Scopus İndeksli Yayınlar Koleksiyonu / Scopus Indexed Publications Collection
Permanent URI for this collectionhttps://hdl.handle.net/11147/7148
Browse
1015 results
Search Results
Conference Object A Study on the Influence of Magnetic Nanoparticle Concentration on Heating Efficiency in Magnetic Hyperthermia(Institute of Electrical and Electronics Engineers Inc., 2025) Savranguler, E.N.; Gümüş, S.; Harmansah, C.; Öztürk, Y.; Magat, H.Despite significant advancements in diagnostic and therapeutic technologies, cancer remains one of the most significant global health problems and continues to be among the leading causes of death. In addition to conventional cancer treatments, alternative and innovative methods are being developed for cancer therapies. One of the promising cancer therapies is magnetic hyperthermia, which is based on the principle of heat generation through the Brownian and Néel relaxation mechanisms of magnetic nanoparticles (MNPs) exposure to an alternating magnetic field. In this method, heat is generated by MNPs that are selectively targeted to tumor tissues, resulting in localized cell death. The heating efficiency of MNPs is directly influenced by their physical and chemical properties, such as particle size, magnetic anisotropy, chemical composition, and colloidal stability. Recent studies have shown that magnetic hyperthermia can be effective in tumor reduction when applied alone or in combination with other conventional treatment modalities. In this study, a low-cost and easily assembled magnetic hyperthermia measurement system was employed to investigate the effect of varying nanoparticle concentrations on heating efficiency. The experimental setup consisted of a thermally insulated sample holder, an 88 kHz magnetic induction heater, and a thermometer. EFH-1 magnetic fluid was diluted with hydrocarbon oil at particle volume concentrations ranging from 7.1% (100% EFH-1) to 0.14% (2% EFH1+%98 hydrocarbon oil), and measurements were taken under an applied field of 6.03 kA/m. Based on the experimental data, the rate of temperature change over time was calculated to be in the range of 0.16 K/s to 0.005 K/s. The resulting heating efficiencies, as a function of nanoparticle concentration, were analyzed and discussed by considering previous experimental and theoretical studies. © 2025 IEEE.Conference Object A Comparative Study of Attention-Augmented YOLO Architectures for Defect Detection in Fused Deposition Modelling(Institute of Electrical and Electronics Engineers Inc., 2025) Cezayirli, H.; Tetik, H.; Dede, M.I.C.; Phone, W.L.; Alkan, B.Additive manufacturing (AM), particularly fused deposition modelling (FDM), facilitates the fabrication of complex geometries with increasing flexibility and efficiency. Ensuring consistent print quality in FDM processes necessitates the development of accurate defect detection mechanisms. Attention-augmented YOLO (You Only Look Once) models have emerged as a promising solution for addressing this challenge. In this study, we systematically benchmark and evaluate the performance of YOLO architectures enhanced with attention mechanisms within the context of FDM 3D printing applications. The models were trained and evaluated using representative defect datasets. The attention-augmented models demonstrate improved detection performance. © 2025 IEEE.Conference Object Collabpersona: A Framework for Collaborative Decision Analysis in Persona Driven LLM-Based Multi-Agent Systems(IEEE Computer Society, 2025) Tamer, O.A.; Gumus, A.Large Language Model (LLM) agents have recently demonstrated impressive capabilities in single agent and adversarial settings, but their ability to collaborate effectively with minimal communication remains uncertain. We introduce CollabPersona, a simulation framework that combines persona-grounded memory with one-shot feedback to study team-based reasoning among LLM agents. In a multi-round variant of the Guess 0.8 of the Average game, agents reason entirely through structured prompts without fine-tuning. Our results show that minimal feedback significantly improves intra-team coordination and stabilizes strategic behavior, while cognitive style remains a primary driver of competitive outcomes. These findings suggest that lightweight scaffolding can elicit emergent collaboration in LLM agents and provide a flexible platform for studying cooperative intelligence. © 2025 IEEE.Conference Object Adapting Language Models to Sentiment Analysis for Automatically Translated and Labelled Turkish News Texts(Institute of Electrical and Electronics Engineers Inc., 2025) Serficeli, S.C.; Udunman, B.; Inan, E.The proliferation of news sources makes it difficult to track current events and social events in real time. In order to interpret