Gang JIN Jingsheng ZHAI Jianguo WEI
In this paper, we propose an end-to-end two-branch feature attention network. The network is mainly used for single image dehazing. The network consists of two branches, we call it CAA-Net: 1) A U-NET network composed of different-level feature fusion based on attention (FEPA) structure and residual dense block (RDB). In order to make full use of all the hierarchical features of the image, we use RDB. RDB contains dense connected layers and local feature fusion with local residual learning. We also propose a structure which called FEPA.FEPA structure could retain the information of shallow layer and transfer it to the deep layer. FEPA is composed of serveral feature attention modules (FPA). FPA combines local residual learning with channel attention mechanism and pixel attention mechanism, and could extract features from different channels and image pixels. 2) A network composed of several different levels of FEPA structures. The network could make feature weights learn from FPA adaptively, and give more weight to important features. The final output result of CAA-Net is the combination of all branch prediction results. Experimental results show that the CAA-Net proposed by us surpasses the most advanced algorithms before for single image dehazing.
Han WANG Ruiliu FU Xuejun ZHANG Jun ZHOU Qingwei ZHAO
Lifelong language learning (LLL) aims at learning new tasks and retaining old tasks in the field of NLP. LAMOL is a recent LLL framework following data-free constraints. Previous works have been researched based on LAMOL with additional computing with more time costs or new parameters. However, they still have a gap between multi-task learning (MTL), which is regarded as the upper bound of LLL. In this paper, we propose Metacognitive Adaptation (Metac-Adapt) almost without adding additional time cost and computational resources to make the model generate better pseudo samples and then replay them. Experimental results demonstrate that Metac-Adapt is on par with MTL or better.
Ruxue GUO Pengxu JIANG Ruiyu LIANG Yue XIE Cairong ZOU
For a long time, the compensation effect of hearing aid is mainly evaluated subjectively, and there are fewer studies of objective evaluation. Furthermore, a pure speech signal is generally required as a reference in the existing objective evaluation methods, which restricts the practicality in a real-world environment. Therefore, this paper presents a non-intrusive speech quality evaluation method for hearing aid, which combines the audiogram and weighted frequency information. The proposed model mainly includes an audiogram information extraction network, a frequency information extraction network, and a quality score mapping network. The audiogram is the input of the audiogram information extraction network, which helps the system capture the information related to hearing loss. In addition, the low-frequency bands of speech contain loudness information and the medium and high-frequency components contribute to semantic comprehension. The information of two frequency bands is input to the frequency information extraction network to obtain time-frequency information. When obtaining the high-level features of different frequency bands and audiograms, they are fused into two groups of tensors that distinguish the information of different frequency bands and used as the input of the attention layer to calculate the corresponding weight distribution. Finally, a dense layer is employed to predict the score of speech quality. The experimental results show that it is reasonable to combine the audiogram and the weight of the information from two frequency bands, which can effectively realize the evaluation of the speech quality of the hearing aid.
Hitoshi KIYA Ryota IIJIMA Aprilpyone MAUNGMAUNG Yuma KINOSHITA
In this paper, we propose a combined use of transformed images and vision transformer (ViT) models transformed with a secret key. We show for the first time that models trained with plain images can be directly transformed to models trained with encrypted images on the basis of the ViT architecture, and the performance of the transformed models is the same as models trained with plain images when using test images encrypted with the key. In addition, the proposed scheme does not require any specially prepared data for training models or network modification, so it also allows us to easily update the secret key. In an experiment, the effectiveness of the proposed scheme is evaluated in terms of performance degradation and model protection performance in an image classification task on the CIFAR-10 dataset.
