Logistic regression is a powerful machine learning tool to classify data. When dealing with sensitive or private data, cares are necessary. In this paper, we propose a secure system for privacy-protecting both the training and predicting data in logistic regression via homomorphic encryption. Perhaps surprisingly, despite the non-polynomial tasks of training and predicting in logistic regression, we show that only additively homomorphic encryption is needed to build our system. Indeed, we instantiate our system with Paillier, LWE-based, and ring-LWE-based encryption schemes, highlighting the merits and demerits of each instantiation. Besides examining the costs of computation and communication, we carefully test our system over real datasets to demonstrate its utility.

In data sharing privacy has become one of the main concerns particularly when sharing datasets involving individuals contain private sensitive information. A model that is widely used to protect the privacy of individuals in publishing micro-data is k-anonymity. It reduces the linking confidence between private sensitive information and specific individual by generalizing the identifier attributes of each individual into at least k-1 others in dataset. K-anonymity can also be defined as clustering with constrain of minimum k tuples in each group. However, the accuracy of the data in k-anonymous dataset decreases due to huge information loss through generalization and suppression. Also most of the current approaches are designed for numerical continuous attributes and for categorical attributes they do not perform efficiently and depend on attributes hierarchical taxonomies, which often do not exist. In this paper we propose a new model for k-anonymization, which is called Similarity-Based Clustering (SBC). It is based on clustering and it measures similarity and calculates distances between tuples containing numerical and categorical attributes without hierarchical taxonomies. Based on this model a bottom up greedy algorithm is proposed. Our extensive study on two real datasets shows that the proposed algorithm in comparison with existing well-known algorithms offers much higher data utility and reduces the information loss significantly. Data utility is maintained above 80% in a wide range of k values.

Location-based services (LBSs) are useful for many applications in internet of things(IoT). However, LBSs has raised serious concerns about users' location privacy. In this paper, we propose a new location privacy attack in LBSs called hidden location inference attack, in which the adversary infers users' hidden locations based on the users' check-in histories. We discover three factors that influence individual check-in behaviors: geographic information, human mobility patterns and user preferences. We first separately evaluate the effects of each of these three factors on users' check-in behaviors. Next, we propose a novel algorithm that integrates the above heterogeneous factors and captures the probability of hidden location privacy leakage. Then, we design a novel privacy alert framework to warn users when their sharing behavior does not match their sharing rules. Finally, we use our experimental results to demonstrate the validity and practicality of the proposed strategy.

This article describes an approximate model of a group of cells in the wireless 4G network with implemented load balancing mechanism. An appropriately modified model of Erlang's Ideal Grading is used to model this group of cells. The model makes it possible to take into account limited availability of resources of individual cells to multi-rate elastic and adaptive traffic streams generated by Erlang and Engset sources. The developed solution allows the basic traffic characteristics in the considered system to be determined, i.e. the occupancy distribution and the blocking probability. Because of the approximate nature of the proposed model, the results obtained based on the model were compared with the results of a digital simulation. The present study validates the adopted assumptions of the proposed model.

This study investigates a real-time joint channel and hyperparameter estimation method for orthogonal frequency division multiplexing mobile communications. The channel frequency response of the pilot subcarrier and its fixed hyperparameters (such as channel statistics) are estimated using a Liu and West filter (LWF), which is based on the state-space model and sequential Monte Carlo method. For the first time, to our knowledge, we demonstrate that the conventional LWF biases the hyperparameter due to a poor estimate of the likelihood caused by overfitting in noisy environments. Moreover, this problem cannot be solved by conventional smoothing techniques. For this, we modify the conventional LWF and regularize the likelihood using a Kalman smoother. The effectiveness of the proposed method is confirmed via numerical analysis. When both of the Doppler frequency and delay spread hyperparameters are unknown, the conventional LWF significantly degrades the performance, sometimes below that of least squares estimation. By avoiding the hyperparameter estimation failure, our method outperforms the conventional approach and achieves good performance near the lower bound. The coding gain in our proposed method is at most 10 dB higher than that in the conventional LWF. Thus, the proposed method improves the channel and hyperparameter estimation accuracy. Derived from mathematical principles, our proposal is applicable not only to wireless technology but also to a broad range of related areas such as machine learning and econometrics.

