Researcher:
Gürsoy, Mehmet Emre

Profile Picture
ORCID

Job Title

Faculty Member

First Name

Mehmet Emre

Last Name

Gürsoy

Name

Name Variants

Gürsoy, Mehmet Emre

Email Address

Birth Date

Filters

2020 - 20228

Advanced Search

Filter by

Settings

Researcher Of Publications: search.filters.isAuthorOfPublication.a9bdc1a3-2301-428e-9c3b-57ff24265cac

Search Results

Now showing 1 - 8 of 8
  • Placeholder
    PublicationMetadata only
    Detection and mitigation of targeted data poisoning attacks in federated learning
    (Institute of Electrical and Electronics Engineers Inc., 2022) Department of Computer Engineering; Department of Computer Engineering; Gürsoy, Mehmet Emre; Erbil, Pınar; Faculty Member; Student; Department of Computer Engineering; College of Engineering; College of Engineering; 330368; N/A
    Federated learning (FL) has emerged as a promising paradigm for distributed training of machine learning models. In FL, several participants train a global model collaboratively by only sharing model parameter updates while keeping their training data local. However, FL was recently shown to be vulnerable to data poisoning attacks, in which malicious participants send parameter updates derived from poisoned training data. In this paper, we focus on defending against targeted data poisoning attacks, where the attacker's goal is to make the model misbehave for a small subset of classes while the rest of the model is relatively unaffected. To defend against such attacks, we first propose a method called MAPPS for separating malicious updates from benign ones. Using MAPPS, we propose three methods for attack detection: MAPPS + X-Means, MAPPS + VAT, and their Ensemble. Then, we propose an attack mitigation approach in which a "clean"model (i.e., a model that is not negatively impacted by an attack) can be trained despite the existence of a poisoning attempt. We empirically evaluate all of our methods using popular image classification datasets. Results show that we can achieve > 95% true positive rates while incurring only < 2% false positive rate. Furthermore, the clean models that are trained using our proposed methods have accuracy comparable to models trained in an attack-free scenario.
  • Placeholder
    PublicationMetadata only
    Utility-aware and privacy-preserving mobile query services
    (2022) Yiğitoğlu, Emre; Liu, Ling; Department of Computer Engineering; Gürsoy, Mehmet Emre; Faculty Member; Department of Computer Engineering; College of Engineering; 330368
    Location-based queries enable fundamental services for mobile users. While the benefits of location-based services (LBS) are numerous, exposure of mobile users' locations to untrusted LBS providers may lead to privacy concerns. This paper proposes StarCloak, a utility-aware and attack-resilient location anonymization service for privacy-preserving LBS usage. StarCloak combines several desirable properties. First, unlike conventional approaches which are indifferent to underlying road network structure, StarCloak uses the concept of stars and proposes cloaking graphs for effective location cloaking on road networks. Second, StarCloak supports user-specified k-user anonymity and $l$-segment indistinguishability, for enabling personalized privacy protection and for serving users with varying privacy preferences. Third, StarCloak achieves strong attack-resilience against replay and query injection attacks through randomized star selection and pruning. Finally, to enable efficient query processing with high throughput and low bandwidth overhead, StarCloak makes cost-aware star selection decisions by considering query evaluation and network communication costs. We evaluate StarCloak on two datasets using real-world road networks, under various privacy and utility constraints. Results show that StarCloak achieves improved query success rate and throughput, reduced anonymization time and network usage, and higher attack-resilience in comparison to XStar, its most relevant competitor.
  • Placeholder
    PublicationMetadata only
    Forecasting daily COVID-19 case counts using aggregate mobility statistics
    (MDPI, 2022) Department of Computer Engineering; Department of Computer Engineering; Boru, Bulut; Gürsoy, Mehmet Emre; Undergraduate Student; Faculty Member; Department of Computer Engineering; College of Engineering; College of Engineering; N/A; 330368