Tatsuya KOYAKUMARU Masahiro YUKAWA Eduardo PAVEZ Antonio ORTEGA
This paper presents a convex-analytic framework to learn sparse graphs from data. While our problem formulation is inspired by an extension of the graphical lasso using the so-called combinatorial graph Laplacian framework, a key difference is the use of a nonconvex alternative to the l1 norm to attain graphs with better interpretability. Specifically, we use the weakly-convex minimax concave penalty (the difference between the l1 norm and the Huber function) which is known to yield sparse solutions with lower estimation bias than l1 for regression problems. In our framework, the graph Laplacian is replaced in the optimization by a linear transform of the vector corresponding to its upper triangular part. Via a reformulation relying on Moreau's decomposition, we show that overall convexity is guaranteed by introducing a quadratic function to our cost function. The problem can be solved efficiently by the primal-dual splitting method, of which the admissible conditions for provable convergence are presented. Numerical examples show that the proposed method significantly outperforms the existing graph learning methods with reasonable computation time.
Construction of resilient Boolean functions in odd variables having strictly almost optimal (SAO) nonlinearity appears to be a rather difficult task in stream cipher and coding theory. In this paper, based on the modified High-Meets-Low technique, a general construction to obtain odd-variable SAO resilient Boolean functions without directly using PW functions or KY functions is presented. It is shown that the new class of functions possess higher resiliency order than the known functions while keeping higher SAO nonlinearity, and in addition the resiliency order increases rapidly with the variable number n.
Antoine BOSSARD Keiichi KANEKO Frederick C. HARRIS, JR.
Reducing the number of link crossings in a network drawn on the plane such as a wiring board is a well-known problem, and especially the calculation of the minimum number of such crossings: this is the crossing number problem. It has been shown that finding a general solution to the crossing number problem is NP-hard. So, this problem is addressed for particular classes of graphs and this is also our approach in this paper. More precisely, we focus hereinafter on the torus topology. First, we discuss an upper bound on cr(T(2, k)) the number of crossings in a 2-dimensional k-ary torus T(2, k) where k ≥ 2: the result cr(T(2, k)) ≤ k(k - 2) and the given constructive proof lay foundations for the rest of the paper. Second, we extend this discussion to derive an upper bound on the crossing number of a 3-dimensional k-ary torus: cr(T(3, k)) ≤ 2k4 - k3 - 4k2 - 2⌈k/2⌉⌊k/2⌋(k - (k mod 2)) is obtained. Third, an upper bound on the crossing number of an n-dimensional k-ary torus is derived from the previously established results, with the order of this upper bound additionally established for more clarity: cr(T(n, k)) is O(n2k2n-2) when n ≥ k and O(nk2n-1) otherwise.
In this letter, we consider a global stabilization problem for a class of feedforward systems by an event-triggered control. This is an extended work of [10] in a way that there are uncertain feedforward nonlinearity and time-varying input delay in the system. First, we show that the considered system is globally asymptotically stabilized by a proposed event-triggered controller with a gain-scaling factor. Then, we also show that the interexecution times can be enlarged by adjusting a gain-scaling factor. A simulation example is given for illustration.
Feng LIU Qianqian WU Conggai LI Fangjiong CHEN Yanli XU
To improve the performance of underwater acoustic communications, this letter proposes a polar coding scheme with adaptive channel equalization, which can reduce the amount of feedback information. Furthermore, a hybrid automatic repeat request (HARQ) mechanism is provided to mitigate the impact of estimation errors. Simulation results show that the proposed scheme outperforms the turbo equalization in bit error rate. Computational complexity analysis is also provided for comparison.
Daiki TODA Ren ANZAI Koichi ICHIGE Ryo SAITO Daichi UEKI
A method of radar-based contactless vital-sign sensing and electrocardiogram (ECG) signal reconstruction using deep learning is proposed. A radar system is an effective tool for contactless vital-sign sensing because it can measure a small displacement of the body surface without contact. However, most of the conventional methods have limited evaluation indices and measurement conditions. A method of measuring body-surface-displacement signals by using frequency-modulated continuous-wave (FMCW) radar and reconstructing ECG signals using a convolutional neural network (CNN) is proposed. This study conducted two experiments. First, we trained a model using the data obtained from six subjects breathing in a seated condition. Second, we added sine wave noise to the data and trained the model again. The proposed model is evaluated with a correlation coefficient between the reconstructed and actual ECG signal. The results of first experiment show that their ECG signals are successfully reconstructed by using the proposed method. That of second experiment show that the proposed method can reconstruct signal waveforms even in an environment with low signal-to-noise ratio (SNR).
