In this paper, we propose a design of reversible adder/subtractor blocks and arithmetic logic units (ALUs). The main concept of our approach is different from that of the existing related studies; we emphasize the function design. Our approach of investigating the reversible functions includes (a) the embedding of irreversible functions into incompletely-specified reversible functions, (b) the operation assignment, and (c) the permutation of function outputs. We give some extensions of these techniques for further improvements in the design of reversible functions. The resulting reversible circuits are smaller than that of the existing design in terms of the number of multiple-control Toffoli gates. To evaluate the quantum cost of the obtained circuits, we convert the circuits to reduced quantum circuits for experiments. The results also show the superiority of our realization of adder/subtractor blocks and ALUs in quantum cost.

Bit-interleaved coded modulation with iterative decoding (BICM-ID) is suitable for correlated Rayleigh fading channels. Additionally, BICM-ID using differential encoding can avoid the pilot overhead. In this paper, we consider BICM-ID using 16-DAPSK (differential amplitude and phase-shift keying). We first derive the probability of receiving signals conditioned on the transmission of input bits for general differential encoding; then we propose two new 16-DAPSK bit labeling methods. In addition, convolutional codes for the new bit labeling are developed. Both the minimum distance and the simulation results show that the proposed labeling has better error performance than that of the original differential encoding, and the searched new codes can further improve the error performance.

In this paper, we propose a novel primary user detection scheme for spectrum sensing in cognitive radio. Inspired by the conventional signal classification approach, the spectrum sensing is translated into a classification problem. On the basis of feature-based classification, the spectral correlation of a second-order cyclostationary analysis is applied as the feature extraction method, whereas a stacked denoising autoencoders network is applied as the classifier. Two training methods for signal detection, interception-based detection and simulation-based detection, are considered, for different prior information and implementation conditions. In an interception-based detection method, inspired by the two-step sensing, we obtain training data from the interception of actual signals after a sophisticated sensing procedure, to achieve detection without priori information. In addition, benefiting from practical training data, this interception-based detection is superior under actual transmission environment conditions. The alternative, a simulation-based detection method utilizes some undisguised parameters of the primary user in the spectrum of interest. Owing to the diversified predetermined training data, simulation-based detection exhibits transcendental robustness against harsh noise environments, although it demands a more complicated classifier network structure. Additionally, for the above-described training methods, we discuss the classifier complexity over implementation conditions and the trade-off between robustness and detection performance. The simulation results show the advantages of the proposed method over conventional spectrum-sensing schemes.

Huge amounts of software appear nowadays. The more the number of software increases, the more increased software vulnerabilities are. Although some automatic methods have been proposed in order to detect and remove software vulnerabilities, they still require a lot of time so they have a limitation in the real world. To solve this problem, we propose BugHunter which automatically tests a binary file compiled with a C++ compiler. It searches for unsafe API calls and automatically executes to the program block that have an unsafe API call. Also, we showed that BugHunter is more efficient than angr through experiments. As a result, BugHunter is very helpful to find a software vulnerability in a short time.

Attribute-based encryption (ABE), a cryptographic primitive, realizes fine-grained access control. Because of its attractive functionality, many systems based on ABE have been constructed to date. In such cryptographic systems, revocation functionality is indispensable to handle withdrawal of users, secret key exposure, and others. Although many ABE schemes with various functionalities have been proposed, only a few of these are revocable ABE (RABE). In this paper, we propose two generic constructions of RABE from ABE. Our first construction employs the pair encoding framework (Attrapadung, EUROCRYPT 2014), and combines identity-based revocation and ABE via the generic conjunctive conversion of Attrapadung and Yamada (CT-RSA 2015). Our second construction converts ABE to RABE directly when ABE supports Boolean formulae. Because our constructions preserve functionalities of the underlying ABE, we can instantiate various fully secure RABE schemes for the first time, e.g., supporting regular languages, with unbounded attribute size and policy structure, and with constant-size ciphertext and secret key.

