19
Arşivlenen Tez
1
DOI Atanmış
5%
DOI Oranı
Anabilim Dalı
Makine öğrenmesinde gradyan inişi optimizasyon algoritmaları üzerine
Bu çalışmada, makine öğrenmesinde önemli bir yere sahip olan gradyan tabanlı optimizasyon algoritmalarının yapısı, çeşitleri, avantaj ve dezavantajlarına yer verilmiştir. Bu amaçla 1. dereceden optimizasyon algoritmalarından literatürde en çok kullanılan; Stokastik Gradyan İniş, Momentum, Nesterov Momentum, AdaGrad, Adadelta, RMSProp, Adam ve Nadam algoritmaları ile 2. Dereceden optimizasyon algoritmalarından Newton, BFGS ve L-BFGS algoritmaları ele alınmıştır. Algoritmaların matematiksel yapıları incelenmiş ve karşılaştırmaları için üç farklı gerçek hayat problemi ele alınmıştır. Bu problemlerin yapay zeka modelleri ile çözümlerinde ResNet50, VGG19 ve lojistik regresyon modelleri kullanılmıştır. Elde edilen sonuçlar çizelge ve şekiller üzerinden değerlendirilmiştir. Algoritmaların performansları metriklerle ölçülerek, algoritmaların hem birbirlerine karşı performansları hem de modellerde ki başarımları tespit edilmiş ve sonuçları yorumlanmıştır.
indexliAçık ErişimTRAğ trafiği kullanarak saldırı tespit edebilen derin öğrenme modellerinin performans değerlendirmesi
Dijital altyapıların hızla büyümesi ve siber saldırıların giderek karmaşıklaşması, ağ tabanlı saldırı tespit sistemlerini modern siber güvenliğin temel bileşenlerinden biri hâline getirmiştir. İmza tabanlı yöntemler yeni tehditlerin tespitinde yetersiz kalabildiğinden, makine öğrenmesi ve derin öğrenme yöntemleri önemli alternatifler sunmaktadır. Bu çalışmada, tek boyutlu evrişimli sinir ağı (CNN) ile hibrit CNN-BiLSTM mimarisinin ikili ve çok sınıflı ağ saldırısı sınıflandırma performansları incelenmiştir. Modeller, CIC-IDS2017 ve UNSW-NB15 veri kümeleri üzerinde tabakalı beş katlı çapraz doğrulama, odak kaybı, QuantileTransformer ile ölçeklendirme ve SMOTE ile aşırı örnekleme içeren ortak bir deneysel düzen kullanılarak değerlendirilmiştir. Ayrıca rastgele orman ve XGBoost, UNSW-NB15 üzerinde klasik karşılaştırma modelleri olarak test edilmiştir. Bulgular, CNN modelinin CIC-IDS2017 ikili sınıflandırma görevinde %99,75 doğruluk ve 0,9999 ROC-AUC değerine ulaştığını göstermiştir. CNN-BiLSTM, UNSW-NB15 ikili sınıflandırma görevinde yanlış negatifleri yaklaşık %36 azaltmış; ancak çıkarım gecikmesi CNN’e göre 4-6 kat artmıştır. Çok sınıflı sınıflandırmada CNN, CIC-IDS2017 üzerinde 0,701 makro F1 skoruyla CNN-BiLSTM’den (0,671) daha yüksek performans göstermiştir. XGBoost ise derin öğrenme modelleriyle benzer F1 skorlarına daha kısa sürede ulaşmıştır. Azınlık saldırı sınıflarının tespiti tüm yapılandırmalarda temel bir güçlük olmayı sürdürmektedir. Sonuçlar, gerçek uygulama kısıtları altında derin öğrenme ve klasik makine öğrenmesi modelleri arasında seçim yapacak araştırmacı ve uygulayıcılar için karşılaştırmalı çıkarımlar sunmaktadır.