Xiang BI Huang HUANG Benhong ZHANG Xing WEI
It is of great significance to design a stable and reliable routing protocol for Vehicular Ad Hoc Networks (VANETs) that adopt Vehicle to Vehicle (V2V) communications in the face of frequent network topology changes. In this paper, we propose a hybrid routing algorithm, RCRIQ, based on improved Q-learning. For an established cluster structure, the cluster head is used to select the gateway vehicle according to the gateway utility function to expand the communication range of the cluster further. During the link construction stage, an improved Q-learning algorithm is adopted. The corresponding neighbor vehicle is chosen according to the maximum Q value in the neighbor list. The heuristic algorithm selects the next-hop by the maximum heuristic function value when selecting the next-hop neighbor node. The above two strategies are comprehensively evaluated to determine the next hop. This way ensures the optimal selection of the next hop in terms of reachability and other communication parameters. Simulation experiments show that the algorithm proposed in this article has better performance in terms of routing stability, throughput, and communication delay in the urban traffic scene.
Souhei YANASE Fujun HE Haruto TAKA Akio KAWABATA Eiji OKI
This paper proposes a migration model for distributed server allocation. In distributed server allocation, each user is assigned to a server to minimize the communication delay. In the conventional model, a user cannot migrate to another server to avoid instability. We develop a model where each user can migrate to another server while receiving services. We formulate the proposed model as an integer linear programming problem. We prove that the considered problem is NP-complete. We introduce a heuristic algorithm. Numerical result shows that the proposed model reduces the average communication delay by 59% compared to the conventional model at most.
Xiangyu MENG Kangfeng WEI Zhiyi YU Xinlun CAI
This paper proposes a low-power 100Gb/s four-level pulse amplitude modulation driver (PAM-4 Driver) based on linear distortion compensation structure for thin-film Lithium Niobate (LiNbO3) modulators, which manages to achieve high linearity in the output. The inductive peaking technology and open drain structure enable the overall circuit to achieve a 31-GHz bandwidth. With an area of 0.292 mm2, the proposed PAM-4 driver chip is designed in a 65-nm process to achieve power consumption of 37.7 mW. Post-layout simulation results show that the power efficiency is 0.37 mW/Gb/s, RLM is more than 96%, and the FOM value is 8.84.
Daisuke KOBAYASHI Ken NAKAMURA Masaki KITAHARA Tatsuya OSAWA Yuya OMORI Takayuki ONISHI Hiroe IWASAKI
This paper describes a novel low-latency 4K 60 fps HEVC (high efficiency video coding)/H.265 multi-channel encoding system with content-aware bitrate control for live streaming. Adaptive bitrate (ABR) streaming techniques, such as MPEG-DASH (dynamic adaptive streaming over HTTP) and HLS (HTTP live streaming), spread widely on Internet video streaming. Live content has increased with the expansion of streaming services, which has led to demands for traffic reduction and low latency. To reduce network traffic, we propose content-aware dynamic and seamless bitrate control that supports multi-channel real-time encoding for ABR, including 4K 60 fps video. Our method further supports chunked packaging transfer to provide low-latency streaming. We adopt a hybrid architecture consisting of hardware and software processing. The system consists of multiple 4K HEVC encoder LSIs that each LSI can encode 4K 60 fps or up to high-definition (HD) ×4 videos efficiently with the proposed bitrate control method. The software takes the packaging process according to the various streaming protocol. Experimental results indicate that our method reduces encoding bitrates obtained with constant bitrate encoding by as much as 56.7%, and the streaming latency over MPEG-DASH is 1.77 seconds.
