Search

Differential privacy has emerged as a practical technique for privacy-preserving deep learning. However, recent studies on privacy attacks have demonstrated vulnerabilities in the existing differential privacy implementations for deep models. While encryption-based methods offer robust security, their computational overheads are often prohibitive. To address these challenges, we propose a novel differential privacy-based image generation method. Our approach employs two distinct noise types: one makes the image unrecognizable to humans, preserving privacy during transmission, while the other maintains features essential for machine learning analysis. This allows the deep learning service to provide accurate results, without compromising data privacy. We demonstrate the feasibility of our method on the CIFAR100 dataset, which offers a realistic complexity for evaluation. [ABSTRACT FROM AUTHOR]

In recent years, the number and sophistication of malware attacks on computer systems have increased significantly. One technique employed by malware authors to evade detection and analysis, known as Heaven's Gate, enables 64-bit code to run within a 32-bit process. Heaven's Gate exploits a feature in the operating system that allows the transition from a 32-bit mode to a 64-bit mode during execution, enabling the malware to evade detection by security software designed to monitor only 32-bit processes. Heaven's Gate poses significant challenges for existing security tools, including dynamic binary instrumentation (DBI) tools, widely used for program analysis, unpacking, and de-virtualization. In this paper, we provide a comprehensive analysis of the Heaven's Gate technique. We also propose a novel approach to bypass the Heaven's Gate technique using black-box testing. Our experimental results show that the proposed approach effectively bypasses and prevents the Heaven's Gate technique and strengthens the capabilities of DBI tools in combating advanced malware threats. [ABSTRACT FROM AUTHOR]

Data are needed to train machine learning (ML) algorithms, and in many cases often include private datasets that contain sensitive information. To preserve the privacy of data used while training ML algorithms, computer scientists have widely deployed anonymization techniques. These anonymization techniques have been widely used but are not foolproof. Many studies showed that ML models using anonymization techniques are vulnerable to various privacy attacks willing to expose sensitive information. As a privacy-preserving machine learning (PPML) technique that protects private data with sensitive information in ML, we propose a new task-specific adaptive differential privacy (DP) technique for structured data. The main idea of the proposed DP method is to adaptively calibrate the amount and distribution of random noise applied to each attribute according to the feature importance for the specific tasks of ML models and different types of data. From experimental results under various datasets, tasks of ML models, different DP mechanisms, and so on, we evaluate the effectiveness of the proposed task-specific adaptive DP method. Thus, we show that the proposed task-specific adaptive DP technique satisfies the model-agnostic property to be applied to a wide range of ML tasks and various types of data while resolving the privacy–utility trade-off problem. [ABSTRACT FROM AUTHOR]

Chronic kidney disease (CKD) progression involves morphological changes in the kidney, such as decreased length and thickness, with associated histopathological alterations. However, the relationship between morphological changes in the kidneys and glomerular filtration rate (GFR) has not been quantitatively and comprehensively evaluated. We evaluated the three-dimensional size and shape of the kidney using computed tomography (CT)-derived features in relation to kidney function. We included 257 patients aged ≥18 years who underwent non-contrast abdominal CT at the Inha University Hospital. The features were quantified using predefined algorithms in the pyRadiomics package after kidney segmentation. All features, except for flatness, significantly correlated with estimated GFR (eGFR). The surface-area-to-volume ratio (SVR) showed the strongest negative correlation (r = −0.75, p < 0.0001). Kidney size features, such as volume and diameter, showed moderate to high positive correlations; other morphological features showed low to moderate correlations. The calculated area under the receiver operating characteristic (ROC) curve (AUC) for different features ranged from 0.51 (for elongation) to 0.86 (for SVR) for different eGFR thresholds. Diabetes patients had weaker correlations between the studied features and eGFR and showed less bumpy surfaces in three-dimensional visualization. We identified alterations in the CKD kidney based on various three-dimensional shape and size features, with their potential diagnostic value. [ABSTRACT FROM AUTHOR]

With intentional feature perturbations to a deep learning model, the adversary generates an adversarial example to deceive the deep learning model. As an adversarial example has recently been considered in the most severe problem of deep learning technology, its defense methods have been actively studied. Such effective defense methods against adversarial examples are categorized into one of the three architectures: (1) model retraining architecture; (2) input transformation architecture; and (3) adversarial example detection architecture. Especially, defense methods using adversarial example detection architecture have been actively studied. This is because defense methods using adversarial example detection architecture do not make wrong decisions for the legitimate input data while others do. In this paper, we note that current defense methods using adversarial example detection architecture can classify the input data into only either a legitimate one or an adversarial one. That is, the current defense methods using adversarial example detection architecture can only detect the adversarial examples and cannot classify the input data into multiple classes of data, i.e., legitimate input data and various types of adversarial examples. To classify the input data into multiple classes of data while increasing the accuracy of the clustering model, we propose an advanced defense method using adversarial example detection architecture, which extracts the key features from the input data and feeds the extracted features into a clustering model. From the experimental results under various application datasets, we show that the proposed method can detect the adversarial examples while classifying the types of adversarial examples. We also show that the accuracy of the proposed method outperforms the accuracy of recent defense methods using adversarial example detection architecture. [ABSTRACT FROM AUTHOR]