Verimli ve etkili öğrenme için rezervuar hesaplama
Resource-efficient modeling of nonlinear dynamical systems is vital for embedded sensing, edge analytics, and neuromorphic hardware. Reservoir Computing (RC) reduces training overhead in recurrent neural networks and transformers by fixing internal weights and training only a linear read-out. However, classical Echo State Networks (sRC) scale capacity by increasing reservoir size, which inflates memory and energy costs. This thesis addresses that bottleneck by improving the quality, rather than the quantity, of reservoir units. We propose the Higher-Order Augmented Reservoir Computer (haRC), which augments reservoir with a subset of multiplicative monomials of co-temporal activations. This polynomial feature space enhances nonlinear expressivity while keeping the reservoir compact. On chaotic benchmarks such as Lorenz and Mackey–Glass, haRC achieves the same or better accuracy with fewer parameters than baseline sRC models.We propose two improved models: Selective haRC, which ranks reservoir units by variance and covariance to discard redundant interactions, and Sparse haRC, which integrates controlled sparsity masks to reduce memory and compute load. Comprehensive experiments evaluate memorization and forecasting tasks under equal total and tunable parameter settings across different sparsity levels. haRC's consistently outperform sRC across all sparsity regimes, with the advantage growing in high sparsity. Noise robustness tests inject Gaussian noise into the reservoir. Selective haRC maintains or exceeds the performance of sRC in terms of root-mean-squared error and variance with 12% of the reservoir units sRC has. Demonstrating up to 120 times memory efficiency, higher-order augmentation emerges as a promising approach for ultra-lightweight sequence learners suited for microcontrollers, FPGA overlays, and in-memory compute systems.
SQTL ve HI-C'nin ortak analizi, birden fazla dokuda Sqtller ile hedef genler arasında mekansal yakınlık ortaya çıkarıyor
Gene expression and regulation with or without alternative splicing are crucial for tissues and cells to properly function. They have been studied from three almost independent perspectives at the genome level: 1- Recognition of splicing quantitative trait loci~(sQTLs), 2- Expression quantitative trait loci~(eQTLs) recognition, and 3- Recognition of longer-range physical chromatin interactions between genome segments which model $3$D dynamics of cells and tissues. Even though the associations between eQTLs and longer range chromatin interactions have been previously studied, similar relationship between sQTLs and chromatin interactions has not been previously analyzed. In this case, it is crucial to analyze whether sQTLs control the alternative splicing of their target genes mRNA via physically-interacting genome segments. Even though chromatin interactions are part of the principal processes governing eQTLs functioning, similar analysis is missing from sQTLs perspective. We have jointly analyzed high-throughput chromatin conformation capture~(Hi-C) and sQTL datasets over 8 different human cancer tissues. We have discovered the existence of positive association between the number of genes having sQTLs and chromatin interaction frequency. Such positive association still exists when we also control for eQTLs. Additionally, sQTLs and their target genes generally exist inside identical topologically associating domain~(TAD). Those findings are observed over the whole set of analyzed cancer types and over different functional subsets of sQTL dataset such as survival-related sQTLs. Furthermore, tissue-specific sQTLs are statistically enriched in tissue-specific frequently interacting regions~(FIREs) in 6 out of 8 human cancer tissues~(Chronic Myeloid Leukemia, Colon Adenocarcinoma, Acute Myeloid Leukemia, Lung Adenocarcinoma, Prostate Cancer, Sarcoma). Our sQTL and Hi-C datasets have shown the existence of closer spatial distance between sQTLs and their target genes with possible alternative splicing across a number of different cancer types in human. Such closer spatial distance also exists independent of whether we integrate eQTLs into the analysis. We found that sQTLs regulate the alternative splicing through chromatin interactions.
Kombinatoryal optimizasyon için grafik sinir ağları tabanlı birincil sezgisel yöntem
By examining the patterns of solutions obtained for varying instances, one can gain insights into the structure and behavior of combinatorial optimization (CO) problems and develop efficient algorithms for solving them. Machine learning techniques, especially Graph Neural Networks (GNNs), have shown promise in parametrizing and automating this laborious design process. The inductive bias of GNNs allows for learning solutions to mixed-integer programming (MIP) formulations of constrained CO problems with a relational representation of decision variables and constraints. The trained GNNs can be leveraged with primal heuristics to construct high-quality feasible solutions to CO problems quickly. However, current GNN-based end-to-end learning approaches have limitations for scalable training and generalization on larger-scale instances; therefore, they have been mostly evaluated over small-scale instances. Addressing this issue, our study builds on end-to-end learning of optimal solutions to the downscaled instances of given large-scale CO problems. We introduce several improvements on a recent GNN model for CO to generalize on instances of a larger scale than those used in the training. We also propose a two-stage primal heuristic strategy based on uncertainty-quantification to automatically configure how solution search relies on the predicted decision values. Our models can generalize on 16x upscaled instances of commonly benchmarked five CO problems. Unlike the regressive performance of existing GNN-based CO approaches as the scale of problems increases, the CO pipelines using our models offer an incremental performance improvement relative to a state-of-the-art MIP solver CPLEX. The proposed uncertainty-based primal heuristics provide 6-75% better optimality gap values and 45-99% better primal gap values for the 16x upscaled instances and brings immense speedup to obtain high-quality solutions. All these gains are achieved in a computationally efficient modeling approach without sacrificing solution quality.