social events in this context quickly and effectively, it is important to translate news texts provided in different natural languages into Turkish and to perform sentiment analysis on them. The aim of this study is to translate multilingual news texts into Turkish and perform sentiment analysis on these texts. The generated labels were compared and the data that were given the same label by all models were separated as automatically labelled data. This automatic labelling process ensured that the data for which different models produced consistent results were reliably labelled. When the results were evaluated, F1 score of 0.946 was achieved for sentiment analysis using the automatic labelling mechanism for texts translated into Turkish. © 2025 IEEE.Conference Object LoRA+ID: Enhancing Identity Preservation in Generative Models Through Face-Conditioned LoRA Training(Institute of Electrical and Electronics Engineers Inc., 2025) Ozbay, K.; Ulusoy, A.E.; Baştanlar, Y.Low-Rank Adaptation (LoRA) has become the standard approach for fine-tuning large-scale generative models like Stable Diffusion XL (SDXL), offering efficiency in compute and memory. However, traditional LoRA methods rely solely on text prompts, limiting their ability to preserve detailed identity features. In this work, we propose a novel training framework, LoRA+ID, that integrates face embeddings-derived from face recognition networks-into the LoRA training loop. Unlike methods such as FaceID or InstantID, which introduce image conditioning only at inference time, our approach conditions LoRA directly on facial features during training. We evaluate our method across four setups involving 8 identities and 18 generations per identity. Experimental results show that LoRA+ID, especially when used with FaceID during inference, significantly improves identity preservation compared to both traditional LoRA and zero shot FaceID generation. © 2025 IEEE.Conference Object A Semantic Search Engine for Turkish and English Research Resources(Institute of Electrical and Electronics Engineers Inc., 2025) Karabacak, O.; Inan, E.Research resources are growing in volume at an exponential rate across disciplines and languages. This exponential increase has created a pressing need for intelligent search systems that can help researchers efficiently access relevant academic material. To overcome this issue, this study introduces a bilingual semantic search engine designed to retrieve academic articles written in both Turkish and English. The primary goal is to improve the accuracy and relevance of academic information retrieval by using modern Natural Language Processing techniques. Instead of relying on traditional keyword-based search methods, the system leverages transformer-based sentence embedding models. To capture semantic meaning more effectively, MiniLM-L6v2, paraphrase-multilingual-MiniLM-L12-v2 and multilingual-e5-base models were chosen for their multilingual capabilities and sentence-level embedding performance. To assess the quality of search results, Mean Average Precision (MAP) and Normalized Discounted Cumulative Gain (nDCG) were used. These metrics were calculated for each model across both language groups. Evaluation results show that the multilingual-e5-base model consistently outperformed the other models in both MAP and nDCG scores, demonstrating superior semantic understanding and multilingual alignment. The system also features a simple and responsive Streamlit-based interface that allows for real-time querying and result display. © 2025 IEEE.Conference Object Investigation of a Magnetic Levitation Density Measurement System(Institute of Electrical and Electronics Engineers Inc., 2025) Gümüş, S.; Öztürk, Y.Magnetic levitation systems are widely used for density measurements in biomedical research and sensor technologies. These systems consist of pairs of magnets with like poles facing each other, creating a repulsive magnetic field. Levitation occurs as diamagnetic particles are suspended in a paramagnetic fluid between the magnets. The force acting on the particles is proportional to the product of the magnetic field and its gradient, while the concentration of the paramagnetic fluid influences the magnitude of the force. To optimize sensor performance, both magnetic field strength and paramagnetic ion concentration must be considered. In this study, two magnets in an anti-Helmholtz configuration (62 × 3 × 12 mm) were used, with a variable gap distance (g). Experimental analysis was conducted to investigate the effect of magnetic field strength and fluid concentration on levitation behavior. Initially, g was set to 1.8 mm, and Gadolinium-based paramagnetic fluid (Gadovist) was prepared at 30 mM, 45 mM, and 60 mM concentrations. Microplastic