In this paper, we propose a new deep learning-based image translation method to predict and generate images after hair transplant surgery from images before hair transplant surgery. Since existing image translation models use a naive strategy that trains the whole distribution of translation, the image translation models using the original image as the input data result in converting not only the hair transplant surgery region, which is the region of interest (ROI) for image translation, but also the other image regions, which are not the ROI. To solve this problem, we proposed a novel generative adversarial network (GAN)-based ROI image translation method, which converts only the ROI and retains the image for the non-ROI. Specifically, by performing image translation and image segmentation independently, the proposed method generates predictive images from the distribution of images after hair transplant surgery and specifies the ROI to be used for generated images. In addition, by applying the ensemble method to image segmentation, we propose a more robust method through complementing the shortages of various image segmentation models. From the experimental results using a real medical image dataset, e.g., 1394 images before hair transplantation and 896 images after hair transplantation, to train the GAN model, we show that the proposed GAN-based ROI image translation method performed better than the other GAN-based image translation methods, e.g., by 23% in SSIM (Structural Similarity Index Measure), 452% in IoU (Intersection over Union), and 42% in FID (Frechet Inception Distance), on average. Furthermore, the ensemble method that we propose not only improves ROI detection performance but also shows consistent performances in generating better predictive images from preoperative images taken from diverse angles. [ABSTRACT FROM AUTHOR]

As the amount of data collected and analyzed by machine learning technology increases, data that can identify individuals is also being collected in large quantities. In particular, as deep learning technology—which requires a large amount of analysis data—is activated in various service fields, the possibility of exposing sensitive information of users increases, and the user privacy problem is growing more than ever. As a solution to this user's data privacy problem, homomorphic encryption technology, which is an encryption technology that supports arithmetic operations using encrypted data, has been applied to various field including finance and health care in recent years. If so, is it possible to use the deep learning service while preserving the data privacy of users by using the data to which homomorphic encryption is applied? In this paper, we propose three attack methods to infringe user's data privacy by exploiting possible security vulnerabilities in the process of using homomorphic encryption-based deep learning services for the first time. To specify and verify the feasibility of exploiting possible security vulnerabilities, we propose three attacks: (1) an adversarial attack exploiting communication link between client and trusted party; (2) a reconstruction attack using the paired input and output data; and (3) a membership inference attack by malicious insider. In addition, we describe real-world exploit scenarios for financial and medical services. From the experimental evaluation results, we show that the adversarial example and reconstruction attacks are a practical threat to homomorphic encryption-based deep learning models. The adversarial attack decreased average classification accuracy from 0.927 to 0.043, and the reconstruction attack showed average reclassification accuracy of 0.888, respectively. [ABSTRACT FROM AUTHOR]

With the development of wireless networks and mobile devices, interest on indoor localization systems (ILSs) has increased. In particular, Wi-Fi-based ILSs are widely used because of the good prediction accuracy without additional hardware. However, as the prediction accuracy decreases in environments with natural noise, some studies were conducted to remove it. So far, two representative methods, i.e., the filtering-based method and deep learning-based method, have shown a significant effect in removing natural noise. However, the prediction accuracy of these methods severely decreased under artificial noise caused by adversaries. In this paper, we introduce a new media access control (MAC) spoofing attack scenario injecting artificial noise, where the prediction accuracy of Wi-Fi-based indoor localization system significantly decreases. We also propose a new deep learning-based indoor localization method using random forest(RF)-filter to provide the good prediction accuracy under the new MAC spoofing attack scenario. From the experimental results, we show that the proposed indoor localization method provides much higher prediction accuracy than the previous methods in environments with artificial noise. [ABSTRACT FROM AUTHOR]

