Graduate School of Informatics - PhD Thesis http://ii.metu.edu.tr/research-news-categories/phd-thesis Ph.D. Thesis en Fatih Ömrüuzun, A Novel Content-Based Retrieval System for Hyperspectral Remote Sensing Imagery http://ii.metu.edu.tr/fatih-omruuzun-novel-content-based-retrieval-system-hyperspectral-remote-sensing-imagery <div class="field-body"> <p class="MsoNormal" style="margin-top: .25in; margin-right: 0in; margin-bottom: .25in; margin-left: 0in; line-height: normal; background: white;"><b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;">Ph.D. Candidate:</span></b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;"> Fatih Ömrüuzun<br /> <b>Program:</b> Information Systems<br /> <b>Date:</b> 24.01.2024 <span style="border: none windowtext 1.0pt; mso-border-alt: none windowtext 0in; padding: 0in;">/ 14:00</span><br /> <b>Place:</b> B-116<p></p></span></p> <p class="MsoNormal"><b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR">Abstract:</span></b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR"> </span><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black; mso-ansi-language: TR;" xml:lang="TR">Due to the increased use of hyperspectral remote sensing payloads, there has been a rise in the number of hyperspectral remote sensing image archives, resulting in a massive amount of collected data. This highlights the need for a content-based image retrieval system that can manage and enable the use hyperspectral remote-sensing images efficiently. The conventional content-based hyperspectral image retrieval (CBHIR) systems define each image by a set of endmembers and then perform image retrieval using pairwise distance measures. However, this approach significantly increases the computational complexity of retrieval, especially when there is a high diversity of materials. Additionally, those systems have difficulties in retrieving images with particular materials whose abundance is extremely low compared to other materials or those that are not considered as an endmember while modeling the image. To address these issues, a novel CBHIR system is proposed that aims to define global hyperspectral image representations based on a semantic approach to differentiate background and foreground image content considering both spatial and spectral information. In this way, two spectral content dictionaries are used in the process of modeling hyperspectral images. While the first dictionary originates in spectral terms related to materials that are rarely encountered in the relevant geographical region, called foreground content, the second dictionary contains spectral terms for materials that are commonly seen in the geographical region, called background content. The proposed system consists of two main modules. The first module characterizes the hyperspectral images in the archive by four global descriptors: 1) two binary spectral descriptors (which represent spectral characteristics of distinct foreground and background materials); 2) two abundance descriptors that model the normalized cumulative fractional abundance of the corresponding materials. The second module retrieves hyperspectral images from the archive that either cover materials that are most similar to the given query signature or query image based on a hierarchical strategy that evaluates the spectral and abundance descriptor similarity. Experiments conducted on a benchmark dataset of hyperspectral images demonstrated the system's effectiveness in terms of retrieval accuracy and time.<p></p></span></p> <p><img src="/system/files/tez_web.png" width="1000" height="750" /></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/fatih-omruuzun-novel-content-based-retrieval-system-hyperspectral-remote-sensing-imagery" title="Fatih Ömrüuzun, A Novel Content-Based Retrieval System for Hyperspectral Remote Sensing Imagery" class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Tue, 23 Jan 2024 11:49:43 +0000 wwwii 1282 at http://ii.metu.edu.tr Hatice Gonca Bulur, Analyzing Decision Making Behaviour Under Risk and Uncertainty with The Help of Computational Cognitive Modeling and Neuroscience Perspectives http://ii.metu.edu.tr/hatice-gonca-bulur-analyzing-decision-making-behaviour-under-risk-and-uncertainty-help-computational <div class="field-body"> <p class="MsoNormal" style="margin-top: .25in; margin-right: 0in; margin-bottom: .25in; margin-left: 0in; line-height: normal; background: white;"><b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;">Ph.D. Candidate:</span></b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;"> Hatice Gonca Bulur<br /> <b>Program:</b> Cognitive Science<br /> <b>Date:</b> 26.01.2024 <span style="border: none windowtext 1.0pt; mso-border-alt: none windowtext 0in; padding: 0in;">/ 11:00</span><br /> <b>Place:</b> A-212<p></p></span></p> <p class="MsoNormal"><b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR">Abstract:</span></b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR"> </span><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black; mso-ansi-language: TR;" xml:lang="TR">People make decisions in almost every moment of their lives. Therefore, understanding the fundamentals of decision making behaviour receives a lot of interest. This study aims to understand individuals’ decision making behaviour under risk and uncertainty by bringing insights from computational cognitive modeling and neuroscience perspectives. Two functional near-infrared spectroscopy (fNIRS)-based Balloon Analogue Risk Task (BART) experiments (perfect gambling, probability learning) are conducted to obtain measures related to risk-taking behaviour and risk probability learning. The collected data from the participants is analyzed with relevant neural and statistical techniques. Behavioural data analysis results indicate significant differences between balloon colours (explosion probabilities), two conditions (gambling vs probability learning) and balloon presentation orders (constant, mix) on the number of pumps, total points earned and adjusted average number of pumps. The mean value of the gambling condition is higher than the probability learning case for all measures, which indicates higher risk taking. fNIRS data analysis findings show that the difference between two conditions for mixed blocks is the strongest at the the right dorsolateral prefrontal cortex (dlPFC) due to more cognitive challenge in the perfect gambling case. The separation between conditions is mostly in the balloon with the highest explosion probability. The acquired data is benefited to jointly model participants’ intended number of pumps in BART so that an explanation for the projections of the fNIRS data can be ensured from the model.<p></p></span></p> <p><img src="/system/files/websitesiimage.png" width="493" height="398" /></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/hatice-gonca-bulur-risk-ve-belirsizlik-altinda-karar-verme-davranislarinin-hesaplamali-bilissel" title="Hatice Gonca Bulur, Risk ve Belirsizlik Altında Karar Verme Davranışlarının Hesaplamalı Bilişsel Modelleme ve Sinirbilim Perspektifleri Yardımıyla Analiz Edilmesi" class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Tue, 23 Jan 2024 09:08:39 +0000 wwwii 1280 at http://ii.metu.edu.tr Utku Civelek, The Conceptual Design and Implementation of a Knowledge Management System for Collaborative Data Science http://ii.metu.edu.tr/utku-civelek-conceptual-design-and-implementation-knowledge-management-system-collaborative-data <div class="field-body"> <p class="MsoNormal" style="margin-top: .25in; margin-right: 0in; margin-bottom: .25in; margin-left: 0in; line-height: normal; background: white;"><b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;">Ph.D. Candidate:</span></b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;"> Utku Civelek<br /> <b>Program:</b> Information Systems<br /> <b>Date:</b> 22.01.2024 <span style="border: none windowtext 1.0pt; mso-border-alt: none windowtext 0in; padding: 0in;">/ 13:00</span><br /> <b>Place:</b> II-06<p></p></span></p> <p class="MsoNormal"><b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR">Abstract:</span></b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR"> </span><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black; mso-ansi-language: TR;" xml:lang="TR">The new industrial revolution heavily depends on digital transformation. Companies strive to ameliorate or redesign their business processes with the help of digital technologies to preserve or attain competitive advantage. The most interactive field of digital transformation is data science, as it entails a longtime active collaboration among multiple partners. On the one hand, data scientists seek domain expertise to understand the structure and environment of the data to be analyzed. On the other hand, business users and managers take pains with data science concepts and sample use cases, to exploit analytical capabilities and opportunities in their organizations. To this end, the extent of collaboration and knowledge share affects the success of data science projects. Nevertheless, the existing literature is dramatically limited in proposing a comprehensive solution to assist organizations for