A comprehensive human-agent negotiation framework: Preferences, emotion & interaction
In today's increasingly interconnected world, human-agent negotiation plays a pivotal role in reaching socially beneficial agreements when stakeholders need to make joint decisions. Developing intelligent agents capable of understanding not only human negotiators' preferences but also attitudes is a significant prerequisite for effective human-agent interactions. Awareness of a human's emotional state and ability to express an agent's mood to influence the human negotiator might significantly affect the negotiation outcome. This thesis presents a comprehensive framework that revolutionizes the field of human-agent negotiation, integrating two critical elements: Emotionally aware negotiation strategy and Conflict-Based Opponent Modeling (CBOM). By combining these novel approaches, the framework enhances negotiation outcomes and fosters cooperation between agents and human negotiators, ultimately leading to mutually advantageous agreements. The thesis establishes the research context and motivation, underscoring the escalating importance of human-agent negotiation in a world where collaborative decision-making is essential for addressing complex challenges. It highlights the need for advanced agents to accurately interpret human preferences and behaviors, enabling admissible settlements that serve joint interests. Shedding light on the limitations of conventional approaches that heavily rely on opponent offers and remaining time. Additionally, it explores the critical role of emotional awareness and opponent modeling strategies in human-agent negotiation. The synthesis of existing research lays the groundwork for developing the proposed comprehensive framework. Emotional awareness takes center stage in the proposed negotiation strategy. Solver Agent: Emotional Extension of the Hybrid Agent bidding strategy is introduced. The Solver Agent considers the opponent's emotional state during negotiation, leading to higher social welfare scores and faster agreement times. The experimental study emphasized the profound impact of emotional awareness on negotiation outcomes, particularly in human-agent settings. CBOM efficiently extracts maximum information from limited interaction rounds in human-agent negotiation settings, surpassing traditional approaches in prediction performance. Experimental analyses confirmed the superiority of CBOM in human-agent and automated negotiation scenarios, even when the exploration of the outcome space is limited. The experimental findings establish CBOM as a powerful tool for modeling human behavior and preferences in negotiation. In conclusion, the comprehensive human-agent negotiation framework presented in this thesis represents a significant advancement in the field. By seamlessly combining Conflict-Based Opponent Modeling and Emotional Awareness, the framework empowers intelligent agents to discern human preferences and behaviors more accurately, facilitating cooperative interactions and achieving mutually beneficial agreements. The framework's effectiveness in human-agent and automated negotiation settings highlights its potential for designing negotiation agents that interact adeptly with human negotiators, fostering understanding and optimizing negotiation outcomes. The future of human-agent negotiation lies in forging a new era of cooperation, where intelligent agents serve as capable partners, promoting social welfare and driving positive change through admissible settlements that incorporate joint interests. This thesis contributes valuable insights towards realizing this vision, marking a significant step forward in the field of human-agent interaction.
