We propose an optimal power allocation scheme that maximizes the transmission rate of device-to-device (D2D) communications underlaying a cellular system based on orthogonal frequency division multiplexing (OFDM). The proposed algorithm first calculates the maximum allowed transmission power of a D2D transmitter to restrict the interference caused to a cellular link that share the same OFDM subchannels with the D2D link. Then, with a constraint on the maximum transmit power, an optimization of water-filling type is performed to find the optimal transmit power allocation across subchannels and within each subchannel. The performance of the proposed power allocation scheme is evaluated in terms of the average achievable rate of the D2D link.

Recently, the ECG-based diagnosis system based on wearable devices has attracted more and more attention of researchers. Existing studies have achieved high classification accuracy by using deep neural networks (DNNs), but there are still some problems, such as: imprecise heart beat segmentation, inadequate use of medical knowledge, the same treatment of features with different importance. To address these problems, this paper: 1) proposes an adaptive segmenting-reshaping method to acquire abundant useful samples; 2) builds a set of hand-crafted features and deep features on the inner-beat, beat and inter-beat scale by integrating enough medical knowledge. 3) introduced a modified channel attention module (CAM) to augment the significant channels in deep features. Following the Association for Advancement of Medical Instrumentation (AAMI) recommendation, we classified the dataset into four classes and validated our algorithm on the MIT-BIH database. Experiments show that the accuracy of our model reaches 96.94%, a 3.71% increase over that of a state-of-the-art alternative.

Training deep learning (DL) is a computationally intensive process; as a result, training time can become so long that it impedes the development of DL. High performance computing clusters, especially supercomputers, are equipped with a large amount of computing resources, storage resources, and efficient interconnection ability, which can train DL networks better and faster. In this paper, we propose a method to train DL networks distributed with high efficiency. First, we propose a hierarchical synchronous Stochastic Gradient Descent (SGD) strategy, which can make full use of hardware resources and greatly increase computational efficiency. Second, we present a two-level parameter synchronization scheme which can reduce communication overhead by transmitting parameters of the first layer models in shared memory. Third, we optimize the parallel I/O by making each reader read data as continuously as possible to avoid the high overhead of discontinuous data reading. At last, we integrate the LARS algorithm into our system. The experimental results demonstrate that our approach has tremendous performance advantages relative to unoptimized methods. Compared with the native distributed strategy, our hierarchical synchronous SGD strategy (HSGD) can increase computing efficiency by about 20 times.

Recurrent neural networks (RNNs) have been proven effective for sequence-based tasks thanks to their capability to process temporal information. In real-world systems, deep RNNs are more widely used to solve complicated tasks such as large-scale speech recognition and machine translation. However, the implementation of deep RNNs on traditional hardware platforms is inefficient due to long-range temporal dependence and irregular computation patterns within RNNs. This inefficiency manifests itself in the proportional increase in the latency of RNN inference with respect to the number of layers of deep RNNs on CPUs and GPUs. Previous work has focused mostly on optimizing and accelerating individual RNN cells. To make deep RNN inference fast and efficient, we propose an accelerator based on a multi-FPGA platform called Flow-in-Cloud (FiC). In this work, we show that the parallelism provided by the multi-FPGA system can be taken advantage of to scale up the inference of deep RNNs, by partitioning a large model onto several FPGAs, so that the latency stays close to constant with respect to increasing number of RNN layers. For single-layer and four-layer RNNs, our implementation achieves 31x and 61x speedup compared with an Intel CPU.

Quantization is an important technique for implementing convolutional neural networks on edge devices. Quantization often requires relearning, but relearning sometimes cannot be always be applied because of issues such as cost or privacy. In such cases, it is important to know the numerical precision required to maintain accuracy. We accurately simulate calculations on hardware and accurately measure the relationship between accuracy and numerical precision.

