Publications without Fulltext
Permanent URI for this collectionhttps://hdl.handle.net/20.500.14288/3
Browse
76 results
Search Results
Publication Metadata only Learning Markov Chain Models from sequential data under local differential privacy(Springer Science and Business Media Deutschland Gmbh, 2024) Department of Computer Engineering; Güner, Efehan; Gürsoy, Mehmet Emre; Department of Computer Engineering; Graduate School of Sciences and Engineering; College of EngineeringMarkov chain models are frequently used in the analysis and modeling of sequential data such as location traces, time series, natural language, and speech. However, considering that many data sources are privacy-sensitive, it is imperative to design privacy-preserving methods for learning Markov models. In this paper, we propose Prima for learning discrete-time Markov chain models under local differential privacy (LDP), a state-of-the-art privacy standard. In Prima, each user locally encodes and perturbs their sequential record on their own device using LDP protocols. For this purpose, we adapt two bitvector-based LDP protocols (RAPPOR and OUE); and furthermore, we develop a novel extension of the GRR protocol called AdaGRR. We also propose to utilize custom privacy budget allocation strategies for perturbation, which enable uneven splitting of the privacy budget to better preserve utility in cases with uneven sequence lengths. On the server-side, Prima uses a novel algorithm for estimating Markov probabilities from perturbed data. We experimentally evaluate Prima using three real-world datasets, four utility metrics, and under various combinations of privacy budget and budget allocation strategies. Results show that Prima enables learning Markov chains under LDP with high utility and low error compared to Markov chains learned without privacy constraints.Publication Metadata only Building quadtrees for spatial data under local differential privacy(Springer Science and Business Media Deutschland Gmbh, 2023) Department of Computer Engineering; Alptekin, Ece; Gürsoy, Mehmet Emre; Department of Computer Engineering; Graduate School of Sciences and Engineering; College of EngineeringSpatial decompositions are commonly used in the privacy literature for various purposes such as range query answering, spatial indexing, count-of-counts histograms, data summarization, and visualization. Among spatial decomposition techniques, quadtrees are a popular and well-known method. In this paper, we study the problem of building quadtrees for spatial data under the emerging notion of Local Differential Privacy (LDP). We first propose a baseline solution inspired from a state-of-the-art method from the centralized DP literature and adapt it to LDP. Motivated by the observation that the baseline solution causes large noise accumulation due to its iterative strategy, we then propose a novel solution which utilizes a single data collection step from users, propagates density estimates to all nodes, and finally performs structural corrections to the quadtree. We experimentally evaluate the baseline solution and the proposed solution using four real-world location datasets and three utility metrics. Results show that our proposed solution consistently outperforms the baseline solution, and furthermore, the resulting quadtrees provide high accuracy in practical tasks such as spatial query answering under conventional privacy levels.Publication Metadata only Implications of node selection in decentralized federated learning(IEEE, 2023) Department of Computer Engineering; Lodhi, Ahnaf Hannan; Akgün, Barış; Özkasap, Öznur; Department of Computer Engineering; Graduate School of Sciences and Engineering; College of EngineeringDecentralized Federated Learning (DFL) offers a fully distributed alternative to Federated Learning (FL). However, the lack of global information in a highly heterogeneous environment negatively impacts its performance. Node selection in FL has been suggested to improve both communication efficiency and convergence rate. In order to assess its impact on DFL performance, this work evaluates node selection using performance metrics. It also proposes and evaluates a time-varying parameterized node selection method for DFL employing validation accuracy and its per-round change. The mentioned criteria are evaluated using both hard and stochastic/soft selection on sparse networks. The results indicate that the bias associated with node selection adversely impacts performance as training progresses. Furthermore, under extreme conditions, soft selection is observed to result in higher diversity for better generalization, while a completely random selection is shown to be preferable with very limited participation.Publication Metadata only Histopathological classification of colon tissue images with self-supervised models(IEEE, 2023) Department of Computer Engineering; Erden, Mehmet Bahadır; Cansız, Selahattin; Demir, Çiğdem Gündüz; Department of Computer Engineering; Graduate School of Sciences and Engineering; College of EngineeringDeep learning techniques have demonstrated their ability to facilitate medical image diagnostics