this scope. This thesis presents the conceptual design and implementation of CoDS (Collaborative Data Science Framework), a knowledge management system for consolidating data science activities in an enterprise, to address these important challenges. The CoDS presents a platform on which business details, data-related knowledge, modeling procedures, and deployment steps are shared for (i) mediating and scaling ongoing projects, (ii) enriching knowledge transfer among stakeholders, (iii) facilitating ideation of new products, and (iv) supporting the onboarding of new employees and developers. A case study is formed to evaluate its ease of creation, usefulness, attraction, and comprehensibility. Accordingly, this study proposes a novel structure and a roadmap for creating a data science knowledge management system for the collaboration of all stakeholders.<p></p></span></p> <p><img src="/system/files/utku_civelek_-_tez_tanitim_imgesi.png" width="960" height="960" /></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/utku-civelek-veri-biliminde-birligi-icin-bir-bilgi-yonetim-sisteminin-kavramsal-tasarim-ve" title="Utku Civelek, Veri Biliminde İş Birliği için Bir Bilgi Yönetim Sisteminin Kavramsal Tasarım ve Uygulaması" class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Wed, 17 Jan 2024 13:04:44 +0000 wwwii 1268 at http://ii.metu.edu.tr Umut Şener, Development of a Maturity Index for Digital Transformation in Organizations http://ii.metu.edu.tr/umut-sener-development-maturity-index-digital-transformation-organizations <div class="field-body"> <p class="MsoNormal" style="margin-top: .25in; margin-right: 0in; margin-bottom: .25in; margin-left: 0in; line-height: normal; background: white;"><b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;">Ph.D. Candidate:</span></b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;"> Umut Şener<br /> <b>Program:</b> Information Systems<br /> <b>Date:</b> 22.01.2024 <span style="border: none windowtext 1.0pt; mso-border-alt: none windowtext 0in; padding: 0in;">/ 14:30</span><br /> <b>Place:</b> II-06<p></p></span></p> <p class="MsoNormal" style="text-align: justify;"><b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR">Abstract:</span></b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR"> </span><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black; mso-ansi-language: TR;" xml:lang="TR">Organizations aim to maximize the benefits of digital transformation (DX) by deploying connected, intelligent, and self-governed systems that leverage diverse technologies, such as the Internet of Things (IoT). In order to stay competitive and succeed in the market, they need to evaluate their organizational DX maturity. Maturity structures, such as maturity index, provide a systematic and structured approach for evaluating maturity and benchmarking with other organizations. However, the literature review reveals limitations in existing DX maturity structures, including restricted public access, insufficient granularity, reliance on domain experts, and subjectivity in assessment, indicating a research gap. To address this gap, this thesis introduces a novel DX maturity index called DX-MI, using design science research. The DX-MI is a self-diagnostic tool that helps organizations measure their DX maturity and guides them toward achieving higher levels of maturity. The tool has a hierarchical structure that includes dimensions, sub-dimensions, and measurement instruments, all underpinned by an assessment approach grounded in evidence or objective quantifiable metrics. Additionally, it employs the Analytical Hierarchy Process (AHP) to determine the weights of its components, resulting in a more accurate measurement. Multiple case studies were conducted to check the applicability and usability of the DX-MI. The results indicate that it provides a rigorous and practical methodology for measuring DX maturity in organizations.