We introduce the design and deployment of the latest version of the VNode infrastructure, VNode-i. We present new extended VNode-i functions that offer high performance and provide convenient deep programmability to network developers. We extend resource abstraction to the transport network and achieve highly precise slice measurement for resource elasticity. We achieve precise resource isolation for VNode-i. We achieve coexistence of high performance and programmability. We also enhance AGW functions. In addition, we extend network virtualization from the core network to edge networks and terminals. In evaluation experiments, we deploy the enhanced VNode-i on the JGN-X testbed and evaluate its performance. We successfully create international federation slices across VNode-i, GENI, and Fed4FIRE. We also present experimental results on video streaming on a federated slice across VNode-i and GENI. Testbed experiments confirm the practicality of the enhanced VNode-i.

Fine-grained power gating (FGPG) is a power-saving technique by switching off circuit blocks while the blocks are idle. Although FGPG can reduce power consumption without compromising computational performance, switching the power supply on and off causes energy overhead. To prevent power increase caused by the energy overhead, in our prior research we proposed an FGPG control method of the operating system(OS) based on pre-analyzing applications' power usage. However, modern computing systems have a wide variety of use cases and run many types of application; this makes it difficult to analyze the behavior of all these applications in advance. This paper therefore proposes a new FGPG control method without profiling application programs in advance. In the new proposed method, the OS monitors a circuit's idle interval periodically while application programs are running. The OS enables FGPG only if the interval time is long enough to reduce the power consumption. The experimental results in this paper show that the proposed method reduces power consumption by 9.8% on average and up to 17.2% at 25°C. The results also show that the proposed method achieves almost the same power-saving efficiency as the previous profile-based method.

This paper addresses the issue of implementing a sequence for restoring fiber links and communication paths that have failed due to a catastrophe. We present a mathematical formulation to minimize the total number of steps needed to restore communication paths. We also propose two heuristic algorithms: Minimum spanning tree - based degree order restoration and Congestion link order restoration. Numerical evaluations show that integer linear programming based order restoration yields the fewest number of restoration steps, and that the proposed heuristic algorithms, when used properly with regard to the accommodation rate, are highly effective for real-world networks.

We have developed software that uses the R statistics software environment to automatically generate tactile graphs — i.e. graphs that can be read by blind people using their sense of touch. We released this software as a Web application to make it available to anyone, from anywhere. This Web application can automatically generate images for tactile graphs from numerical data in a CSV file. It is currently able to generate four types of graph — scatter plots, line graphs, bar charts and pie charts. This paper describes the Web application's functions, operating procedures and the results of evaluation experiments.

This paper proposes an interference suppression scheme based on linear combining for multiple relay systems. Interference from base stations and relays in neighboring cells degrades the bit error rate (BER) performance of mobile stations (MSs) near cell boundaries. To suppress such interference for half-duplex relay systems, the proposed scheme linearly combines received signals of the first and second phases at MS. Without channel state information (CSI) feedback, weight coefficients for the linear combining are estimated by the recursive least-squares (RLS) algorithm, which requires only information on preamble symbols of the target MS. Computer simulations of orthogonal frequency-division multiplexing (OFDM) transmission under two-cell and frequency selective fading conditions are conducted. It is demonstrated that the RLS-based linear combining with decision directed estimation is superior to the RLS-based linear combining using only the preamble and can outperform the minimum mean-squared error (MMSE) combining with estimated CSI when the number of preamble symbols is two and four that correspond to the minimum requirements for MMSE and RLS, respectively.