The-state-of-the-art neural quality estimation (QE) of machine translation model consists of two sub-networks that are tuned separately, a bidirectional recurrent neural network (RNN) encoder-decoder trained for neural machine translation, called the predictor, and an RNN trained for sentence-level QE tasks, called the estimator. We propose to combine the two sub-networks into a whole neural network, called the unified neural network. When training, the bidirectional RNN encoder-decoder are initialized and pre-trained with the bilingual parallel corpus, and then, the networks are trained jointly to minimize the mean absolute error over the QE training samples. Compared with the predictor and estimator approach, the use of a unified neural network helps to train the parameters of the neural networks that are more suitable for the QE task. Experimental results on the benchmark data set of the WMT17 sentence-level QE shared task show that the proposed unified neural network approach consistently outperforms the predictor and estimator approach and significantly outperforms the other baseline QE approaches.

8B/10B is an encoding technique largely used in different communication protocols, with several advantages such as zero DC bias. In the last years transmission rates have grown rapidly, thus the need of encoders with better performance in terms of throughput, area and power consumption raised rapidly. In this article we will present and discuss the architecture of two symbols parallel encoder, comparing it with a classical pipelined solution.

Channel capacity is a useful numerical index not only for grasping the upper limit of the transmission bit rate but also for comparing the abilities of various digital transmission schemes commonly used in radio-wave propagation environments because the channel capacity does not depend on specific communication methods such as modulation/demodulation schemes or error correction schemes. In this paper, modeling of the noncoherent capacity in a highly underspread WSSUS channel is investigated using a new approach. Unlike the conventional method, namely, the information theoretic method, a very straightforward formula can be obtained in a statistical manner. Although the modeling in the present study is carried out using a somewhat less rigorous approach, the result obtained is useful for roughly understanding the channel capacity in doubly selective fading environments. We clarify that the radio wave propagation parameter of the spread factor, which is the product of the Doppler spread and the delay spread, can be related quantitatively to the effective maximum signal-to-interference ratio by a simple formula. Using this model, the physical limit of wireless digital transmission is discussed from a radio wave propagation perspective.

In this paper, we propose a novel self-supervised learning of video representation which is capable to anticipate the video category by only reading its short clip. The key idea is that we employ the Siamese convolutional network to model the self-supervised feature learning as two different image matching problems. By using frame encoding, the proposed video representation could be extracted from different temporal scales. We refine the training process via a motion-based temporal segmentation strategy. The learned representations for videos can be not only applied to action anticipation, but also to action recognition. We verify the effectiveness of the proposed approach on both action anticipation and action recognition using two datasets namely UCF101 and HMDB51. The experiments show that we can achieve comparable results with the state-of-the-art self-supervised learning methods on both tasks.

In mobile and wireless networks, controlling data delivery latency is one of open problems due to the stochastic nature of wireless channels, which are inherently unreliable. This paper explores how the current best-effort throughput-oriented wireless services might evolve into latency-sensitive enablers of new mobile applications such as remote three-dimensional (3D) graphical rendering for interactive virtual/augmented-reality overlay. Assuming that the signal propagation delay and achievable throughput meet the standard latency requirements of the user application, we examine the idea of trading excess/federated bandwidth for the elimination of non-negligible delay of data re-ordering, caused by temporal transmission failures and buffer overflows. The general system design is based on (i) spatially diverse data delivery over multiple paths with uncorrelated outage likelihoods; and (ii) forward packet-loss protection (FPP), creating encoding redundancy for proactive recovery of intolerably delayed data without end-to-end retransmissions. Analysis and evaluation are based on traces of real life traffic, which is measured in live carrier-grade long term evolution (LTE) networks and campus WiFi networks, due to no such system/environment yet to verify the importance of spatial diversity and encoding redundancy. Analysis and evaluation reveal the seriousness of the latency problem and that the proposed FPP with spatial diversity and encoding redundancy can minimize the delay of re-ordering. Moreover, a novel FPP effectiveness coefficient is proposed to explicitly represent the effectiveness of EPP implementation.