<p></p></span></p> <p><img src="/system/files/umut_sener.png" width="868" height="700" /></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/umut-sener-kuruluslardaki-dijital-donusum-icin-bir-olgunluk-indeksinin-gelistirilmesi" title="Umut Şener, Kuruluşlardaki Dijital Dönüşüm İçin Bir Olgunluk İndeksinin Geliştirilmesi" class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Wed, 17 Jan 2024 12:59:10 +0000 wwwii 1266 at http://ii.metu.edu.tr Özgür Korkmaz, Hyperspectral Imaging Applications for Steel Production http://ii.metu.edu.tr/ozgur-korkmaz-hyperspectral-imaging-applications-steel-production <div class="field-body"> <p class="MsoNormal" style="margin-top: .25in; margin-right: 0in; margin-bottom: .25in; margin-left: 0in; line-height: normal; background: white;"><b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;">Ph.D. Candidate:</span></b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;"> Özgür Korkmaz<br /> <b>Program:</b> Information Systems<br /> <b>Date:</b> 11.01.2024 / 09:30<br /> <b>Place:</b> B-116<p></p></span></p> <p style="text-align: justify;"><b><span lang="TR" style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR">Abstract:</span></b><span lang="TR" style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-ansi-language: TR;" xml:lang="TR"> <span style="color: black;">Steel production serves as the backbone of countless infrastructure projects and industrial applications worldwide. In order to maintain and improve its productivity, quality and environmental sustainability, the steel industry is constantly looking for innovative technologies and methods. Hyperspectral imaging is a promising technology in this context. A novel, non-destructive approach is presented to quantify the free lime content in steel slag by utilizing an integrated algorithm applied to hyperspectral images. This method includes spectral unmixing for mixture component quantification and endmember extraction of mixture. The methodology involved various experiments with both fresh and six-month-aged steel slag, demonstrating its accuracy compared to the Rietveld Analysis of X-ray Diffraction patterns, with an RMSE of 5.57% for aged slag and 6.51% for fresh slag. Furthermore, the thesis explores the application of hyperspectral imaging in identifying foreign substances in iron ore and detecting copper accumulations in continuous pickling lines. As a result of the experiments, the impurities in the raw iron ore were clearly identified and copper accumulation areas were detected at various sensitivities on the steel sheet produced in the continuous pickling line process. An in-depth analysis of the challenges and limitations associated with the use of hyperspectral imaging in steelmaking are provided. The findings of this research advance the technological capabilities in steel production.</span></span><span lang="TR" style="font-family: 'TimesNewRomanPSMT',serif; mso-no-proof: yes;" xml:lang="TR"> </span><span style="mso-no-proof: yes;"><p></p></span></p> <p><img src="/system/files/ozgur_korkmaz_tez.png" width="1415" height="430" /></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/ozgur-korkmaz-celik-uretimi-icin-hiperspektral-goruntuleme-uygulamalari" title="Özgür Korkmaz, Çelik Üretimi içı̇n Hiperspektral Görüntüleme Uygulamaları" class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Mon, 08 Jan 2024 11:28:14 +0000 wwwii 1256 at http://ii.metu.edu.tr Müslüm Kaan Arıcı, Uncovering Hidden Connections and Functional Modules via pyPARAGON: A Hybrid Approach for Network Contextualization http://ii.metu.edu.tr/muslum-kaan-arici-uncovering-hidden-connections-and-functional-modules-pyparagon-hybrid-approach <div class="field-body"> <p class="MsoNormal" style="margin-top: .25in; margin-right: 0in; margin-bottom: .25in; margin-left: 0in; line-height: normal; background: white;"><b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;">Ph.D. Candidate:</span></b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;"> Müslüm Kaan Arıcı<br /> <b>Program:</b> Medical Informatics<br /> <b>Date:</b> 22.01.2024 / 14:00<br /> <b>Place:</b> A-108<p></p></span></p> <p class="MsoNormal"><b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR">Abstract:</span></b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR"> <span style="color: black;">State-of-the-art omics technologies provide molecular insights into various biological contexts, such as disease states, patients, and drug perturbations. Network inference and reconstruction methods use several omics datasets to build contextualized networks with specific biomolecular interactions and activities. In this thesis, we conducted a benchmark analysis to identify deficiencies in reference networks, comparing the coverage of reference networks across various types of knowledge, such as pathways, three-dimensional structure, and publication counts. Additionally, we examined the limitations of reconstruction algorithms by inferring signaling pathways. Contextualized network inference has several challenging issues: i) Hits of omics datasets are sparse in reference networks. ii) Interpretation methods can miss hidden knowledge that connects significant hits in omics datasets while evaluating multi-omics datasets. iii) Well-studied proteins in reference