Nowadays recommender systems (RS) keep drawing attention from academia, and collaborative filtering (CF) is the most successful technique for building RS. To overcome the inherent limitation, which is referred to as data sparsity in CF, various solutions are proposed to incorporate additional social information into recommendation processes, such as trust networks. However, existing methods suffer from multi-source data integration (i.e., fusion of social information and ratings), which is the basis for similarity calculation of user preferences. To this end, we propose a social collaborative filtering method based on novel trust metrics. Firstly, we use Graph Convolutional Networks (GCNs) to learn the associations between social information and user ratings while considering the underlying social network structures. Secondly, we measure the direct-trust values between neighbors by representing multi-source data as user ratings on popular items, and then calculate the indirect-trust values based on trust propagations. Thirdly, we employ all trust values to create a social regularization in user-item rating matrix factorization in order to avoid overfittings. The experiments on real datasets show that our approach outperforms the other state-of-the-art methods on usage of multi-source data to alleviate data sparsity.

In recent years, the use of big data has attracted more attention, and many techniques for data analysis have been proposed. Big data analysis is difficult, however, because such data varies greatly in its regularity. Heterogeneous mixture machine learning is one algorithm for analyzing such data efficiently. In this study, we propose online heterogeneous learning based on an online EM algorithm. Experiments show that this algorithm has higher learning accuracy than that of a conventional method and is practical. The online learning approach will make this algorithm useful in the field of data analysis.

Deep neural network (DNN)-based speech synthesis became popular in recent years and is expected to soon be widely used in embedded devices and environments with limited computing resources. The key intention of these systems in poor computing environments is to reduce the computational cost of generating speech parameter sequences while maintaining voice quality. However, reducing computational costs is challenging for two primary conventional DNN-based methods used for modeling speech parameter sequences. In feed-forward neural networks (FFNNs) with maximum likelihood parameter generation (MLPG), the MLPG reconstructs the temporal structure of the speech parameter sequences ignored by FFNNs but requires additional computational cost according to the sequence length. In recurrent neural networks, the recursive structure allows for the generation of speech parameter sequences while considering temporal structures without the MLPG, but increases the computational cost compared to FFNNs. We propose a new approach for DNNs to acquire parameters captured from the temporal structure by backpropagating the errors of multiple attributes of the temporal sequence via the loss function. This method enables FFNNs to generate speech parameter sequences by considering their temporal structure without the MLPG. We generated the fundamental frequency sequence and the mel-cepstrum sequence with our proposed method and conventional methods, and then synthesized and subjectively evaluated the speeches from these sequences. The proposed method enables even FFNNs that work on a frame-by-frame basis to generate speech parameter sequences by considering the temporal structure and to generate sequences perceptually superior to those from the conventional methods.

We propose a speech analysis-synthesis and deep neural network (DNN)-based text-to-speech (TTS) synthesis framework using Gaussian mixture model (GMM)-based approximation of full-band spectral envelopes. GMMs have excellent properties as acoustic features in statistic parametric speech synthesis. Each Gaussian function of a GMM fits the local resonance of the spectrum. The GMM retains the fine spectral envelope and achieve high controllability of the structure. However, since conventional speech analysis methods (i.e., GMM parameter estimation) have been formulated for a narrow-band speech, they degrade the quality of synthetic speech. Moreover, a DNN-based TTS synthesis method using GMM-based approximation has not been formulated in spite of its excellent expressive ability. Therefore, we employ peak-picking-based initialization for full-band speech analysis to provide better initialization for iterative estimation of the GMM parameters. We introduce not only prediction error of GMM parameters but also reconstruction error of the spectral envelopes as objective criteria for training DNN. Furthermore, we propose a method for multi-task learning based on minimizing these errors simultaneously. We also propose a post-filter based on variance scaling of the GMM for our framework to enhance synthetic speech. Experimental results from evaluating our framework indicated that 1) the initialization method of our framework outperformed the conventional one in the quality of analysis-synthesized speech; 2) introducing the reconstruction error in DNN training significantly improved the synthetic speech; 3) our variance-scaling-based post-filter further improved the synthetic speech.

