Bilkent University Institutional Repository (BUIR)

Bilkent University Institutional Repository (BUIR), a service of Bilkent University Libraries, collects, preserves, and distributes the intellectual output of Bilkent University. Faculty, staff, and students are invited to deposit their research and scholarship. Departments, administrative units, programs, and centers are invited to use the Institutional Repository to distribute their working papers, technical reports, conference proceedings, and other research material. For assistance in depositing documents. For more information, please contact us.


Scholarly Publications


University Library


Bilkent Theses

Recent Submissions

ItemOpen Access
Progressive gameplay: emergent anticapitalism in story-driven video games
(Bilkent University, 2023-11) Doğan, Taylanumut; Özkul McGeoch, Didem
This thesis considers the potential of story-driven video games in promoting liberatory discourses. It examines Cyberpunk 2077, an action/roleplaying game set in a cyberpunk city rife with violence and social inequality; and Night in the Woods, an adventure game about a gothic mystery set in a small Rust Belt town. The thesis adopts a textual and formal approach in conjunction while analyzing the discourses of social liberation and their effectiveness in the interactive yet solitary experience of single-player gameplay through these two video games. In doing so, the thesis adopts a critical theoretical framework that considers these two video games in relation to the Spectacle of capitalist society and the forms of resistance adopted by the Situationist International, as well as the cultural materialist approach of Raymond Williams. Observing these approaches in narrative flow as well as player interactivity, this thesis aims to contribute to the body of thought that evaluates video games as potential sources for discourses of social liberation.
ItemOpen Access
Federated learning and distributed inference over wireless channels
(Bilkent University, 2023-11) Tegin, Büşra; Duman, Tolga Mete
In an era marked by massive connectivity and a growing number of connected devices, we have gained unprecedented access to a wealth of information, enhancing the reliability and precision of intelligent systems and enabling the de-velopment of learning algorithms that are more capable than ever. However, this proliferation of data also introduces new challenges for centralized learning algorithms for the training and inference processes of these intelligent systems due to increased traffic loads and the necessity of substantial computational resources. Consequently, the introduction of federated learning (FL) and distributed inference systems has become essential. Both FL and distributed inference necessitate communication within the network, specifically, the transmission of model updates and intermediate features. This has led to a significant emphasis on their utilization over wireless channels, underscoring the pivotal role of wireless communications in this context. In pursuit of a practical implementation of federated learning over wireless fading channels, we direct our focus towards cost-effective solutions, accounting for hardware-induced distortions. We consider a blind transmitter scenario, wherein distributed workers operate without access to channel state information (CSI). Meanwhile, the parameter server (PS) employs multiple antennas to align received signals. To mitigate the increased power consumption and hardware cost, we leverage complex-valued, low-resolution digital-to-analog converters (DACs) at the transmitter and analog-to-digital converters (ADCs) at the PS. Through a combination of theoretical analysis and numerical demonstrations, we establish that federated learning systems can effectively operate over fading channels, even in the presence of low-resolution ADCs and DACs. As another aspect of practical implementation, we investigate federated learning with over-the-air aggregation over time-varying wireless channels. In this scenario, workers transmit their local gradients over channels that undergo time variations, stemming from factors such as worker or PS mobility and other transmission medium fluctuations. These channel variations introduce inter-carrier interference (ICI), which can notably degrade the system performance, particularly in cases of rapidly varying channels. We examine the effects of the channel time variations on FL with over-the-air aggregation, and show that the resulting undesired interference terms have only limited destructive effects, which do not prevent the convergence of the distributed learning algorithm. Focusing on the distributed inference concept, we also consider a multi-sensor wireless inference system. In this configuration, several sensors with constrained computational capacities observe common phenomena and engage in collaborative inference efforts alongside a central device. Given the inherent limitations on the computational capabilities of the sensors, the features extracted from the front part of the network are transmitted to an edge device, which necessitates sensor fusion for the intermediate features. We propose Lp-norm inspired and LogSumExp approximations for the maximum operation as a sensor fusion method, resulting in the acquisition of transformation-invariant features that also enable bandwidth-efficient feature transmission. As a further enhancement of the proposed method, we introduce a learnable sensor fusion technique inspired by the Lp-norm. This technique incorporates a trainable parameter, providing the flexibility to customize the sensor fusion according to the unique network and sensor distribution characteristics. We show that by encompassing a spectrum of behaviors, this approach enhances the adaptability of the system and contributes to its overall performance improvement.