networks come along with bias in contextualization. iv) Highly connected nodes, or hubs, are driven to unspecific and noisy interactions in inferred networks. We developed pyPARAGON (PAgeRAnk-flux on Graphlet-guided network for multi-Omics data integratioN) to address these challenges, we developed pyPARAGON (PAgeRAnk-flux on Graphlet-guided network for multi-Omics data integratioN), combining network propagation with graphlets to integrate multi-Omics data. pyPARAGON improves precision and reduces the presence of non-specific interactions in signaling networks by using network motifs in the benchmark tests: reconstruction of cancer signaling pathways and contextualized cancer models. Moreover, pyPARAGON has promising performance in case studies such as tumor-specific networks with significant biological processes and pathways, shared pathways, and different signal strengths in cancer and neurodevelopmental disorders models based on contextualized networks.</span></span><span style="font-size: 9.0pt; mso-bidi-font-size: 11.0pt; line-height: 107%; font-family: 'Helvetica Neue';"><p></p></span></p> <p><img src="/system/files/arici_tez.png" width="1385" height="851" /></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/muslum-kaan-arici-gizli-etkilesimler-ve-fonksiyonel-modullerin-hibrit-bir-ag-baglamsallastirma-araci" title="Müslüm Kaan Arıcı, Gizli Etkileşimler ve Fonksiyonel Modüllerin Hibrit Bir Ağ Bağlamsallaştırma Aracı pyPARAGON ile Açığa Çıkarılması" class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Fri, 05 Jan 2024 13:17:30 +0000 wwwii 1254 at http://ii.metu.edu.tr Alper Sarıkaya, A Robust Machine Learning Based IDS Design Against Adversarial Attacks in SDN http://ii.metu.edu.tr/alper-sarikaya-robust-machine-learning-based-ids-design-against-adversarial-attacks-sdn <div class="field-body"> <p class="MsoNormal" style="margin-top: .25in; margin-right: 0in; margin-bottom: .25in; margin-left: 0in; line-height: normal; background: white;"><b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;">Ph.D. Candidate:</span></b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;"> Alper Sarıkaya<br /> <b>Program:</b> Information Systems<br /> <b>Date:</b> 17.01.2024 / 14:00<br /> <b>Place:</b> A-108<p></p></span></p> <p class="MsoNormal"><b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR">Abstract:</span></b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR"> <span style="color: black;">Machine learning-based intrusion detection systems (IDS) are essential security functions in conventional and software-defined networks alike. Their success and the security of the networks they protect depend on the accuracy of their classification results. Adversarial attacks against machine learning, which seriously threaten any IDS, are still not countered effectively. In this thesis, firstly, a method is designed that employs generative adversarial networks to produce adversarial attack data. Then, RAIDS, a robust IDS model, is proposed which is resilient against adversarial attacks. In RAIDS, an autoencoder’s reconstruction error is used as a prediction value for a classifier. Also, to prevent the attacker from guessing about the feature set, multiple feature sets are created and used to train baseline machine learning classifiers. A LightGBM classifier is then trained with the results produced by two autoencoders and an ensemble of baseline machine learning classifiers. The results show that the proposed robust model can increase overall accuracy by at least 13.2% and F1-score by more than 110% against adversarial attacks without the need for adversarial training.</span></span><span style="font-size: 9.0pt; mso-bidi-font-size: 11.0pt; line-height: 107%; font-family: 'Helvetica Neue';"><p></p></span></p> <p><img src="/system/files/alper_sarikaya_tez.png" width="3520" height="1604" /></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/alper-sarikaya-yazilim-tanimli-agda-yaniltici-saldirilara-karsi-makine-ogrenimi-tabanli-direncli-bir" title="Alper Sarıkaya, Yazılım Tanımlı Ağda Yanıltıcı Saldırılara Karşı Makine Öğrenimi Tabanlı Dirençli Bir Saldırı Tespit Sistemi Tasarımı" class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Thu, 28 Dec 2023 10:57:37 +0000 wwwii 1244 at http://ii.metu.edu.tr Serkan Özdemir, Development of a Decision-Support Tool for Managing Drinking Water Reservoir by Using Machine Learning and Deep Learning Methods http://ii.metu.edu.tr/serkan-ozdemir-development-decision-support-tool-managing-drinking-water-reservoir-using-machine <div class="field-body"> <p class="MsoNormal" style="margin-top: 