Massive Open Online Courses (MOOC) have been invented to support Virtual Learning Environment (VLE) for higher education. While numerous learning courses and contents were authored, most of the existing resources are now hard to reuse/redistribute among instructors due to the privacy of the contents. Therefore, Open Educational Resources (OER) and the Creative Commons license (CC) are interesting solutions available to alleviate such problems of MOOC. This research presents a new framework that effectively connects OER and MOOC for a life-long e-Learning platform for Thai people. We utilize the Fedora Commons repository for an OER back-end, and develop a new front-end to manage OER resources. In addition, we introduce a “FedX API” - including a packet encapsulation and a data transmission module - that organizes educational resources between both systems. We also proposed the CC declaring function to help participants on-the-fly declare their content license; therefore, any resources must be granted as an open licensing. Another important function is a Central Authorized System (CAS) which is applied to develop single signing-on to facilitate the OER-MOOC connection. Since the framework is designed to support the massive demand, the concurrent access capability is also evaluated to measure the performance of the proposed framework. The results show that the proposed framework can provide up to 750 concurrencies without any defects. The FedX API does not produce bottleneck trouble on the interoperability framework in any cases. In addition, resources can be exchanged among the third-party OER repositories by an OAI-PMH harvesting tool.

In this paper, we propose a single-channel speech enhancement method for a push-to-talk enabled wireless communication device. The proposed method is based on adaptive weighted β-order spectral amplitude estimation under speech presence uncertainty and enhanced instantaneous phase estimation in order to achieve flexible and effective noise reduction while limiting the speech distortion due to different noise conditions. Experimental results confirm that the proposed method delivers higher voice quality and intelligibility than the reference methods in various noise environments.

Radio Frequency Identification (RFID) plays a crucial role in IoT development. With the extensive use of RFID, the fact that a single RFID tag integrates multiple applications has become a mainstream. To facilitate users to use the multi-application RFID tag and revoke some applications in the tag securely and efficiently, a secure RFID application revocation scheme is proposed in this paper. In the scheme, each response for the challenge between tag and reader is different, and a group has the feature of many tags. Even if the group index number and corresponding group are revealed, a specific tag does not be precisely found and tracked. Users are anonymous completely. The scheme also allows users to set the validity period for an application or some applications. If the application contains the validity period and expires, the server will remove the validity period and revoke the application automatically in the tag when the RFID tag accesses server again. The proposed scheme cannot only be used in multi-application RFID tag but also be used in one-application RFID tag. Furthermore, compared with other existing schemes, the scheme provides a higher level of security and has an advantage of performance. Our scheme has the ability of mutual authentication and Anti-replay by adding a random number r2, and it is easy to against synchronization attack. Security proof is given in our paper and performance advantage are mainly reflected in the following points such as forward security, synchronization, storage complexity, computational complexity, etc. Finally, the proposed scheme can be used in multi-application RFID tag to promote the development of the IoT.

Oblivious RAM is a technique for hiding the access patterns between a client and an untrusted server. However, current ORAM algorithms incur large communication or storage overhead. We propose a novel ORAM construction using a matrix logical structure for server storage where a client downloads blocks from each row, choosing the column randomly to hide the access pattern. Both a normal construction and recursive construction, where a position map normally stored on the client is also stored on the server, are presented. We show our matrix ORAM achieves constant bandwidth cost for the normal construction, uses similar storage to the existing Path ORAM, and improves open the bandwidth cost compared to Path ORAM under certain conditions in the recursive construction.

In this paper, we propose a multi-cell structure backscatter based wireless-powered communication network (WPCN) where a number of backscatter cells are locally separated, each containing a subset of users around a carrier emitter. The multi-cell structure backscatter based WPCN can be implemented in two ways, namely time-division multiplexing (TDM) and frequency-division multiplexing (FDM). Here users harvest energy from the carrier signal transmitted by the carrier emitter, and then transmit their own information in a passive way via the reflection of the carrier signal using frequency-shift keying modulation. We characterize the energy-free condition and the signal-to-noise ratio (SNR) outage zone in a backscatter based WPCN. Also, a backscatter based harvest-then-transmit protocol is adopted to maximize the sum-throughput of all users by optimally allocating time for energy harvesting and information transmission. Numerical results demonstrate that the backscatter based WPCN ensures an increased long-range coverage and a diminished SNR outage zone compared to conventional radio based WPCNs. Also, comparing the two types of multi-cell structure backscatter based WPCN, TDM within each backscatter cell and FDM across backscatter cells versus FDM within each backscatter cell and TDM across backscatter cells, numerical results confirm that which one yields a better performance.