    The COVID-19 pandemic has impacted the whole world profoundly. For managing the pandemic, the ability to forecast daily COVID-19 case counts would bring considerable benefit to governments and policymakers. In this paper, we propose to leverage aggregate mobility statistics collected from Google's Community Mobility Reports (CMRs) toward forecasting future COVID-19 case counts. We utilize features derived from the amount of daily activity in different location categories such as transit stations versus residential areas based on the time series in CMRs, as well as historical COVID-19 daily case and test counts, in forecasting future cases. Our method trains optimized regression models for different countries based on dynamic and data-driven selection of the feature set, regression type, and time period that best fit the country under consideration. The accuracy of our method is evaluated on 13 countries with diverse characteristics. Results show that our method's forecasts are highly accurate when compared to the real COVID-19 case counts. Furthermore, visual analysis shows that the peaks, plateaus and general trends in case counts are also correctly predicted by our method.
  • Placeholder
    PublicationMetadata only
    Physical activity recognition using deep transfer learning with convolutional neural networks
    (Institute of Electrical and Electronics Engineers Inc., 2022) Department of Electrical and Electronics Engineering; Department of Computer Engineering; N/A; N/A; Gürsoy, Beren Semiz; Gürsoy, Mehmet Emre; Ataseven, Berke; Madani, Alireza; Faculty Member; Faculty Member; Master Student; Master Student; Department of Electrical and Electronics Engineering; Department of Computer Engineering; College of Engineering; College of Engineering; N/A; Graduate School of Sciences and Engineering; 332403; 330368; N/A; N/A
    Current wearable devices are capable of monitoring various health indicators as well as fitness and/or physical activity types. However, even on the latest models of many wearable devices, users need to manually enter the type of work-out or physical activity they are performing. In order to automate real-time physical activity recognition, in this study, we develop a deep transfer learning-based physical activity recognition framework using acceleration data acquired through inertial measurement units (IMUs). Towards this goal, we modify a pre-trained version of the GoogLeNet convolutional neural network and fine-tune it with data from IMUs. To make IMU data compatible with GoogLeNet, we propose three novel data transform approaches based on continuous wavelet transform: Horizontal Concatenation (HC), Acceleration-Magnitude (AM), and Pixelwise Axes-Averaging (PA). We evaluate the performance of our approaches using the real-world PAMAP2 dataset. The three approaches result in 0.93, 0.95 and 0.98 validation accuracy and 0.75, 0.85 and 0.91 test accuracy, respectively. The PA approach yields the highest weighted F1 score (0.91) and activity-specific true positive ratios. Overall, our methods and results show that accurate real-time physical activity recognition can be achieved using transfer learning and convolutional neural networks.
  • Placeholder
    PublicationMetadata only
    Automatic subject identification using scale-based ballistocardiogram signals
    (Springer Science and Business Media Deutschland GmbH, 2022) Shandhi, Md Mobashir Hasan; Orlandic, Lara; Mooney, Vincent J.; Inan, Omer T.; Department of Computer Engineering; Department of Electrical and Electronics Engineering; Gürsoy, Mehmet Emre; Gürsoy, Beren Semiz; Faculty Member; Faculty Member; Department of Computer Engineering; Department of Electrical and Electronics Engineering; College of Engineering; College of Engineering; 330368; 332403
    Many electronic devices such as weighing scales, fitness equipment and medical devices are nowadays shared by multiple users. In such devices, automatic identification of device users becomes an important step towards improved user convenience and personalized service. In this paper, we propose a novel approach for subject identification using ballistocardiogram (BCG) signals collected unobtrusively from a modified weighing scale. Our approach first segments BCG signals into heartbeats using signal filtering and beat detection techniques, and averages beats to obtain smoother ensemble averaged BCG frames that are more robust to noise. Second, it extracts features related to subjects’ cardiovascular performance and musculoskeletal system from their BCG frames. Finally, it trains a machine learning model for predicting the owner of an unlabeled BCG recording based on its features. We evaluated our approach through a pilot experimental study with subjects’ BCG signals recorded at rest and following different physiological modulation. Our approach achieves up to 97% identification accuracy at rest conditions and incurs a 15–20% accuracy drop on average under physiological modulation. © 2022, ICST Institute for Computer Sciences, Social Informatics and Telecommunications Engineering.
  • Placeholder
    PublicationMetadata only