Network virtualization and slicing technologies create opportunity for infrastructure-less virtual network operators (VNOs) to enter the market anytime and provide diverse services. Multiple VNOs compete to provide the same kinds of services to end users (EUs). VNOs lease virtual resources from the infrastructure provider (InP) and sell services to the EUs by using the leased resources. The difference between the selling and leasing is the gross profit for the VNOs. A VNO that leases resources without precise knowledge of future demand, may not consume all the leased resources through service offers to EUs. Consequently, the VNO experiences loss and resources remain unused. In order to improve resource utilization and ensure that new entrant VNOs do not face losses, proper estimation of initial resource demand is important. In this paper, we propose a Bayesian game with Cournot oligopoly model to properly estimate the optimal initial resource demands for multiple entrant competing VNOs (players) with the objective of maximizing the expected profit for each VNO. The VNOs offer the same kinds of services to EUs with different qualities (player's type), which are public information. The exact service quality with which a VNO competes in the market is private information. Therefore, a VNO assumes the type of its opponent VNOs with certain probability. We derive the Bayesian Nash equilibrium (BNE) of the presented game and evaluate numerically the effect of service qualities and prices on the expected profit and market share of the VNOs.

Rough set theory is an important branch of data mining and granular computing, among which neighborhood rough set is presented to deal with numerical data and hybrid data. In this paper, we propose a new concept called inconsistent neighborhood, which extracts inconsistent objects from a traditional neighborhood. Firstly, a series of interesting properties are obtained for inconsistent neighborhoods. Specially, some properties generate new solutions to compute the quantities in neighborhood rough set. Then, a fast forward attribute reduction algorithm is proposed by applying the obtained properties. Experiments undertaken on twelve UCI datasets show that the proposed algorithm can get the same attribute reduction results as the existing algorithms in neighborhood rough set domain, and it runs much faster than the existing ones. This validates that employing inconsistent neighborhoods is advantageous in the applications of neighborhood rough set. The study would provide a new insight into neighborhood rough set theory.

High dynamic range (HDR) images contain more details of the scene as compared to commonly used low dynamic range (LDR) images. The additional information in the HDR images is important for applications such as high-quality graphics rendering, sensing, scene analysis, and surveillance etc. Moreover, HDR images would provide better visualization experience on HDR displays, which might become more common in near future. Therefore, it is important to encode the entire dynamic range of the HDR images. In this paper, a new lossless, four-channel, eight bits per channel, format for encoding floating-point HDR images is proposed. The format is similar to the well-known RGBE format but constructs the E channel differently for better accuracy. Experimental results show that our technique could reduce the rounding error of the RGBE by more than 88%. In addition, there was a reduction of 44.3% in average error for all 33 images in the database used for this study.

In this paper, a multichannel speech enhancement system which adopts a denoising auto-encoder as part of the beamformer is proposed. The proposed structure of the generalized sidelobe canceller generates enhanced multi-channel signals, instead of merely one channel, to which the following denoising auto-encoder can be applied. Because the beamformer exploits spatial information and compensates for differences in the transfer functions of each channel, the proposed system is expected to resolve the difficulty of modelling relative transfer functions consisting of complex numbers which are hard to model with a denoising auto-encoder. As a result, the modelling capability of the denoising auto-encoder can concentrate on removing the artefacts caused by the beamformer. Unlike conventional beamformers, which combine these artefacts into one channel, they remain separated for each channel in the proposed method. As a result, the denoising auto-encoder can remove the artefacts by referring to other channels. Experimental results prove that the proposed structure is effective for the six-channel data in CHiME, as indicated by improvements in terms of speech enhancement and word error rate in automatic speech recognition.