In this paper, we propose a novel single-template strategy based on a mean template set and locally/globally weighted dynamic time warping (LG-DTW) to improve the performance of online signature verification. Specifically, in the enrollment phase, we implement a time series averaging method, Euclidean barycenter-based DTW barycenter averaging, to obtain a mean template set considering intra-user variability among reference samples. Then, we acquire a local weighting estimate considering a local stability sequence that is obtained analyzing multiple matching points of an optimal match between the mean template and reference sets. Thereafter, we derive a global weighting estimate based on the variable importance estimated by gradient boosting. Finally, in the verification phase, we apply both local and global weighting methods to acquire a discriminative LG-DTW distance between the mean template set and a query sample. Experimental results obtained on the public SVC2004 Task2 and MCYT-100 signature datasets confirm the effectiveness of the proposed method for online signature verification.

How to restore virtual network against substrate network failure (e.g. link cut) is one of the key challenges of network virtualization. The traditional virtual network recovery (VNR) methods are mostly based on the idea of centralized control. However, if multiple virtual networks fail at the same time, their recovery processes are usually queued according to a specific priority, which may increase the average waiting time of users. In this letter, we study distributed virtual network recovery (DVNR) method to improve the virtual network recovery efficiency. We establish exclusive virtual machine (VM) for each virtual network and process recovery requests of multiple virtual networks in parallel. Simulation results show that the proposed DVNR method can obtain recovery success rate closely to centralized VNR method while yield ~70% less average recovery time.

In this paper, we address the problem of analyzing topics, included in a social video group, to improve the retrieval performance of videos. Unlike previous methods that focused on an individual visual aspect of videos, the proposed method aims to leverage the “mutual reinforcement” of heterogeneous modalities such as tags and users associated with video on the Internet. To represent multiple types of relationships between each heterogeneous modality, the proposed method constructs three subgraphs: user-tag, video-video, and video-tag graphs. We combine the three types of graphs to obtain a heterogeneous graph. Then the extraction of latent features, i.e., topics, becomes feasible by applying graph-based soft clustering to the heterogeneous graph. By estimating the membership of each grouped cluster for each video, the proposed method defines a new video similarity measure. Since the understanding of video content is enhanced by exploiting latent features obtained from different types of data that complement each other, the performance of visual reranking is improved by the proposed method. Results of experiments on a video dataset that consists of YouTube-8M videos show the effectiveness of the proposed method, which achieves a 24.3% improvement in terms of the mean normalized discounted cumulative gain in a search ranking task compared with the baseline method.

Tweakable block cipher (TBC) is an extension of conventional block cipher. We study how to build a TBC based on generalized Feistel structure (GFS), a classical block cipher construction. While known dedicated TBC proposals are based on substitution-permutation network (SPN), GFS has not been used for building TBC. In particular, we take 64-bit GFS block cipher TWINE and try to make it tweakable with a minimum change. To find a best one from a large number of candidates, we performed a comprehensive search with a help of mixed integer linear programming (MILP) solver. As a result, our proposal TWINE is quite efficient, has the same number of rounds as TWINE with extremely simple tweak schedule.

Autonomous mobility machines, such as self-driving cars, transportation robots, and automated construction machines, are promising to support or enrich human lives. To further improve such machines, they will be connected to the network via wireless links to be managed, monitored, or remotely operated. The autonomous mobility machines must have self-status based on their positioning system to safely conduct their operations without colliding with other objects. The self-status is not only essential for machine operation but also it is valuable for wireless link quality management. This paper presents self-status-based wireless link quality prediction and evaluates its performance by using a prototype mobility robot combined with a wireless LAN system. The developed robot has functions to measure the throughput and receive signal strength indication and obtain self-status details such as location, direction, and odometry data. Prediction performance is evaluated in offline processing by using the dataset gathered in an indoor experiment. The experiments clarified that, in the 5.6 GHz band, link quality prediction using self-status of the robot forecasted the throughput several seconds into the future, and the prediction accuracies were investigated as dependent on time window size of the target throughput, bandwidth, and frequency gap.