particles with densities of 1.05 g/cc and 1.09 g/cc were added into the solutions. Levitation heights, measured relative to the bottom magnet, increased with concentration: 0.60-0.51 mm and 0.43-0.39 mm at 30 mM; 0.70-0.66 mm and 0.49-0.47 mm at 45 mM; and 0.76-0.71 mm and 0.63-0.61 mm at 60 mM for 1.05 g/cc and 1.09 g/cc particles, respectively. In the second stage, g was increased to 2.4 mm and 3 mm using 60 mM fluid. Levitation heights were 0.69-0.68 mm and 0.55-0.49 mm at 2.4 mm; and 0.65-0.64 mm and 0.48-0.47 mm at 3 mm, respectively. These results were compared with theoretical calculations, and sensor performance was evaluated for different application scenarios, contributing to the development of future levitation-based sensing systems. © 2025 IEEE.Conference Object Analysis and Optimization of Spiking Neural Network Simulations on GPUs(Institute of Electrical and Electronics Engineers Inc., 2025) Kocatepe, Doǧu; Öz, IşilArtificial neural networks (ANNs) have experienced remarkable growth over the past 20 years, propelled by advancements in GPU devices and the creation of accessible tools for constructing complex models. Concurrently, another type of neural network, spiking neural networks (SNNs), has been under development, particularly within neuroscience. SNNs, which are inspired by the brain, mimic biological nervous systems and incorporate neuronal dynamics described by differential equations. Recently, SNNs have gained traction in machine learning due to their potential for energy efficiency and relevance to discussions on artificial general intelligence (AGI). Over the past few decades, substantial research has been conducted on SNN simulation, leading to significant performance improvements, primarily through GPU-based implementations. However, there is a lack of benchmark studies involving the most recent GPU-based SNN simulators. In this work, we systematically analyze the state-of-the-art GPU-based SNN simulations and develop a novel simulator, cuSpike, which outperforms existing simulators on selected benchmark models by preserving accurate simulation results and providing energy-efficient executions. © 2025 Elsevier B.V., All rights reserved.Conference Object Strengthening of Reinforced Concrete Columns Using Recycled Polyethylene Terephthalate Fibers: A Preliminary Numerical Study(fib. The International Federation for Structural Concrete, 2025) Dalgic, K.D.; Gozun, U.; Simsek, B.; Sencar, I.; Ispir, M.; Ilki, A.Strengthening of reinforced concrete (RC) columns, which have inadequate capacities of deformation and axial/lateral load, using carbon fiber reinforced polymers (CFRP) has become widespread. However, concerns about cost, energy sustainability and environmental impact have led to increased interest in alternative fibers, such as recycled polyethylene terephthalate (PET) fibers, instead of high-tech, carbon-intensive materials. This study presents preliminary numerical analyses on the use of PET fibers recycled from tire industry waste in Türkiye to strengthen substandard RC columns. The numerical analyses of the column models were performed under axial and horizontal loads. The results show that even small amounts of PET-FRP fibers can significantly improve both lateral load and deformation capacities of RC column, indicating the development of strengthening strategies for upcoming column tests. Based on the results of numerical studies, an experimental program for quasistatic testing of substandard RC columns has been planned. © 2025 Elsevier B.V., All rights reserved.Conference Object Measuring the Size of Change Requests in Microservice-Based Software Projects(Springer Science and Business Media Deutschland GmbH, 2026) Yenel, M.; Ünlu, H.; Demirors, O.Accurately estimating the effort required for implementing change requests remains a critical challenge in software engineering, especially in microservice-based software architectures (MSSA). Traditional functional size measurement methods often fail to capture the distinct characteristics of MSSAs. To address this limitation, we propose a change size measurement method based on MicroM, a size measurement approach specifically developed for MSSAs. The proposed method counts added, deleted, and modified events across functional, architectural, and algorithmic levels, and includes the number of affected initial requirements. We conducted an exploratory case study with 18 change requests and built four regression-based effort estimation models. The results show that combining event counts with the number of affected requirements improves estimation accuracy. Our method provides a more precise and context-aware way to estimate change-related effort in MSSA projects. © 2025 Elsevier B.V., All rights reserved.