Atılım Academic Archive (Atılım University)
Not a member yet
9347 research outputs found
Sort by
Studies of Hadronic Showers in SND@LHC
No full textThe SND@LHC experiment was built for observing neutrinos arising from LHC pp collisions. The detector consists of two sections: a target instrumented with SciFi modules and a hadronic calorimeter/muon detector. Energetic νN collisions in the target produce hadronic showers. Reconstruction of the shower total energy requires an estimate of the fractions deposited in both the target and the calorimeter. In order to calibrate the SND@LHC response, a replica of the detector was exposed to hadron beams with 100 to 300 GeV in the CERN SPS H8 test beam line in Summer 2023. This report describes the methods developed to tag the presence of a shower, to locate the shower origin in the target, and to combine the target SciFi and the calorimeter signals so to measure the shower total energy. © 2025 Elsevier B.V., All rights reserved
Observations on Nist Sp 800-90b Entropy Estimators (Jan, 10.1007/S12095-025-00778-7, 2025)
No full textScience Citation Index Expande
Architectural Development and Analysis of a Deep Learning Model for 3D Medical Image Processing
Get PDFGünümüzde medikal görüntü segmentasyonuna yönelik geliştirilen derin öğrenme modelleri, yüksek doğruluk sunmalarına rağmen; aşırı hesaplama maliyeti, karmaşık yapılar ve donanım bağımlılığı nedeniyle pratik kullanımda çeşitli sınırlılıklar barın-dırmaktadır. Bu doğrultuda, kullanıcı dostu, düşük donanım gereksinimiyle çalışabi-len, sade ancak derin yapıda, sınırlı veri setlerinde de etkili sonuçlar verebilen, genellenebilir ve güçlü mimarilere duyulan ihtiyaç giderek artmaktadır. Bu tezde, herhangi bir fine-tuning veya dışsal optimizasyona ( pruning, quantization, attention vb.) ihtiyaç duymadan, yalnızca yapısal mimari iyileştirmelerle yüksek doğruluk elde eden donanım dostu bir 3B CNN modeli geliştirilmiştir. Model mimarisi kapsamlı biçimde ele alınmış; katman derinliği, filtre boyutu, kanal sayısı, aktivasyon ve normalizasyon sıralaması gibi birçok parametre sistematik olarak analiz edilmiştir. Farklı çekirdek boyutlarına sahip konvolüsyon filtreleri hem paralel yollarla aynı blok içinde, hem de ardışık katmanlar arasında dağıtılarak farklı mimari konfigürasyonlarla yapılandırılmıştır. Bu yapılarda tek ve çok katmanlı, simetrik ve asimetrik tasarımlar denenmiştir. Ayrıca model tasarımı sürecinde NAS (Neural Architecture Search) yöntemi uygulanmış; elde edilen mimari varyantlar performans açısından değerlendirilmiştir. Geliştirilen model, klasik U-Net'e kıyasla eğitim süresini 2.5 ila 10 kat arasında kısaltmış, FLOPs değerini yaklaşık yarı yarıya düşürmüş ve benzer Dice Benzerlik Katsayısı (DSC) ile segmentasyon doğruluğunu korumayı başarmıştır. Ayrıca yapılan analizlerde, FLOPs'un gerçek zamanlı performansı belirlemede tek başına yeterli bir ölçüt olmadığı ortaya konmuştur. Bu tez kapsamında yürütülen çalışmalar, yalnızca mimari düzeyde gerçekleştirilen iyileştirmelerle yüksek doğruluk ve donanım verimliliğine ulaşılabileceğini göstermekte; geliştirilen yapının sade fakat derin mimarisi-yle genellenebilirliği, sınırlı veri setlerinde başarımı ve hangi mimari parametrelerin modele belirgin katkı sağladığı detaylı biçimde ortaya konmuştur.Deep learning based models developed for medical image segmentation achieve high accuracy, yet their practical implementation remains limited due to excessive computational cost, architectural complexity, and dependency on hardware resources. Accordingly, there is an increasing demand for models that are structurally compact, hardware efficient, sufficiently deep, generalizable, and capable of performing effectively even with limited datasets. This thesis presents a 3D convolutional neural network architecture that achieves high segmentation accuracy through architectural modifications alone, without requiring any external optimization techniques. The architecture was thoroughly investigated, and key parameters such as layer, kernel size, number of channels, and the ordering of activation and normalization were systematically analyzed. Convolutional kernels with varying receptive field sizes were employed in different configurations, including parallel branches within the same block and sequential layers across the architecture. The design process involved comprehensive experimentation using neural architecture search through which various architectural configurations were evaluated. Compared to the classical U-Net model, the proposed architecture reduced training time by a factor of 2.5 to 10, halved the number of floating point operations, and maintained comparable segmentation accuracy in terms of the Dice Similarity Coefficient.Furthermore, it was observed that although FLOPs is widely used as a computational cost metric, it does not directly correlate with actual inference time. The findings of this thesis demonstrate that it is possible to achieve high segmentation performance and hardware efficiency through architectural design alone, and offer a detailed investigation into the structural parameters that most significantly affect performance in compact and generalizable CNN models