Illumination-guided inverse rendering benchmark: Learning real objects with few cameras
The field of 3D computer vision and graphics has seen a rapid expansion of late, making the development of realistic virtual environments and digital representations of real-world objects possible. Fundamental to this development are 3D reconstruction techniques that enable the transition of physical objects' form, color, and surface particulars to the virtual domain. Current approaches mainly rely on neural scene representations, which, despite being effective, face challenges such as the need for a large quantity of captured images and the complexity associated with converting these representations into explicit geometric forms. An alternative strategy that has gained traction is the deployment of methods such as physically-based differentiable rendering (PBDR) and inverse rendering. These approaches require fewer viewpoints, yield explicit format results, and ensure a smoother transition to other representation methods. However, in order to effectively assess the performance of the available methods in 3D reconstruction, it is imperative to utilize standard benchmark scenes for comparison. Although there are standard objects and scenes available in existing research, there is a noticeable deficiency of real-world benchmark data that simultaneously captures camera, lighting, and scene parameters — all of which are essential for high-quality 3D reconstructions using methods based on PBDR and inverse rendering. In this study, we present a method for capturing real-world scenes as virtual environments, incorporating lighting parameters along with camera and scene parameters to enhance the veracity of virtual representations. In addition, we provide a set of ten real-world scenes, each with its corresponding virtual counterparts, purposely designed as benchmarks. These benchmarks cover a basic assortment of geometric structures, such as convex, concave, flat, and composite surfaces. Furthermore, we showcase the 3D reconstruction results of cutting-edge 3D reconstruction techniques using PBDR in real-world scenes, using both established methodologies and our proposed one.
Derı̇n transformatör tabanlı varlık fı̇yatı ve yön tahmı̇nı̇
The field of algorithmic trading, driven by deep learning methodologies, has garnered substantial attention in recent times. Within this domain, transformers, convolutional neural networks, and patch embedding-based techniques have emerged as popular choices within the computer vision community. Here, inspired by the latest cutting-edge computer vision methodologies and the existing work showing the capability of image-like conversion for time-series datasets, we apply more advanced transformer-based and patch-based approaches for predicting asset prices and directional price movements. The employed transformer models include Vision Transformer~(ViT), Data Efficient Image Transformers~(DeiT), and Swin. We use ConvMixer for a patch embedding-based convolutional neural network architecture without a transformer. Our tested transformer-based and patch-based methodologies aim to predict asset prices and directional movements using historical price data by leveraging the inherent image-like properties within the historical time-series dataset. Before the implementation of attention-based architectures, the historical time series price dataset is transformed into two-dimensional images. This transformation is facilitated through the incorporation of various common technical financial indicators, each contributing to the data for a fixed number of consecutive days. Consequently, a diverse set of two-dimensional images is constructed, reflecting various dimensions of the dataset. Subsequently, the original images depicting market valleys and peaks are annotated with labels such as Hold, Buy, or Sell. According to the experiments, trained attention-based models consistently outperform the baseline convolutional architectures, particularly when applied to a subset of frequently traded Exchange-Traded Funds~(ETFs). This better performance of attention-based architectures, especially ViT, is evident in terms of both accuracy and other financial evaluation metrics, particularly during extended testing and holding periods. These findings underscore the potential of transformer-based approaches to enhance predictive capabilities in asset price and directional forecasting. Our code and processed datasets are available at~https://github.com/seferlab/price_transformer
İkili müzakerelerde zaman serisi tahmin modelleri
This thesis explores the dynamics of agent-based negotiations, with a focus on understanding opponent's behavior and predicting their offering patterns to make strategic decisions. Guessing the utility of the opponent's upcoming offers valuable insights for the agent's subsequent moves. The research aims to predict the opponent's future offers by employing diverse learning algorithms in various situations to measure their effectiveness in comprehending negotiation behavior. The prediction study comprises two parts; one investigating the impact of these models in one-to-one negotiations, specifically tailored for the agent's own experiences, while the other examines the performance of predictive models in a tournament setting for all agents. A learning process with three distinct targets have been established to assess the prediction models: (i) estimating the agent's utility of the opponent's next offer by considering only its offer history, (ii) estimating the agent's utility considering opponent-related variables, and (iii) estimating the opponent's utility using opponent-related variables. According to the experimented results, the best learning approach is incorporated into an agent design to observe the impacts of having predictions of future utility values on the agent's negotiation success. The thesis evaluates these models in diverse negotiation scenarios and highlights promising outcomes for the proposed methods. It also introduces a novel negotiation strategy called `Negoformer', which incorporates predictions into the offering strategy and investigates their impact on the outcome of negotiations. The experiments showcased the success of Negoformer compared to other agents in various negotiation success metrics, such as individual utility value and social welfare score.
