sürecinin modellenmesi ve optimizasyonuna yönelik yenilikçi bir yaklaşım sunmaktadır. Çelik numuneler, farklı kaplama parametreleri ile kaplanmış ve yüksek hassasiyetli ekipmanlarla karakterize edilmiştir. Verileri tanıyabilmek amacıyla, keşifsel veri analizi gerçekleştirilmiştir. Üç farklı kaplama özelliğini tahmin etmek üzere altı farklı mimari kullanılarak makine öğrenmesi modelleri geliştirilmiş ve regresyon değerlendirme metrikleri ile karşılaştırılmıştır. Son olarak, en başarılı modeller, yeni veri setleri üzerinde tahminlerde bulunmak amacıyla kullanılmış ve sonuçlar görselleştirilmiştir. YZ tabanlı yaklaşım, karar verme süresini azaltarak istenilen özelliklere göre en uygun parametrelerinin belirlenmesini sağlamaktadır.Conference Object Citation - WoS: 1An Undergraduate Curriculum for Deep Learning(Ieee, 2018) Tirkes, Guzin; Ekin, Cansu Cigdem; Sengul, Gokhan; Bostan, Atila; Karakaya, MuratDeep Learning (DL) is an interesting and rapidly developing field of research which has been currently utilized as a part of industry and in many disciplines to address a wide range of problems, from image classification, computer vision, video games, bioinformatics, and handwriting recognition to machine translation. The starting point of this study is the recognition of a big gap between the sector need of specialists in DL technology and the lack of sufficient education provided by the universities. Higher education institutions are the best environment to provide this expertise to the students. However, currently most universities do not provide specifically designed DL courses to their students. Thus, the main objective of this study is to design a novel curriculum including two courses to facilitate teaching and learning of DL topic. The proposed curriculum will enable students to solve real-world problems by applying DL approaches and gain necessary background to adapt their knowledge to more advanced, industry-specific fields.Conference Object Securing the Internet of Things: Challenges and Complementary Overview of Machine Learning-Based Intrusion Detection(Institute of Electrical and Electronics Engineers Inc., 2024) Isin, L.I.; Dalveren, Y.; Leka, E.; Kara, A.The significant increase in the number of IoT devices has also brought with it various security concerns. The ability of these devices to collect a lot of data, including personal information, is one of the important reasons for these concerns. The integration of machine learning into systems that can detect security vulnerabilities has been presented as an effective solution in the face of these concerns. In this review, it is aimed to examine the machine learning algorithms used in the current studies in the literature for IoT network security. Based on the authors' previous research in physical layer security, this research also aims to investigate the intersecting lines between upper layers of security and physical layer security. To achieve this, the current state of the area is presented. Then, relevant studies are examined to identify the key challenges and research directions as an initial overview within the authors' ongoing project. © 2024 IEEE.Article Citation - WoS: 2Citation - Scopus: 4Exploiting Visual Features in Financial Time Series Prediction(Igi Global, 2020) Karacor, Adil Gursel; Erkan, Turan ErmanThe possibility to enhance prediction accuracy for foreign exchange rates was investigated in two ways: first applying an outside the box approach to modeling price graphs by exploiting their visual properties, and secondly employing the most efficient methods to detect patterns to classify the direction of movement. The approach that exploits the visual properties of price graphs which make use of density regions along with high and low values describing the shape; hence, the authors propose the name 'Finance Vision.' The data used in the predictive model consists of 1-hour past price values of 4 different currency pairs, between 2003 and 2016. Prediction performances of state-of-the-art methods; Extreme Gradient Boosting, Artificial Neural Network and Support Vector Machines are compared over the same data with the same sets of features. Results show that density based visual features contribute considerably to prediction performance.Conference Object Hybrid AI-Driven Decision Model for Test Automation in Agile Software Development(Institute of Electrical and Electronics Engineers Inc., 2025) Bon, Mohammad; Yazici, AliTest automation plays an essential role in Agile Software Development (ASD), but its implementation remains complex. This study conducts a Systematic Literature Review (SLR) to identify key points of test automation and recent developments in Artificial Intelligence (AI). Based on 21 factors