The Educational Use of Facebook: A Phenomenological Exploration of Faculty Members' and Students' Lived Experiences
No full textSocial media has become a pervasive element of higher education; however, the experiences of faculty and students with its educational use remain underexplored. This study employed a transcendental phenomenological design to investigate the lived experiences of three faculty members and 12 students who used Facebook as a learning tool. Data were collected through in-depth, semi-structured interviews and analyzed using Moustakas' transcendental phenomenological method, which allowed for the identification of the essence of participants' experiences while minimizing researcher bias. Findings indicate that Facebook facilitates communication, supports collaborative learning, and provides flexible access to educational resources, enhancing academic engagement. Participants reported benefits, including immediate information sharing and interactive peer support. However, challenges were also noted, including privacy concerns, potential distractions, and the informal nature of interactions, which may affect structured learning. These findings suggest that while Facebook can supplement formal education meaningfully, it should not replace traditional instruction. The study offers practical guidance on balancing engagement, privacy, and curriculum integration, providing educators with insight into thoughtfully leveraging social media in higher education. By presenting both benefits and limitations, this research contributes to a nuanced understanding of Facebook's educational role and informs strategies for its practical use
Impact of Water Consumption on Structural Members in RC Frames Using Multi-Objective Metaheuristics Algorithm
No full textThe construction industry has a significant global water footprint because buildings incorporate large quantities of embedded materials, such as concrete and steel, whose production consumes substantial amounts of water throughout their life cycle. The grey wolf optimizer (GWO) is particularly suitable for this problem because it is a population-based metaheuristic with strong exploration and exploitation balance, which makes it effective in navigating large discrete search spaces such as structural design variables. GWO has demonstrated robustness in multi-objective problems by efficiently approximating Pareto fronts and avoiding local optima. A numerical structural analysis and design model was developed via an application programming interface. This study is indeed the first optimization of the weight of reinforced concrete (RC) structural elements considering virtual water, with the given structural specifications and using the proposed multi-objective metaheuristic methodology. Results showed the optimal structural weight to be 266 tons, with virtual water usage reaching approximately 253 m3. These findings provide actionable insights for sustainable structural design, guiding material selection and early-stage decision-making to minimize virtual water consumption in RC buildings. This study addresses the research gap by introducing virtual water, alongside structural weight, as a novel objective function within multi-objective metaheuristic optimization of RC frames
Çocuk Şarkılarının Yapay Zekâ Destekli Görselleştirilmesine Yönelik Bir İnceleme: Ali Babanın Çiftliği Örneği
No full textThis study aims to visualize children’s songs, which are part of primary-level music education, using AI-supported tools. The objectives of the Ministry of National Education’s music course curriculum were examined, and both the themes to be emphasized in song selection and the pedagogical functions of children’s songs were analyzed. In the literature review, the Web of Science and Google Scholar databases were used. The obtained source data were analyzed with the VOSviewer software to generate conceptual maps, through which thematic trends in the field were identified. In the practical part of the study, the children’s song “Old MacDonald’s Farm” was visualized in detail using two different AI-supported tools: RunwayML and WZRD.ai. In RunwayML, prompt-based scenes were generated using the “text-to-video” feature, and visuals compatible with the lyrics of the song were created. On the WZRD.ai platform, visuals were automatically generated in response to sound waves, and the limitations of the platform were examined. Based on the findings, it was concluded that RunwayML offers more effective results for pedagogical content production, while WZRD. ai, despite its technical capabilities, falls short in delivering child-appropriate visual stimuli. The study also provides a theoretical foundation on synesthesia and discusses how AI tools can be integrated into music education in early childhood and primary school levels. The findings indicate that AI-supported visualization tools have the potential to provide engaging and flexible educational materials that support learning at the primary school level. It is recommended that teacher training programs develop hands-on modules for these tools, and that future research focus on how these technologies can be adapted to various songs, age groups, and learning domains. © © 2025 The Author(s). This is an open access article distributed under the Creative Commons Attribution License (CC BY), which permits unrestricted use, distribution, and reproduction in any medium or format, provided the original work is properly cited