18.0pt; margin-right: 0cm; margin-bottom: 18.0pt; margin-left: 0cm; line-height: normal; background: white;"><b><span lang="EN-US" style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;" xml:lang="EN-US">Ph.D. Candidate:</span></b><span lang="EN-US" style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;" xml:lang="EN-US"> Serkan Özdemir<br /> <b>Program:</b> Information Systems<br /> <b>Date:</b> 19.12.2023 / 13:30<br /> <b>Place:</b> A-212<p></p></span></p> <p class="MsoNormal" style="text-align: justify;"><b><span style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;">Abstract:</span></b><span style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;"> <span style="color: black;">Global climate change has led to large fluctuations in lake levels in recent years, due to both changing meteorological parameters and intensive water use. A shift in input or output variables can easily alter the water balance equation and move water levels in the opposite direction. To understand the continuing trend and to create an action plan for dramatic water balance and water quality management, scientists use a variety of models to analyze several variables recorded. In this thesis, the predictive models used for the climatic and hydrologic variables are discussed and their relationships to Lake Water Level (LWL) and water quality are presented. Based on the technological progress, three different types of algorithms a) Naive Method, b) Artificial Neural Networks (ANN) and finally c) Recurrent Neural Network (RNN) models are used to predict water level in lakes. The prediction results from the thesis show that Long Short Term Memory (LSTM) has the highest accuracy with respect to the Root Mean Squared Error (RMSE) evaluation metric. The models were also compared with the performance of the Naïve Method, and the results show that ANN and RNN algorithms are superior in prediction accuracy as the prediction horizon increases. The prediction performances were assessed with Diebold Mariano Test to decide significant differences. It also reveals the water quality of the lake is highly correlated with temperature and evaporation. The models and evaluation metrics are constructed to build a prototype of decision support tool in order water managers to use in operational transactions.</span></span><span lang="EN-US" style="font-size: 9.0pt; mso-bidi-font-size: 11.0pt; line-height: 107%; font-family: 'Helvetica Neue';" xml:lang="EN-US"><p></p></span></p> <p><img src="/system/files/serkan_tez_imge.png" width="728" height="728" /></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/serkan-ozdemir-makine-ogrenmesi-ve-derin-ogrenme-yontemlerin-kullanarak-icme-suyu-reservuarinin" title="Serkan Özdemir, Makine Öğrenmesi ve Derin Öğrenme Yöntemlerin Kullanarak İçme Suyu Reservuarının Yönetimi için Bir Karar Destek Aracının Geliştirilmesi" class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Wed, 06 Dec 2023 16:00:23 +0000 wwwii 1242 at http://ii.metu.edu.tr Burak Demiralay, Efficient Primer Design for Genotype and Subtype Detection of Highly Divergent Viruses in Large Scale Genome Datasets http://ii.metu.edu.tr/burak-demiralay-efficient-primer-design-genotype-and-subtype-detection-highly-divergent-viruses <div class="field-body"> <p></p> <p class="MsoNormal" style="margin: 0.25in 0in; line-height: normal; background-image: initial; background-position: initial; background-size: initial; background-repeat: initial; background-attachment: initial; background-origin: initial; background-clip: initial;"><b><span style="font-size: 10pt; font-family: Verdana, sans-serif;">Ph.D. Candidate:</span></b><span style="font-size: 10pt; font-family: Verdana, sans-serif;"> Burak Demiralay<br /> <b>Program:</b> Health Informatics<br /> <b>Date:</b> 11.09.2023 / 17:00<br /> <b>Place:</b> A-212<p></p></span></p> <p class="MsoNormal" style="margin: 0.25in 0in; line-height: normal; background-image: initial; background-position: initial; background-size: initial; background-repeat: initial; background-attachment: initial; background-origin: initial; background-clip: initial;"></p> <p class="MsoNormal"><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR"></span></p> <p class="MsoNormal"><b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR">Abstract:</span></b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR"> <span style="color: black;">Identification of microorganisms is a crucial step in diagnostics, pathogen screening, biomedical research, evolutionary studies, agriculture, and biological threat assessment. While progress has been made in