by offering more precise and accurate predictions. Convolutional neural network (CNN) architectures have been employed for a decade as the primary approach to enable automated diagnosis. On the other hand, recently proposed vision transformers (ViTs) based architectures have shown success in various computer vision tasks. However, their efficacy in medical image classification tasks remains largely unexplored due to their requirement for large datasets. Nevertheless, significant performance gains can be achieved by leveraging self-supervised learning techniques through pretraining. This paper analyzes performance of self-supervised pretrained networks in medical image classification tasks. Results on colon histopathology images revealed that CNN based architectures are more effective when trained from scratch, while pretrained models could achieve similar levels of performance with limited data.Publication Metadata only Role allocation through haptics in physical human-robot interaction(Institute of Electrical and Electronics Engineers (IEEE), 2013) N/A; N/A; Department of Computer Engineering; Department of Mechanical Engineering; Küçükyılmaz, Ayşe; Sezgin, Tevfik Metin; Başdoğan, Çağatay; PhD Student; Faculty Member; Faculty Member; Department of Computer Engineering; Department of Mechanical Engineering; Graduate School of Sciences and Engineering; College of Engineering; College of Engineering; N/A; 18632; 125489This paper presents a summary of our efforts to enable dynamic role allocation between humans and robots in physical collaboration tasks. A major goal in physical human-robot interaction research is to develop tacit and natural communication between partners. In previous work, we suggested that the communication between a human and a robot would benefit from a decision making process in which the robot can dynamically adjust its control level during the task based on the intentions of the human. In order to do this, we define leader and follower roles for the partners, and using a role exchange mechanism, we enable the partners to negotiate solely through force information to exchange roles. We show that when compared to an “equal control” condition, the role exchange mechanism improves task performance and the joint efficiency of the partners.Publication Metadata only Distributed deep reinforcement learning with wideband sensing for dynamic spectrum access(Ieee, 2020) Ucar, Seyhan; N/A; Department of Computer Engineering; Department of Electrical and Electronics Engineering; Kaytaz, Umuralp; Akgün, Barış; Ergen, Sinem Çöleri; PhD Student; Faculty Member; Faculty Member; Department of Computer Engineering; Department of Electrical and Electronics Engineering; Graduate School of Sciences and Engineering; College of Engineering; College of Engineering; N/A; 258784; 7211Wideband spectrum sensing (WBS) has been a critical issue for communication system designers and specialists to monitor and regulate the wireless spectrum. Detecting and identifying the existing signals in a continuous manner enable orchestrating signals through all controllable dimensions and enhancing resource usage efficiency. This paper presents an investigation on the application of deep learning (DL)-based algorithms within the WBS problem while also providing comparisons to the conventional recursive thresholding-based solution. For this purpose, two prominent object detectors, You Only Learn One Representation (YOLOR) and Detectron2, are implemented and fine-tuned to complete these tasks for WBS. The power spectral densities (PSDs) belonging to over-the-air (OTA) collected signals within the wide frequency range are recorded as images that constitute the signal signatures (i.e., the objects of interest) and are fed through the input of the above-mentioned learning and evaluation processes. The main signal types of interest are determined as the cellular and broadcast types (i.e., GSM, UMTS, LTE and Analogue TV) and the single-tone. With a limited amount of captured OTA data, the DL-based approaches YOLOR and Detectron2 are seen to achieve a classification rate of 100% and detection rates of 85% and 69%, respectively, for a nonzero intersection over union threshold. The preliminary results of our investigation clearly show that both object detectors are promising to take on the task of wideband signal detection and identification, especially after an extended data collection campaign.Publication Metadata only Seed-based distributed group key selection algorithm for ad hoe networks(IEEE, 2007) N/A; Department of Computer Engineering; Atsan, Emre; Özkasap, Öznur; Master Student; Faculty Member; Department of Computer Engineering; Graduate School of Sciences and Engineering; College of Engineering; N/A; 113507Key establishment has a significant role in providing secure infrastructure for ad hoc networks. For this purpose, several key pre-distribution schemes have been proposed, but majority of the existing schemes rely on a trusted third party which causes a constraint in ad hoc platforms. We propose a seed-based distributed key selection algorithm, namely SeeDKS, for groups of nodes in ad hoc