Impact of Kannan Contraction on Khan Contraction and its Generalizations
No full textCvetkovic, Marija/0000-0003-0691-3428; Rakocevic, Vladimir/0000-0002-4182-4458;Several concepts of rational contractions originated during the last 50 years with vivid discussion on possible applications and practical examples. The main topic of this article is a Khan contraction, its modifications and generalizations, with the emphasis on its relation to some well-known classes of contractive mappings. We prove that a Khan contraction is an example of Bianchini and, consequently, a Ciric contraction. Further, some of its generalizations present a special type of Kannan contraction.Ministry of Education, Science and Technological Development, Republic of Serbia [451-03-65/2024-03/200124]The first author is supported by Ministry of Education, Science and Technological Development, Republic of Serbia, no. 451-03-65/2024-03/200124
Effects of a Parent-Implemented Shared Book Reading Program on Turkish Children: Low-SES Families
No full textİşlek, Sinem/0000-0002-8595-8001The effects of language and literacy skills developed during early childhood on subsequent literacy achievement are well-known. As children's first teachers are their parents, supporting these skills in the home environment is necessary. This study investigated the effect of the Shared Book Reading (SBR) program based on print awareness on children's language and early literacy skills in children from low socioeconomic households. Thirty - four mother-child dyads were randomly assigned to an intervention or a control group. Mothers in the intervention group attended training sessions on using shared book reading strategies through picture books for 12 weeks. The analyses revealed positive effects of the parent-implemented shared book reading program on early literacy skills and the early literacy environment. These findings highlight the potential positive impact of a parent-implemented shared book-reading program on the early literacy experiences and skills of children from low socioeconomic households. The findings show that the mothers in the intervention group were able to transform their children's home education environment into one that valued and contributed to the children's early literacy experiences and had a positive effect on early literacy skills
Mild Solutions for Neutral Conformable Fractional Order Functional Evolution Equations Using Meir-Keeler Type Fixed Point Theorem
No full textOur mission is to demonstrate the existence, uniqueness, attractiveness, and controllability of mild solutions to neutral conformable fractional-order functional evolution equations, specifically of order between 1 and 2. These intriguing equations encompass finite delay, all while adhering to local conditions within a separable Banach space. By invoking Meir-Keeler’s fixed-point Theorem and enhancing it with measures of noncompactness, we establish the existence of these solutions. To highlight the potency of our approach, we present a captivating example. © 2025, Politechnica University of Bucharest. All rights reserved
Otonom Araç Radarları için 79 GHz Fazlı Mikroşerit Anten Dizisinin Besleme Analizi
No full textOtomotiv radarı, güvenilirliği nedeniyle otonom araçlarda umut vadeden bir algılama teknolojisi olarak bilinmektedir. Günümüz otonom araçlarında, 77 – 81 GHz frekans bandı otomotiv radarları için ana çalışma bandıdır. Otomotiv radarlarının verimli çalışabilmesi için radar anteninin son derece hassas olması gerekir. Ancak, yüksek çalışma frekansları, yüksek kazanç, geniş bant genişliği ve düşük yan lob seviyeleri (SLL) gerektiren radar anteni tasarımında zorluklar ortaya çıkarabilmektedir. Bu sorunu ele almak için, bu çalışmada, eş düzlemli boşluk kaynak portu, dikey toprak köprüsü ve dalga portu dahil olmak üzere üç farklı topraklanmış eş düzlemli dalga kılavuzu (GCPW) besleme konfigürasyonu kullanılarak düzlemsel seri beslemeli doğrusal bir anten dizisinin 79 GHz otomotiv radar uygulamalarına uyarlaması amaçlanmaktadır. Antenin besleme yapılandırmalarıyla performansını değerlendirmek için benzetimler yürütülmüştür. Elde edilen sonuçlara göre, dalga portu beslemeli antenin en iyi empedans bant genişliğini (>3 GHz) elde ettiği, eş düzlemli boşluk kaynak portu veya dikey toprak köprüsü konfigürasyonları beslemeli antenin ise daha iyi ana lob faz merkezlemesi ve daha yüksek bir kazanç (>18,4 dBi) sergilediği, yan lob seviyelerinin (SLL) -16,28 dB’nin altında olduğu gösterilmiştir. Bu bulguların, yeni nesil otonom araçlar için yüksek performanslı radar antenlerinin geliştirilmesine katkıda bulunabileceği düşünülmektedir