ItemOpen Access
Transier: space, movement, and transience
(Bilkent University, 2023-10) Beyazıt, Müge İrem; McGeoch, Didem Özkul
Transier is both the title of a growing series of interactive kinetic sculptures made of recycled plastic strips and electro-mechanical components aka. specimens, and their defined framework. These specimens are taking the source of their movement from their viewer regarded as the passer-by. They aim to disrupt their transient state as they pass-by and affect their consciousness with their being. This dynamic as the interaction between the specimen and its viewer is integral to both this artwork and this research while the nature of being is investigated. Taking the first two specimens of Transier consisting my current thesis project as its considered body of work, this thesis aims to provide an in-depth theoretical exploration within the complex web of relations between space, movement, being, and materiality with a particular existential emphasis on transience within the context of my artistic practice. Throughout this theoretical exploration, an inherent interconnectedness between the notions of space, movement, and being is revealed and emphasized; an existential philosophical standpoint where a dynamic, ever-changing nature of being that is related to a constant state of flux is established; and eventually the transience of being is argued.
Zebrafish glioma xenograft models: in vıvo investıgatıon of injection methods and development of a streamlit application
(Bilkent University, 2023-10-20) Tombuloğlu, Rüya; Karakayalı, Özlen Konu
Glioblastoma (GBM) is one of the most aggressive and lethal forms of primary brain cancer, posing significant challenges to effective treatment and patient outcomes. Despite extensive research efforts, our understanding of GBM biology and the development of novel therapeutic strategies remains limited. Zebrafish xenograft models have emerged as a promising tool in cancer research, offering unique advantages in studying GBM. This thesis explores the utility of zebrafish xenograft models in advancing our understanding of GBM. Due to their genetic and physiological similarities to humans, zebrafish provide an excellent platform for studying GBM pathogenesis, tumor progression, and drug screening. Their transparency during early development allows for real-time visualization of tumor growth, invasion, and response to treatments. Moreover, zebrafish models enable rapid and cost-effective high-throughput drug screening, accelerating the identification of potential GBM therapeutics. In this thesis, I focused on creating an application named ZenofishDb Glioma, which is a more evolved and focused version of our previous database called ZenofishDb. ZenofishDb Glioma has been created using Python Streamlit, and comprises only the glioma studies and uses Natural Language Processing (NLP) to classify better, and effectively find and summarize information about zebrafish glioma xenograft models. In addition, after searching ZenofishDb Glioma, I decided to investigate an experimental protocol for injection of glioblastoma cells in the zebrafish model to test effects of injected cell numbers. Using MGG-119-GFP cells and Casper zebrafish, I injected different numbers of cells at different locations and stages, i.e., blastula and 2 days post fertilization, and observed there were significant differences between groups at 5 dpf using multiple quantification strategies. In conclusion, ZenofishDb Glioma can help design effective xenotransplantation strategies and make comparisons to understand how different experimental parameters affect the outcome of zebrafish glioma xenograft models.
Computational analysis of 3D genome organization and its effect on nuclear morphology and mechanics
(Bilkent University, 2023-10) Attar, Ali Göktuğ; Erbaş, Aykut
Several disorders, including progeria, cancer, and Emery-Dreifuss muscular dystrophy, share abnormalities in eukaryotic cells' nuclear structure and mechanics. One of the contributors to nuclear morphology and mechanics is the chromatin filling the 10-micron elastic nucleus. The polymer physics principles behind the relationship between chromatin and nuclear morphology and its mechanics need to be clarified. To elucidate this relationship between chromatin and polymer and nuclear morphology and mechanics, we concentrate on chromatin phase separation utilizing a coarse-grained polymer model encapsulated in an elastic shell. Our approach can capture the conventional and inverted nucleus organization while allowing nuclear deformability. Heterochromatin can be one of the key determinants of the nuclear shape by revealed by examining heterochromatin heterochromatin interactions, as well as the interaction between chromatin and lamina inspecting through the biologically relevant volume fractions. The simulations showed that the heterochromatin-nuclear shell interactions influence the variation in the nuclear shape fluctuations, thus leading to nuclear deformations. The interplay between heterochromatin-heterochromatin interactions and its interaction with the nuclear shell plays a role in phase separation and nuclear shape fluctuations. Higher heterochromatin concentration resulted in abnormal morphology in lower volume fraction, in contrast to some experiments suggesting the opposite trend. The volume fraction exhibits a suppressing effect on the nuclear shape fluctuations in all examinations of heterochromatin interactions. Additionally, the tethering and crosslinking of the heterochromatin provide a chromatin-based stiffness to the nuclear shell revealed by force-strain relationships. Altogether, our results imply that chromatin, mainly heterochromatin, considerably contributes to nuclear morphology and mechanics.