proposed by Butt et al., we construct a three-phase decision-support model addressing software, tools, tests, human, and economic dimensions. To improve this model, modern AI techniques - including natural language processing (NLP), machine learning (ML), Mabl (a self-healing, AI-based test automation tool) and Parasoft Selenic - are used. These technologies automate test case generation, prioritization, and maintenance, aligning with Agile's fast-paced demands. Our proposed hybrid model applies NLP to identify effecting factors, ML for impact scoring, and reinforcement learning (RL) for guiding automation strategies. The goal is to decrease manual processes, improve decision accuracy, and to adapt to evolving requirements. However, challenges such as data quality and the need for AI expertise remain. Future work should focus on practical validation and explore applications in non-functional testing. This study offers a practical, AI-enhanced framework to support Agile teams in streamlining test automation. © 2025 IEEE.Master Thesis Makine Öğrenimi Yardımıyla Akış Optimizasyonu Problemlerinin Hızlandırılması(2025) Güngör, Orhan; Akay, Hasan Umur; Ak, Süleyman EmreHesaplamalı Akışkanlar Dinamiği (HAD) analizleri tasarım iyileştirmelerinde vazgeçilmez olsa da her tasarım noktasında HAD çalışması yürütmek yüksek maliyet ve zaman gerektirir. Bu tezde, HAD tabanlı optimizasyon sürecine makine öğrenmesi (ML) tabanlı yaklaşım modelleri entegre edilerek akış problemlerinin çözüm süresini kısaltmak hedeflenmiştir. Çalışmada önce, literatürde kapsamlı deneysel verisi bulunan Francis-99 türbini için HAD doğrulama analizi yapılmıştır. Ardından türbinin çekiş borusu (draft tube) geometrisi üzerinde biçim optimizasyonu gerçekleştirilmiştir. Optimizasyon sırasında, geleneksel matematiksel yaklaşım modelleri ve farklı ML modelleri (ör. XGBoost, Yapay Sinir Ağları) aynı veri seti üzerinde eğitilerek basınç geri kazanım faktörü (Pressure Recovery Factor) ile performansları kıyaslanmıştır. Sonuçlar, ML modellerinin hâlihazırdaki matematiksel modellere benzer doğruluk sunarken tasarım uzayının taranmasında önemli zaman tasarrufu sağladığını göstermiştir. Ayrıca veri seti hacmi değiştirildiğinde ML yaklaşımlarının esnekliği ve genel başarımı incelenmiş, hangi koşullarda üstünlük sağladıkları ortaya konmuştur.Conference Object Citation - Scopus: 2An Undergraduate Curriculum for Deep Learning(Institute of Electrical and Electronics Engineers Inc., 2018) Tirkes,G.; Ekin,C.C.; Engul,G.; Bostan,A.; Karakaya,M.Deep Learning (DL) is an interesting and rapidly developing field of research which has been currently utilized as a part of industry and in many disciplines to address a wide range of problems, from image classification, computer vision, video games, bioinformatics, and handwriting recognition to machine translation. The starting point of this study is the recognition of a big gap between the sector need of specialists in DL technology and the lack of sufficient education provided by the universities. Higher education institutions are the best environment to provide this expertise to the students. However, currently most universities do not provide specifically designed DL courses to their students. Thus, the main objective of this study is to design a novel curriculum including two courses to facilitate teaching and learning of DL topic. The proposed curriculum will enable students to solve real-world problems by applying DL approaches and gain necessary background to adapt their knowledge to more advanced, industry-specific fields. © 2018 IEEE.Conference Object Autonomous Drone System for Afforestation with Swarm Intelligence(Institute of Electrical and Electronics Engineers Inc., 2025) Yildirim, B.E.; Durmaz, S.; Yildiz, A.; Kucukkomurcu, B.; Ozbek, T.; Türkmen, G.This project explores tree detection and tracking using drones coordinated by swarm intelligence. By enabling autonomous coordination and real-time communication between multiple drones, swarm-based systems significantly enhance area coverage, reduce redundancy, and increase data reliability compared to traditional single-drone approaches. Drones equipped with high-resolution cameras collect aerial imagery, which is then processed through image analysis and machine learning algorithms to identify tree locations accurately. Dynamic task allocation and route optimization enable efficient regional coverage while minimizing error rates. The entire system is developed and evaluated in a simulation environment, allowing for controlled testing and iterative refinement of the swarm behaviors. This framework offers scalable and adaptive solutions for applications in forest conservation, environmental monitoring, and ecosystem management. © 2025 IEEE.