Binggang ZHUO Masaki MURATA Qing MA
Paragraph segmentation is a text segmentation task. Iikura et al. achieved excellent results on paragraph segmentation by introducing focal loss to Bidirectional Encoder Representations from Transformers. In this study, we investigated paragraph segmentation on Daily News and Novel datasets. Based on the approach proposed by Iikura et al., we used auxiliary loss to train the model to improve paragraph segmentation performance. Consequently, the average F1-score obtained by the approach of Iikura et al. was 0.6704 on the Daily News dataset, whereas that of our approach was 0.6801. Our approach thus improved the performance by approximately 1%. The performance improvement was also confirmed on the Novel dataset. Furthermore, the results of two-tailed paired t-tests indicated that there was a statistical significance between the performance of the two approaches.
Constructing accurate similarity graph is an important process in graph-based clustering. However, traditional methods have three drawbacks, such as the inaccuracy of the similarity graph, the vulnerability to noise and outliers, and the need for additional discretization process. In order to eliminate these limitations, an entropy regularized unsupervised clustering based on maximum correntropy criterion and adaptive neighbors (ERMCC) is proposed. 1) Combining information entropy and adaptive neighbors to solve the trivial similarity distributions. And we introduce l0-norm and spectral embedding to construct similarity graph with sparsity and strong segmentation ability. 2) Reducing the negative impact of non-Gaussian noise by reconstructing the error using correntropy. 3) The prediction label vector is directly obtained by calculating the sparse strongly connected components of the similarity graph Z, which avoids additional discretization process. Experiments are conducted on six typical datasets and the results showed the effectiveness of the method.
Mitsuki ITO Fujun HE Kento YOKOUCHI Eiji OKI
This paper proposes a robust optimization model for probabilistic protection under uncertain capacity demands to minimize the total required capacity against multiple simultaneous failures of physical machines. The proposed model determines both primary and backup virtual machine allocations simultaneously under the probabilistic protection guarantee. To express the uncertainty of capacity demands, we introduce an uncertainty set that considers the upper bound of the total demand and the upper and lower bounds of each demand. The robust optimization technique is applied to the optimization model to deal with two uncertainties: failure event and capacity demand. With this technique, the model is formulated as a mixed integer linear programming (MILP) problem. To solve larger sized problems, a simulated annealing (SA) heuristic is introduced. In SA, we obtain the capacity demands by solving maximum flow problems. Numerical results show that our proposed model reduces the total required capacity compared with the conventional model by determining both primary and backup virtual machine allocations simultaneously. We also compare the results of MILP, SA, and a baseline greedy algorithm. For a larger sized problem, we obtain approximate solutions in a practical time by using SA and the greedy algorithm.
Xiangyu MENG Yecong LI Zhiyi YU
This paper proposes a design of high-speed interconnection between optical modules and electrical modules via bonding-wires and coplanar waveguide transmission lines on printed circuit boards for 400 Gbps 4-channel optical communication systems. In order to broaden the interconnection bandwidth, interdigitated capacitors were integrated with GSG pads on chip for the first time. Simulation results indicate the reflection coefficient is below -10 dB from DC to 53 GHz and the insertion loss is below 1 dB from DC to 45 GHz. Both indicators show that the proposed interconnection structure can effectively satisfy the communication bandwidth requirements of 100-Gbps or even higher data-rate PAM4 signals.
This letter proposes a novel intelligent dynamic channel assignment (DCA) scheme with small-cells to improve the system performance for uplink machine-type communications (MTC) based on OFDMA-FDD. Outdoor MTC devices (OMDs) have serious interference from indoor MTC devices (IMDs) served by small-cell access points (SAPs) with frequency reuse. Thus, in the proposed DCA scheme, the macro base station (MBS) first measures the received signal strength from both OMDs and IMDs after setting the transmission power. Then, the MBS dynamically assigns subchannels to each SAP with consideration of strong interference from IMDs to the MBS. Through simulation results, it is shown that the proposed DCA scheme outperforms other schemes in terms of the capacity of OMDs and IMDs.
Jisoo KIM Seonjoo CHOI Jaesung LIM
In time difference of arrival-based signal source location estimation, geometrical errors are caused by the location of multiple unmanned aerial vehicles (UAV). Herein, we propose a divide-and-conquer algorithm to determine the optimal location for each UAV. Simulations results confirm that multiple UAVs shifted to an optimal position and the location accuracy improved.