Gösterimden öğrenme için etkili sinir ağları
Learning from demonstration (LfD) is a powerful technique for teaching robots new behaviors by mimicking human demonstrations. One widely used approach in LfD is Behavior Cloning (BC) with human-in-the-loop control. In this method, data collected from humans demonstrations is utilized to create a non-linear controller by learning a mapping from states to desired actions. In this study, we propose a novel BC system that its structure mirrors the structure of an error-based feedback controller. This design choice is based on the premise that embedding such a structure within the learning model can endow our system with a prior bias, resulting in an inherent advantage over controller-agnostic BC systems. This thesis details the components of this novel BC model and demonstrates its application on a two degrees-of-freedom robotic system across various tasks as a proof of concept. To assess the system's effectiveness, we conducted systematic experiments comparing it to a controller-agnostic BC system. The results reveal that our proposed model significantly outperforms the baseline, suggesting it is a promising candidate for LfD tasks where the demonstration data can be assumed to be generated by a feedback controller, especially in resource-scarce conditions.
Akademik kalite ölçümü için yapay zeka tabanli bir yaklaşim
In the context of academic recruitment at universities and research institutions, establish- ing consistent and effective evaluation criteria remains a complex challenge. Identifying robust metrics that align with globally recognized standards of academic quality is essen- tial to ensure merit-based evaluation of researchers and maintain institutional credibility. In this thesis, we address this challenge through an AI-based solution aimed at developing a data-driven approach to quantify academic quality. As a benchmark of academic excellence, we use Nobel laureates' profiles in Physics, Chemistry, Physiology or Medicine, and Economics as a reference cohort. Comparison group includes researchers from the same fields affiliated with universities with an average ranking based on the Times Higher Education World University Rankings. By defining bibliometric features of the academic profiles of both Nobel laureates and the comparison group, we aim to develop machine learning models to quantify academic quality and identify the key features that define academic excellence. The ultimate goal is to support the decision-making process of universities and research institutions in academic recruitment, creating a fair and objective evaluation criteria.
Beyin BT görüntülerinden inme tespiti ve sınıflandırılması için gürültü azaltma yöntemleri
Halk arasında inme ya da felç olarak bilinen hemipleji, beynin çeşitli sebeplerden dolayı herhangi bir bölümünün beslenememesi sonucu bölgedeki hücrelerin ölmesi olayına denir. Dünyada ölüm nedenleri arasında ikinci sırada yer almaktadır. Bununla birlikte DSÖ'ye göre her yıl 15 milyondan fazla insana inme tanısı konmakta, 5 milyon insan inme sebebiyle hayatını kaybetmektedir. İnmede zaman en önemli faktördür. İnme etkisinin zamana bağlı olarak artması hem hastanın sağlığı hem de uygulanabilecek tedavi çeşitliliği açısından kritik bir değer taşımaktadır. Bu sebeple inmenin en hızlı şekilde tespiti hastalığın tedavisi açısından önem arz etmektedir. Tıbbi görüntüleme yöntemleri; hastalıkların teşhisinde, tedavinin seyrinde çok önemli bir yer tutmaktadır. Hekim, radyolog veya karar verici, radyolojik görüntüler üzerinden hastalığı tespit edip, hastalığın seyri, olası tedaviler ve yaklaşımlar hakkında değerlendirmelerde bulunmaktadır. Bilgisayarlı Tomografi (BT), inme hastalığının tespitinde kullanılan tıbbi görüntüleme yöntemlerinden biridir. Tüm tıbbı görüntüleme yöntemlerinde olduğu gibi BT görüntülerinin çekimlerinde çekim kalitesini bozan kirlilik/gürültüler ortaya çıkmaktadır. BT görüntü üzerindeki bu gürültü kimi zaman radyolog ya da karar vericiyi hem hastalığın tanısında hem de ileri tetkiklerinde yanıltabilmektedir. Bu sebeple BT görüntülerinden bu gürültünün temizlenmesi, bu projenin motivasyon kaynağı olmuştur. Genel olarak inme, tıkayıcı (iskemik) ve kanamalı (hemorajik) olmak üzere ikiye ayrılır. Bu çalışmada kanamalı inme özelinde bilgisayarlı tomografi görüntülerinde gürültüyü azaltarak söz konusu medikal görüntüleri daha iyi hale getirip, görüntülerin inme var/yok sınıflandırılmasında yapılan gürültü temizleme işleminin etkisi araştırılmıştır. Çalışmada incelenen gürültü temizleme metotları; uzamsal alan filtrelerinden Medyan ve Gauss filtresi, dönüşüm alan filtrelerinden Wiener filtresi ve Dalgacık Dönüşümüdür. Çalışmada Kuzey Amerika Radyoloji Derneği (KARD) tarafından düzenlenen yarışmada paylaşılan veri setinden faydalanılmıştır. Elde edilen veri seti üzerinde Gauss gürültüsü 1, 5, 10, 25, 50, 75 ve 100 standart sapma seviyelerinde eklenerek adı geçen gürültü temizleme metotları ile temizlenmiştir. Gürültü temizleme performansını ölçmek için SSIM, PSNR ve MSE metrikleri kullanılmıştır. Çalışmada inmenin sınıflandırması için derin sinir ağı modellerinden ResNet50, DenseNet121, InceptionV3 ve AlexNet kullanılmıştır. Modellerin eğitilecek ağırlıkları, ImageNet ağırlıkları ve rassal olmak üzere iki farklı şekilde seçilmiştir. Sınıflandırma modellerinin başarımını ölçmek için Doğruluk, Kesinlik, Duyarlılık, AUC ve F1 Skoru kullanılmıştır. Gürültü temizleme ve model sınıflandırma performansları grafikler üzerinden değerlendirilmiş, gürültü temizleme ve model başarımları arasındaki ilişki ortaya konmuştur. Çalışma sonucunda; Gürültünün ve gürültü temizlemenin BT görüntülerini sınıflandırmada kullanılan derin sinir ağ modellerinin başarımı üzerindeki etkisi ortaya konmuştur. Gürültü, BT görüntüsünü bozmakta ve derin sinir ağlarının bilgi çıkarımını zorlaştırmaktadır. Derin sinir ağlarının başarımı gürültüyle olumsuz etkilenmiş ve düşmüştür. Gürültü temizleme metotları ile gürültünün etkisi azaltılmış ve gürültü temizleme süreci sınıflandırıcı modellerin başarımını gürültülü yapıya göre yükseltmiştir. Bu çalışma ile araştırmacılara, çalışmalarına gürültüyü ve gürültü temizlemeyi bir aşama olarak eklemeleri önerilmektedir. Model başarımlarını en iyi hale getirmek için gürültü temizleme aşamasında farklı gürültü temizleme yaklaşımlarının kullanılması, kullanılan metotların parametre optimizasyonunun yapılması, çalışmalarda farklı yaklaşımlar içeren derin sinir ağ modellerinin kullanılması ve gürültü temizlemenin bu derin sinir ağ modelleri ile test edilmesi önerilen diğer başlıklardır.
Enhanced e-commerce decision-making using iot and machine learning
This master's thesis explores the integration of Machine Learning (ML) and Internet of Things (IoT) in e-commerce decision-making, emphasizing their potential to improve customer satisfaction and revenue. It categorizes prior research in the field, addresses challenges in data quality, privacy, and algorithmic bias, and conducts empirical experiments with various ML and Deep Learning models. This work not only contributes to the existing literature by categorizing prior research but also showcases the practical applicability of ML and IoT in e-commerce decision-making, paving the way for future advancements in this domain.
Recommending ancillary products in aviation industry: A comparative study on recommender systems using online customer reviews
Increasing competition in the aviation industry forces airline companies to find new ways to increase their profitability. Airline companies try to do that by offering ancillary products beside their main service. However, there is the problem of offering a suitable product to the customer who really needs it. Recent improvements in the recommender systems area, especially in the e-commerce industry, raise the question of whether those systems are efficient for recommending ancillary products in aviation industry. In this study we aim to build a recommender system for ancillary products for the airline industry using online customer reviews. Customer reviews from various web sites were separated by their topics using BERTOPIC topic modelling algorithm. Expert labeled customer reviews were fed into the algorithms to build recommender system. The aim of the study is to compare recommender systems using different machine learning algorithms. As the result of the study Neural networks gave the highest accuracy results of 0.85.