Basic safety message (BSM) are messages that contain core elements of a vehicle such as vehicle's size, position, speed, acceleration and others. BSM are lightweight messages that can be regularly broadcast by the vehicles to enable a variety of applications. On the other hand, event-driven message (EDM) are messages generated at the time of occurrence such as accidents or roads sliding and can contain much more heavy elements including pictures, audio or videos. Security, architecture and communication solutions for BSM use cases have been largely documented on in the literature contrary to EDM due to several concerns such as the variant size of EDM, the appropriate architecture along with latency, privacy and security. In this paper, we propose a secure and blockchain based EDM protocol for 5G enabled vehicular edge computing. To offer scalability and latency for the proposed scenario, we adopt a 5G cellular architecture due to its projected features compared to 4G tong-term evaluation (LTE) for vehicular communications. We consider edge computing to provide local processing of EDM that can improve the response time of public agencies (ambulances or rescue teams) that may intervene to the scene. We make use of lightweight multi-receiver signcryption scheme without pairing that offers low time consuming operations, security, privacy and access control. EDM records need to be kept into a distributed system which can guarantee reliability and auditability of EDM. To achieve this, we construct a private blockchain based on the edge nodes to store EDM records. The performance analysis of the proposed protocol confirms its efficiency. [ABSTRACT FROM AUTHOR]

There is a strong devotion in the automotive industry to be part of a wider progression towards the Fifth Generation (5G) era. In-vehicle integration costs between cellular and vehicle-to-vehicle networks using Dedicated Short Range Communication could be avoided by adopting Cellular Vehicle-to-Everything (C-V2X) technology with the possibility to re-use the existing mobile network infrastructure. More and more, with the emergence of Software Defined Networks, the flexibility and the programmability of the network have not only impacted the design of new vehicular network architectures but also the implementation of V2X services in future intelligent transportation systems. In this paper, we define the concepts that help evaluate software-defined-based vehicular network systems in the literature based on their modeling and implementation schemes. We first overview the current studies available in the literature on C-V2X technology in support of V2X applications. We then present the different architectures and their underlying system models for LTE-V2X communications. We later describe the key ideas of software-defined networks and their concepts for V2X services. Lastly, we provide a comparative analysis of existing SDN-based vehicular network system grouped according to their modeling and simulation concepts. We provide a discussion and highlight vehicular ad-hoc networks' challenges handled by SDN-based vehicular networks. [ABSTRACT FROM AUTHOR]

Helicobacter pylori (H. pylori) may be involved in diabetes and other insulin-related processes. This study aimed to investigate the associations between H. pylori infection and the risks of type 2 diabetes, impaired glucose tolerance (IGT), diabetic nephropathy, and poor glycemic control. We retrospectively evaluated 16,091 subjects without diabetes at baseline who underwent repeated health examinations. Subjects were categorized according to whether they were seropositive and seronegative for H. pylori infection. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using Cox proportional hazard models. The serological results were validated using an independent cohort (n = 42,351) based on a histological diagnosis of H. pylori infection. During 108,614 person-years of follow-up, 1338 subjects (8.3%) developed newly diagnosed diabetes, although the cumulative incidence of diabetes was not significantly related to serological H. pylori status. The multivariate Cox proportional-hazards regression models revealed that H. pylori seropositivity was not significantly associated with diabetes (HR: 1.01, 95% CI: 0.88–1.16; p = 0.854), IGT (HR: 0.98, 95% CI: 0.93–1.04; p = 0.566), diabetic nephropathy (HR: 0.99, 95% CI: 0.82–1.21; p = 0.952), or poor glycemic control (HR: 1.05, 95% CI: 0.90–1.22; p = 0.535). Similarly, histopathological findings of H. pylori infection were not significantly associated with diabetes (p = 0.311), diabetic nephropathy (p = 0.888), or poor glycemic control (p = 0.989). The findings from these large Korean cohorts indicate that there does not appear to be a role for past H. pylori infection in the development of diabetes, IGT, diabetic nephropathy, or poor glycemic control. [ABSTRACT FROM AUTHOR]

The association between obesity and peptic ulcer disease (PUD) is inconclusive. To evaluate the association of obesity and metabolically healthy obesity (MHO) with PUD risk, we performed a retrospective cohort study of 32,472 subjects without PUD at baseline who underwent repeated health examinations. Participants were stratified by body mass index (BMI) and metabolically healthy state. Hazard ratios (HRs) and 95% confidence intervals (CIs) were calculated using Cox proportional hazard modelling. During the follow-up period, 1940 PUD cases occurred. PUD, particularly gastric ulcer (GU), had significantly higher cumulative incidence in obese subjects compared to non-obese subjects (p value < 0.001). The HR for developing GU was 1.32 (95% CI, 1.16–1.49; p value <0.001); after adjusting for confounding factors (lifestyle, metabolic, and Helicobacter pylori status), the association was no more significant (p value = 0.789). For duodenal ulcer (DU), cumulative incidence between obese and non-obese groups was not significantly different (p value = 0.464). The risk of developing DU in the obese group was not significantly different from the non-obese group (HR 0.95; 95% CI, 0.83–1.09; p value = 0.469) and consistently showed no association after adjusting for metabolic parameters (p value = 0.199). Furthermore, MHO subjects had no increase in GU or DU risks. In this large cohort study, PUD risk was not associated with obesity or MHO. [ABSTRACT FROM AUTHOR]