The development of multi-view video has paved the way for emerging 3D applications. In general multi-view video streaming, video frames for all viewpoints, i.e., cameras, must be transmitted to viewers because the view-switching demands of all viewers are unpredictable. However, existing transmission schemes are highly vulnerable to frame loss. Specifically, frame loss in one viewpoint can induce a collapse in decoding for other viewpoints. To improve loss-resilience, this paper proposes a multi-path based multi-view video transmission scheme. Our scheme encodes video frames into multiple versions that are independent of each other, using inter-view prediction. The scheme then transmits each version using multiple transmission paths. Our scheme makes three contributions: 1) it reduces video traffic even for a large number of cameras, 2) it prevents an increase in the number of undecoded video frames caused by single-frame loss, and 3) it conceals frame loss by taking video frames from other paths. Evaluations show that our proposed scheme improves video quality by 3 dB, as compared to existing transmission schemes in loss-prone environments.

Index generation functions model content-addressable memory, and are useful in virus detectors and routers. Linear decompositions yield simpler circuits that realize index generation functions. This paper proposes a balanced decision tree based heuristic to efficiently design linear decompositions for index generation functions. The proposed heuristic finds a good linear decomposition of an index generation function by using appropriate cost functions and a constraint to construct a balanced tree. Since the proposed heuristic is fast and requires a small amount of memory, it is applicable even to large index generation functions that cannot be solved in a reasonable time by existing heuristics. This paper shows time and space complexities of the proposed heuristic, and experimental results using some large examples to show its efficiency.

This paper investigates the effect of noises added to hidden units of AutoEncoders linked to multilayer perceptrons. It is shown that internal representation of learned features emerges and sparsity of hidden units increases when independent Gaussian noises are added to inputs of hidden units during the deep network training. It is also shown that the weights that connect the contaminated hidden units with the next layer have smaller values and outputs of hidden units tend to be more definite (0 or 1). This is expected to improve the generalization ability of the network through this automatic structuration by adding the noises. This network structuration was confirmed by experiments for MNIST digits classification via a deep neural network model.

In this letter, we propose two robust and distributed game-based algorithms, which are the modifications of two algorithms proposed in [1], to solve the joint base station selection and resource allocation problem with imperfect information in heterogeneous cellular networks (HCNs). In particular, we repeatedly sample the received payoffs in the exploitation stage of each algorithm to guarantee the convergence when the payoffs of some users (UEs) in [1] cannot accurately be acquired for some reasons. Then, we derive the rational sampling number and prove the convergence of the modified algorithms. Finally, simulation results demonstrate that two modified algorithms achieve good convergence performances and robustness in the incomplete information scheme.

For a multiuser multi-input multi-output (MU-MIMO) multicell network, the Han-Kobayashi strategy aims to improve the achievable rate region by splitting the data information intended to a serviced user (UE) into a common message and a private message. The common message is decodable by this UE and another UE from an adjacent cell so that the corresponding intercell interference is cancelled off. This work aims to design optimal precoders for both common and private messages to maximize the network sum-rate, which is a highly nonlinear and nonsmooth function in the precoder matrix variables. Existing approaches are unable to address this difficult problem. In this paper, we develop a successive convex quadratic programming algorithm that generates a sequence of improved points. We prove that the proposed algorithm converges to at least a local optimum of the considered problem. Numerical results confirm the advantages of our proposed algorithm over conventional coordinated precoding approaches where the intercell interference is treated as noise.

In order to prevent the synonym substitution breaking the balance among frequencies of synonyms and improve the statistical undetectability, this paper proposed a novel linguistic steganography based on synonym run-length encoding. Firstly, taking the relative word frequency into account, the synonyms appeared in the text are digitized into binary values and expressed in the form of runs. Then, message are embedded into the parities of runs' lengths by self-adaptively making a positive or negative synonym transformation on boundary elements of two adjacent runs, while preserving the number of relative high and low frequency synonyms to reduce the embedding distortion. Experimental results have shown that the proposed synonym run-length encoding based linguistic steganographic algorithm makes fewer changes on the statistical characteristics of cover texts than other algorithms, and enhances the capability of anti-steganalysis.