In this paper, we consider the collaborative editing of two-dimensional (2D) data such as handwritten letters and illustrations. In contrast to the editing of 1D data, which is generally realized by the combination of insertion/deletion of characters, overriding of strokes can have a specific meaning in editing 2D data. In other words, the appearance of the resulting picture depends on the reflection order of strokes to the shared canvas in addition of the absolute coordinate of the strokes. We propose a Peer-to-Peer (P2P) collaborative drawing system consisting of several nodes with replica canvas, in which the consistency among replica canvases is maintained through data channel of WebRTC. The system supports three editing modes concerned with the reflection order of strokes generated by different users. The result of experiments indicates that the proposed system realizes a short latency of around 120 ms, which is a half of a cloud-based system implemented with Firebase Realtime Database. In addition, it realizes a smooth drawing of pictures on remote canvases with a refresh rate of 12 fps.

An inpainting method via sparse representation based on a new phaseless quality metric is presented in this paper. Since power spectra, phaseless features, of local regions within images enable more successful representation of their texture characteristics compared to their pixel values, a new quality metric based on these phaseless features is newly derived for image representation. Specifically, the proposed method enables spare representation of target signals, i.e., target patches, including missing intensities by monitoring errors converged by phase retrieval as the novel phaseless quality metric. This is the main contribution of our study. In this approach, the phase retrieval algorithm used in our method has the following two important roles: (1) derivation of the new quality metric that can be derived even for images including missing intensities and (2) conversion of phaseless features, i.e., power spectra, to pixel values, i.e., intensities. Therefore, the above novel approach solves the existing problem of not being able to use better features or better quality metrics for inpainting. Results of experiments showed that the proposed method using sparse representation based on the new phaseless quality metric outperforms previously reported methods that directly use pixel values for inpainting.

In this letter, a quantitative evaluation index of contrast improvement of color images for dichromats is proposed. The index is made by adding two parameters to an existing index to make evaluation results consistent with human evaluation results. The effectiveness and validity of the proposed index are verified by experiments.

In emergency communication systems research, aerial wireless relay networks (AWRNs) using multicopter unmanned aerial vehicles (UAVs) have been proposed. The main issue of the AWRNs is how to minimize the delay time of packet transmissions since it is not easy to supply many multicopters to cover a wide area. In this paper, we review the flight schemes and their delay time for the AWRNs. Furthermore, the network has specific issues such as multicopters' drops due to their battery capacity depletion and inclination of moving multicopters. The inclination of multicopters affects the received power, and the communication range changes based on the inclination as well. Therefore, we clarify the effect of these issues on the delay time.

The volume of digital information is growing at an extremely fast pace which, in turn, exacerbates the need of efficient mechanisms to find the presence of a pattern in an input text or a set of input strings. Combining the processing power of Graphics Processing Unit (GPU) with matching algorithms seems a natural alternative to speedup the string-matching process. This work proposes a Parallel Rabin-Karp implementation (PRK) that encompasses a fast-parallel prefix-sums algorithm to maximize parallelization and accelerate the matching verification. Given an input text T of length n and p patterns of length m, the proposed implementation finds all occurrences of p in T in O(m+q+n/τ+nm/q) time, where q is a sufficiently large prime number and τ is the available number of threads. Sequential and parallel versions of the PRK have been implemented. Experiments have been executed on p≥1 patterns of length m comprising of m=10, 20, 30 characters which are compared against a text string of length n=227. The results show that the parallel implementation of the PRK algorithm on NVIDIA V100 GPU provides speedup surpassing 372 times when compared to the sequential implementation and speedup of 12.59 times against an OpenMP implementation running on a multi-core server with 128 threads. Compared to another prominent GPU implementation, the PRK implementation attained speedup surpassing 37 times.

Change-point detection is the problem of finding points of time when a probability distribution of samples changed. There are various related problems, such as estimating the number of the change-points and estimating magnitude of the change. Though various statistical models have been assumed in the field of change-point detection, we particularly deal with i.p.i.d. (independent-piecewise-identically-distributed) sources. In this paper, we formulate the related problems in a general manner based on statistical decision theory. Then we derive optimal estimators for the problems under the Bayes risk principle. We also propose efficient algorithms for the change-point detection-related problems in the i.p.i.d. sources, while in general, the optimal estimations requires huge amount of calculation in Bayesian setting. Comparison of the proposed algorithm and previous methods are made through numerical examples.