This paper proposes a real-time position accuracy improvement method for a low-cost global positioning system (GPS), which uses geographic data for forming a digital road database in the digital map information. We link the vehicle's location to the position on the digital map using the map-matching algorithm to improve the position accuracy. In the proposed method, we can distinguish the vehicle direction on the road and enhance the horizontal accuracy using the geographic data composed of the vector point set of the digital map. We use the iterative closest point (ICP) algorithm that calculates the rotation matrix and the translation vector to compensate for the disparity between the GPS and the digital map information. We also use the least squares method to correct the error caused by the rotation of the ICP algorithm and link on the digital map to eliminate the residual disparity. Finally, we implement the proposed method in real time with a low-cost embedded system and demonstrate the effectiveness of the proposed method through various experiments. [ABSTRACT FROM AUTHOR]

Basic safety message (BSM) are messages that contain core elements of a vehicle such as vehicle's size, position, speed, acceleration and others. BSM are lightweight messages that can be regularly broadcast by the vehicles to enable a variety of applications. On the other hand, event-driven message (EDM) are messages generated at the time of occurrence such as accidents or roads sliding and can contain much more heavy elements including pictures, audio or videos. Security, architecture and communication solutions for BSM use cases have been largely documented on in the literature contrary to EDM due to several concerns such as the variant size of EDM, the appropriate architecture along with latency, privacy and security. In this paper, we propose a secure and blockchain based EDM protocol for 5G enabled vehicular edge computing. To offer scalability and latency for the proposed scenario, we adopt a 5G cellular architecture due to its projected features compared to 4G tong-term evaluation (LTE) for vehicular communications. We consider edge computing to provide local processing of EDM that can improve the response time of public agencies (ambulances or rescue teams) that may intervene to the scene. We make use of lightweight multi-receiver signcryption scheme without pairing that offers low time consuming operations, security, privacy and access control. EDM records need to be kept into a distributed system which can guarantee reliability and auditability of EDM. To achieve this, we construct a private blockchain based on the edge nodes to store EDM records. The performance analysis of the proposed protocol confirms its efficiency.

In this paper, we are interested in computing ZIP code proximity from two perspectives, proximity between two ZIP codes ( Ad-Hoc ) and neighborhood proximity ( Top-K ). Such a computation can be used for ZIP code-based target marketing as one of the smart city applications. A naïve approach to this computation is the usage of the distance between ZIP codes. We redefine a distance metric combining the centroid distance with the intersecting road network between ZIP codes by using a weighted sum method. Furthermore, we prove that the results of our combined approach conform to the characteristics of distance measurement. We have proposed a general and heuristic approach for computing Ad-Hoc proximity, while for computing Top-K proximity, we have proposed a general approach only. Our experimental results indicate that our approaches are verifiable and effective in reducing the execution time and search space., Competing Interests: The authors declare no conflict of interest.

In the vehicular networks, the real-time video reporting service is used to send the recorded videos in the vehicle to the cloud. However, when facilitating the real-time video reporting service in the vehicular networks, the usage of the fourth generation (4G) long term evolution (LTE) was proved to suffer from latency while the IEEE 802.11p standard does not offer sufficient scalability for a such congested environment. To overcome those drawbacks, the fifth-generation (5G)-enabled vehicular network is considered as a promising technology for empowering the real-time video reporting service. In this paper, we note that security and privacy related issues should also be carefully addressed to boost the early adoption of 5G-enabled vehicular networks. There exist a few research works for secure video reporting service in 5G-enabled vehicular networks. However, their usage is limited because of public key certificates and expensive pairing operations. Thus, we propose a secure and lightweight protocol for cloud-assisted video reporting service in 5G-enabled vehicular networks. Compared to the conventional public key certificates, the proposed protocol achieves entities' authorization through anonymous credential. Also, by using lightweight security primitives instead of expensive bilinear pairing operations, the proposed protocol minimizes the computational overhead. From the evaluation results, we show that the proposed protocol takes the smaller computation and communication time for the cryptographic primitives than that of the well-known Eiza-Ni-Shi protocol., Competing Interests: All the authors confirm that there is no conflict of interest.