networks. Our approach is inspired by the earlier work on distributed key selection (DKS) and is based on the idea of common group key pool generated with group seed value for each different group. Simulation results show that using very small key ring sizes compared to DKS, we can achieve satisfactory results which DKS cannot accomplish in means of finding at least one common key among group members.Publication Metadata only Equilibrium analysis for linear and nonlinear aggregation in network models: applied to mental model aggregation in multilevel organisational learning(Taylor & Francis Ltd, 2022) Treur, Jan; Department of Computer Engineering; Canbaloğlu, Gülay; Undergraduate Student; Department of Computer Engineering; College of Engineering; N/AIn this paper, equilibrium analysis for network models is addressed and applied in particular to a network model of multilevel organisational learning. The equilibrium analysis addresses properties of aggregation characteristics and connectivity characteristics of a network. For aggregation characteristics, it is shown how certain classes of nonlinear functions enable equilibrium analysis of the emerging dynamics within the network like linear functions do. For connectivity characteristics, by using a form of stratification for the network's strongly connected components, it is shown how equilibrium analysis results can be obtained relating equilibrium values in any component to equilibrium values in (independent) components without incoming connections. In addition, concerning aggregation characteristics, two specific types of nonlinear functions for aggregation in networks (weighted euclidean functions and weighted geometric functions) are analysed. It is illustrated in detail how by using certain function transformations also methods for equilibrium analysis based on a symbolic linear equation solver, can be applied to make predictions about equilibrium values for them. All these results are applied to a network model for organisational learning. Finally, it is analysed in some depth how the function transformations applied can be described by the more general notion of function conjugate relation, also often used for coordinate transformations.Publication Metadata only Modeling structural protein interaction networks for betweenness analysis(Springer-Verlag Berlin, 2014) N/A; Department of Chemical and Biological Engineering; Department of Computer Engineering; Demircioğlu, Deniz; Keskin, Özlem; Gürsoy, Attila; Researcher; Faculty Member; Faculty Member; Department of Chemical and Biological Engineering; Department of Computer Engineering; N/A; College of Engineering; College of Engineering; N/A; 26605; 8745Protein-protein interactions are usually represented as interaction networks (graphs), where the proteins are represented as nodes and the connections between the interacting proteins are shown as edges. Proteins or interactions with high betweenness are considered as key connector members of the network. The interactions of a protein are dictated by its structure. In this study, we propose a new protein interaction network model taking structures of proteins into consideration. With this model, it is possible to reveal simultaneous and mutually exclusive interactions of a protein. Effect of mutually exclusive interactions on information flow in a network is studied with weighted edge betweenness analysis and it is observed that a total of 68% of bottlenecks found in p53 pathway network differed from bottlenecks found via regular edge betweenness analysis. The new network model favored the proteins which have regulatory roles and take part in cell cycle and molecular functions like protein binding, transcription factor binding, and kinase activity.Publication Metadata only Framework for traffic proportional energy efficiency in software defined networks(IEEE, 2018) N/A; Department of Computer Engineering; Assefa, Beakal Gizachew; Özkasap, Öznur; PhD Student; Faculty Member; Department of Computer Engineering; Graduate School of Sciences and Engineering; College of Engineering; N/A; 113507Software Defined Networking (SDN) achieves programmability of network elements by separating the control and the forwarding planes, and provides efficiency through optimized routing and flexibility in network management. As the energy costs contribute largely to the overall costs in networks, energy efficiency is a significant design requirement for modern networking mechanisms. However, designing energy efficient solutions is complicated since there is a trade-off between energy efficiency and network performance. In this paper, we propose traffic proportional energy efficient framework for SDN and heuristics algorithm that maintains the tradeoff between efficiency and performance. We also present IP formulation for traffic proportional energy efficiency problem. Comprehensive experiments conducted on Mininet emulator and PDX controller using Abilene, Atlanta, and Nobel-Germany real-world topologies and traffic traces show that our approach saves up to 50% energy while achieving a performance closer to the algorithms prioritizing performance.