The interplay of multiple kernel learning in gans via robust optimization
This study proposes a novel model using multiple kernel learning (MKL) in generated adversarial networks using robust optimization. Furthermore, we integrated stochastic functional gradient (SFG) and (MKL). We rigorously compared this model with empirical risk minimization (ERM), which is also known as sample average approximation and the SFG RKHS. The latter combines reproducing kernel Hilbert spaces (RKHS) which is the key component of support vector machines (SVM) with (SFG). The comparison was performed on challenging datasets, including CIFAR 10, CIFAR 100, MNIST, Fashion MNIST, EMNIST, and SVHN. The SFG MKL model consistently has lower error rates and better performance in adversarial attack scenarios. This shows that it can handle changes from adversaries, which is an important trait. This resilience is a testament to the effectiveness of the functional gradient approach when combined with MKL. In comparison, the ERM model is highly susceptible to perturbation. In contrast, the SFG MKL model shows competitive efficacy with the SFG RKHS in more standard scenarios. This suggests that integrating MKL into the functional gradient framework is a way to enhance model resilience. Our results confirm that the SFG MKL model is a contender in machine learning applications requiring accuracy and resilience against adversarial perturbations. The possibility of exploring kernels within the MKL framework opens opportunities for future progress, especially in critical areas like biomedical imaging and autonomous systems. Combining gradients with MKL holds v great potential to advance the development of reliable machine-learning algorithms, establishing a new standard in the field
Derinlik ve segmentasyon kullanarak tek görüntülerden bilinen nesnelerin hacim tahmini
This thesis delves into the challenging task of estimating the volume of known objects from a single-view image perspective. In this thesis, we proposed a new model for estimating the volume of objects from just one image. Leveraging state-of-the-art deep learning methodologies, the image undergoes a comprehensive analysis involving depth and segmentation networks. The depth network is responsible for estimating the object's depth map, while the segmentation network determines the pose of the objects within the image. Combining these outputs with the intrinsic data of the camera results in the creation of a detailed point cloud. This point cloud serves as the foundational data source for precise volume estimation, contributing to advancements in the field of computer vision. The experiments on publicly available datasets show that our method outperforms the other methods and achieves the state-of-the-art performance.
Noncontractual churn analysis for a private bank customers
The study includes data for a bank in the interest-free finance sector for 2020 and 2021. 6,844 corporate segment customers of a private participation bank in Turkey who exchanged foreign currency and precious metals via a mobile banking application were investigated. For customer churn analysis, first the RFM model was developed, then the CLV value was calculated. A threshold was determined with the statistical method in the transactional data, which was unlabeled, and the customers below the threshold were labeled as "lost customers." After all, customers were labeled according to this rule, and supervised methods were used to predict the most probable churners.
Benchmarking study for customer churn prediction: A case study in the e-commerce industry
The e-commerce sector has grown and continues to grow with an acceleration in the last decade, and the data produced and used in this sector has also prepared the environment for important work areas for the big data world. The data produced and collected in the e-commerce sector is used in many decision-making mechanisms. One of them is customer churn prediction. This study aimed to prepare a benchmark for estimating customer churn using traditional and current machine learning approaches, using data obtained from an e-commerce company operating in Turkey. It also aimed to enrich the benchmarking study by offering alternative solutions to the target class imbalance problem, which is frequently encountered in data sets used in customer churn prediction. A total of 24 different models were created, including 6 forecasting models and 4 sampling strategies. In this study, the best performing model pair was the XGBoost Algorithm and the No Sampling. This model duo produced values of 91%, 79% on the basis of accuracy, F2-Score respectively.
Deep learning approaches for autism spectrum dis-order diagnosis: Ensemble archtectures and multi-modal analysis
This thesis presents novel multi-modal deep learning approaches for enhancing the precision and efficiency of Autism Spectrum Disorder (ASD) diagnosis in children. The first approach, ASD-CVH, combines a hybrid vision transformer and convolu-tional neural network (CNN) architecture to extract high-level features and attention maps from audio samples, achieving excellent accuracy in differentiating between ASD and typically developing (TD) children. The second approach, ASD-EVNet, utilizes ensemble learning with state-of-the-art Vision Transformer (ViT) models fine-tuned on a face-based ASD dataset for children (FADC), achieving state-of-the-art results in ASD diagnosis based on facial expressions. By integrating audio-based and visual-based deep learning models, this research establishes a comprehensive framework for ASD diagnosis, providing a multi-modal perspective that enhances accuracy and relia-bility. The findings contribute to the development of objective and efficient diagnostic tools for ASD, supporting early intervention and improved care for children with ASD.