When a massive network disruption occurs, repair of the damaged network takes time, and the recovery process involves multiple stages. We propose a fast and flow-controlled multi-stage network recovery method for determining the pareto-optimal recovery order of failed physical components reflecting the balance requirement between maximizing the total amount of traffic on all logical paths, called total network flow, and providing adequate logical path flows. The pareto-optimal problem is formulated by mixed integer linear programming (MILP). A heuristic algorithm, called the grouped-stage recovery (GSR), is also introduced to solve the problem when the problem formulated by MILP is computationally intractable in a large-scale failure. The effectiveness of the proposed method was numerically evaluated. The results show that the pareto-optimal recovery order can be determined from the balance between total network flow and adequate logical path flows, the allocated minimum bandwidth of the logical path can be drastically improved while maximizing total network flow, and the proposed method with GSR is applicable to large-scale failures because a nearly optimal recovery order with less than 10% difference rate can be determined within practical computation time.

This paper investigates a frame aggregation (FA) technique in the medium access control (MAC) layer for downlink multi-user multiple input multiple output (MU-MIMO) channels in wireless local area networks (WLANs), and proposes a high-efficient FA scheme that ehances system performance: transmission performance and fairness in communication between mobile terminals (MTs). The proposed FA scheme employs novel criteria for selecting receiving MTs and wireless frame setting with a frame size adaptation mechanism for MU-MIMO transmissions. The proposed receiving MT selection gives higher priority to the MTs expecting higher throughput in the next MU-MIMO transmission and having large amount transmission data while reducing signaling overhead, leading to improvements in system throughput and fairness in communication. The proposed wireless frame setting, which employs hybrid A-MSDU/A-MPDU FA, achieves frame error rate (FER) better than the requirement from communication services by using A-MSDU frame size adaptation. Through system-level simulation, the effectiveness of the proposed scheme is validated for downlink MU-MIMO channels in WLANs.

Biphase periodic sequences having elements +1 or -1 with the two-level autocorrelation function are desirable in communications and radars. However, in case of the biphase orthogonal periodic sequences, Turyn has conjectured that there exist only sequences with period 4, i.e., there exist the circulant Hadamard matrices for order 4 only. In this paper, it is described that the conjecture is proved to be true by means of the isomorphic mapping, the Chinese remainder theorem, the linear algebra, etc.

The finite-difference time-domain (FDTD) method has been widely used in recent years to analyze the propagation and scattering of electromagnetic waves. Because the FDTD method has second-order accuracy in space, its numerical dispersion error arises from truncated higher-order terms of the Taylor expansion. This error increases with the propagation distance in cases of large-scale analysis. The numerical dispersion error is expressed by a dispersion relation equation. It is difficult to solve this nonlinear equation which have many parameters. Consequently, a simple formula is necessary to substitute for the dispersion relation error. In this study, we have obtained a simple formula for the numerical dispersion error of 2-D and 3-D FDTD method in free space propagation.

Digital phase shifters are widely used to achieve space scanning in phased array antenna, and beam pointing accuracy depends on the bit number and resolution of the digital phase shifter. This paper proposes a novel phase feeding method to reduce the phase quantization error effects. A linear formula for the beam pointing deviation of a linear uniform array in condition of phase quantization error is derived, and the linear programming algorithm is introduced to achieve the minimum beam pointing deviation. Simulations are based on the pattern of the phased array, which gives each element a certain quantization phase error to find the beam pointing deviation. The novel method is then compared with previous methods. Examples show that a 32-element uniform linear array with 5-bit phase shifters using the proposed method can achieve a higher beam-steering accuracy than the same array with 11-bit phase shifters.