studying larger organisms, there is a need for an efficient and scalable method that can handle thousands of whole genomes for organisms with high mutation rates and genetic diversity such as single stranded viruses. In this study, we developed a method to extract sequences that would detect the presence of a given species/subspecies using the PCR method. Species detection in any analysis depends highly on the measurement method and since thermodynamic interactions are critical in PCR, thermodynamics is the main driving force in the proposed methodology. We applied our method to three highly divergent viruses; 1) HCV, where the subtypes differ in 31%-33% of nucleotide sites on the average, 2) HIV, for which, 25-35% between-subtype and 15-20% within-subtype variation is observed, and 3) the Dengue virus, whose respective genomes (only DENV 1–4) share 60% sequence identity to each other. Using our method, we were able to select oligonucleotides that can identify in silico 99.9% of 1657 HCV genomes, 99.7% of 11838 HIV genomes, and 95.4% of 4016 Dengue genomes. We also show subspecies identification on genotypes 1-6 of HCV and genotypes 1-4 of the Dengue virus with &gt;99.5% true positive and &lt;0.05% false positive rate, on average. None of the state- of-the-art methods can produce oligonucleotides with this specificity and sensitivity on highly divergent viral genomes like the ones we studied in this thesis.<a name="_Hlk145075026" id="_Hlk145075026"></a><p></p></span></span></p> <p class="MsoNormal"><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR"><span style="color: black;"><img src="/system/files/demiralay_imge.png" width="1200" height="630" /></span></span></p> <p></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/burak-demiralay-buyuk-olcekli-genom-veri-kumelerinde-farklilasma-orani-yuksek-viruslerin-genotip-ve" title="Burak Demiralay, Büyük Ölçekli Genom Veri Kümelerinde Farklılaşma Oranı Yüksek Virüslerin Genotip ve Alttip Tespiti için Etkili Primer Tasarımı" class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Fri, 08 Sep 2023 11:33:03 +0000 wwwii 1210 at http://ii.metu.edu.tr Mine Yoldaş Orhon, MutEXP: A Tool to Identify SNPs That Affect Gene Expression http://ii.metu.edu.tr/mine-yoldas-orhon-mutexp-tool-identify-snps-affect-gene-expression <div class="field-body"> <p class="MsoNormal" style="margin-top: .25in; margin-right: 0in; margin-bottom: .25in; margin-left: 0in; line-height: normal; background: white;"><b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;">Ph.D. Candidate:</span></b><span style="font-size: 10.0pt; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; color: black;"> Mine Yoldaş Orhon<br /><b>Program:</b> Medical Informatics<br /><b>Date:</b> 11.09.2023 / 14:00<br /><b>Place:</b> A-212<p></p></span></p> <p class="MsoNormal"><b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR">Abstract:</span></b><span lang="TR" style="font-size: 10.0pt; line-height: 107%; font-family: 'Verdana',sans-serif; mso-fareast-font-family: 'Times New Roman'; mso-bidi-font-family: 'Times New Roman'; mso-ansi-language: TR;" xml:lang="TR"> <span style="color: black;">Most of the variants in the genome are at the non-coding region. While variations in the coding region effect the protein, variations in non-coding region effect the regulatory mechanism. Therefore, observation of non-coding variations may ensure to identify variations that effect gene expression. eQTL is a popular method used for the purpose to determine the SNPs that effect the gene expression. We have implemented a python based, easy-to-use tool to understand the relationship between the somatic SNPs and gene expression based on eQTL analysis.</span></span><span style="font-size: 9.0pt; mso-bidi-font-size: 11.0pt; line-height: 107%; font-family: 'Helvetica Neue';"><p></p></span></p> <p><img src="/system/files/mine_yoldas.png" width="1700" height="1400" /></p> </div> <div class="form-item form-type-item"> <label>Language </label> English </div> <h3 class="field-label"> Research News Category </h3> <div class="field-research-news-category"> <a href="/research-news-categories/phd-thesis" typeof="skos:Concept" property="rdfs:label skos:prefLabel">PhD Thesis</a> </div> <ul class="links inline"><li class="translation_tr first last"><a href="/tr/mine-yoldas-orhon-mutexp-gen-eksprasyonunu-etkilene-snpleri-belirlemek-icin-bir-arac" title="Mine Yoldaş Orhon, MutEXP: Gen Eksprasyonunu Etkilene SNPleri Belirlemek için Bir Araç " class="translation-link" xml:lang="tr">Türkçe</a></li> </ul> Tue, 05 Sep 2023 15:25:09 +0000 wwwii 1206 at http://ii.metu.edu.tr