Browsing by Author "Köse, Abdulkadir"

Now showing 1 - 3 of 3

Biyomedikal Bilgi Tabanları Üzerinde Hesaplamalı İlaç Yeniden Konumlandırması İçin Bilgi Grafiği Temsil Yaklaşımları
(2026) Erkantarcı, Betül; Bakal, Mehmet Gökhan; Köse, Abdulkadir
Drug repositioning, or the strategy of finding new medical applications to an existing drug, is a safer and cheaper alternative to development of a new drug. The thesis constructs a unified biomedical knowledge graph by integrating biomedical information resources such as SemMedDB, repoDB, and UMLS, and presents a comprehensive evaluation of seven knowledge graph embedding models: TransE, TransH, TransR, TransD, TransF, ProjE, and RESCAL. To have a biological relevance, a semantic validation pipeline was created by refining the PubMed-based biomedical language model to an accuracy of around 96% in order to determine the plausibility of the suggested drug-disease associations. TransF had the best quantitative performance with a macro area under the precision-recall curve of 0.767, whereas ProjE produced the largest number of semantically plausible hypotheses with 14 literature-supported drug-disease pairs. Moreover, an automated update system, which leverages internet of things, is used to retrieve fresh PubMed evidence on a daily basis to update semantic plausibility scores. Altogether, the knowledge graph embeddings and semantic validation workflow showed numerous new and literature-based drug-disease relations and indicated its potential to become a strong, explainable and data-driven model in computational drug repositioning. Keywords: Computational Drug Repositioning, Knowledge Graph Embedding, Biomedical Knowledge Graph, Semantic Validation, Internet of Things
Machine Learning Based Beamwidth Adaptation for mmWave Vehicular Communications
(Institute of Electrical and Electronics Engineers Inc., 2023) Manic, Setinder; Heng Foh, Chuan; Köse, Abdulkadir; Lee, Haeyoung; Leow, Chee Yen; Chatzimisios, Periklis; Suthaputchakun, Chakkaphong; Foh, Chuan Heng
The incorporation of mmWave technology in vehicular networks has unlocked a realm of possibilities, propelling the advancement of autonomous vehicles, enhancing interconnectedness, and facilitating communication for intelligent transportation systems (ITS). Despite these strides in connectivity, challenges such as high path-loss have arisen, impacting existing beam management procedures. This work aims to address this issue by improving beam management techniques, specifically focusing on enhancing the service time between vehicles and base stations through adaptive mmWave beamwidth adjustments, accomplished using a Contextual Multi-Armed Bandit Algorithm. By leveraging various conditions to train the ML agent of the Contextual Multi-Armed Bandit Algorithm, it seeks to learn about vehicle mobility profiles and optimize the usage of different antenna beamwidth settings to maximize seamless connection time. The extensive simulation results showcase the effectiveness of an adaptive beamwidth for mobility profiles, extending the connection time a vehicle experiences with a base station when compared to the existing strategies. © 2024 Elsevier B.V., All rights reserved.
Citation - Scopus: 1
Semantic-Forward Relaying for 6G: Performance Boosts With ResNet-18 and GoogleNet Plus
(Institute of Electrical and Electronics Engineers Inc., 2024) Erkantarci, Betul; Çoban, Mert Korkut; Bozoǧlu, Abdulkadir; Köse, Abdulkadir
This paper investigates the integration of advanced deep learning architectures, namely ResNet-18, GoogleNet and enhanced GoogleNet (GoogleNet Plus), into the Semantic-Forward (SF) relaying framework for cooperative communications in 6G networks. The SF relaying framework enhances transmission efficiency and robustness by leveraging semantic information at relay nodes. We analyze and compare the performance of these deep learning models in terms of validation accuracy, semantic accuracy, and Euclidean distance (ED) metrics on the CIFAR-10 dataset. Results indicate that ResNet-18 achieves the highest performance due to its residual learning architecture. GoogleNet Plus, incorporating Automatic Mixed Precision (AMP) training and the Adam optimizer, demonstrates improved stability and efficiency compared to the original GoogleNet. The results highlights the potential of deep learning models to enhance semantic processing capabilities in SF relaying, contributing to the development of more efficient, resilient, and adaptive cooperative communication systems in 6G networks. © 2025 Elsevier B.V., All rights reserved.