    The tsc- pfed architecture for privacy-preserving fl
    (IEEE Computer Soc, 2021) Truex, Stacey; Liu, Ling; Wei, Wenqi; Chow, Ka Ho; Department of Computer Engineering; Gürsoy, Mehmet Emre; Faculty Member; Department of Computer Engineering; College of Engineering; 330368
    In this paper we will introduce our system for trust and (s) under bar eurity enhanced (c) under bar ustomizable (p) under bar rivate federated learning: TSC-PFed. We combine secure mUItiparty computation and differential privacy to allow participants to leverage known trust dynamics which allow for increased ML model accuracy while preserving privacy guarantees and introduce an update auditor to protect against malicious participants launching dangerous label Dipping data poisoning. We additionally introduce customizable modules into the TSC-PFed ecosystem which (a) allow users to customize the type of privacy protection provided and (b) provide a tiered participant selection approach which considers variation in privacy budgets.
  • Thumbnail Image
    PublicationOpen Access
    An adversarial approach to protocol analysis and selection in local differential privacy
    (Institute of Electrical and Electronics Engineers (IEEE), 2022) Liu, L.; Chow, K.H.; Truex, S.; Wei, W.; Department of Computer Engineering; Gürsoy, Mehmet Emre; Faculty Member; Department of Computer Engineering; College of Engineering
    Local Differential Privacy (LDP) is a popular standard for privacy-preserving data collection. Numerous LDP protocols have been proposed in the literature which differ in how they provide higher utility in different settings. Yet, few have engaged in analyzing the privacy relationships of these protocols under varying settings, and consequently, it is non-trivial to select which LDP protocol is best to use in a newly emerging application. In this paper, we present an adversarial approach to protocol analysis and selection and make three original contributions. First, we introduce a Bayesian adversary to analyze the privacy relationships of LDP protocols under varying settings. We show that different protocols have substantially different responses to the attack effectiveness of the Bayesian adversary, measured in terms of Adversarial Success Rate (ASR). Second, we provide a formal and empirical analysis on a set of privacy and utility-critical factors, including encoding parameters, privacy budget, data domain, adversarial knowledge, and statistical distribution. We show that different settings of these factors have significant effects on the ASRs of LDP protocols, and no protocol provides consistently low ASR across all settings. Third, we design and develop LDPLens, a prototype implementation of our proposed framework. Given a data collection scenario with various factors and constraints, LDPLens enables optimized selection of a desirable LDP protocol for the given scenario. We evaluate the effectiveness of LDPLens using three case studies with real-world datasets. Results show that LDPLens recommends a different protocol in each case study, and the protocol recommended by LDPLens can yield up to 1.5-2 fold reduction in utility loss, ASR or privacy budget compared to a randomly selected protocol.
  • Thumbnail Image
    PublicationOpen Access
    Utility-optimized synthesis of differentially private location traces
    (Institute of Electrical and Electronics Engineers (IEEE), 2020) Department of Computer Engineering; Gürsoy, Mehmet Emre; Faculty Member; Department of Computer Engineering; College of Engineering
    Differentially private location trace synthesis (DPLTS) has recently emerged as a solution to protect mobile users' privacy while enabling the analysis and sharing of their location traces. A key challenge in DPLTS is to best preserve the utility in location trace datasets, which is non-trivial considering the high dimensionality, complexity and heterogeneity of datasets, as well as the diverse types and notions of utility. In this paper, we present OptaTrace: a utility-optimized and targeted approach to DPLTS. Given a real trace dataset D, the differential privacy parameter ϵ controlling the strength of privacy protection, and the utility/error metric Err of interest; OptaTrace uses Bayesian optimization to optimize DPLTS such that the output error (measured in terms of given metric Err) is minimized while ϵ-differential privacy is satisfied. In addition, OptaTrace introduces a utility module that contains several built-in error metrics for utility benchmarking and for choosing Err, as well as a front-end web interface for accessible and interactive DPLTS service. Experiments show that OptaTrace's optimized output can yield substantial utility improvement and error reduction compared to previous work.