Peer assessment in education has pedagogical benefits and is a promising method for grading a large number of submissions. At the same time, student reliability has been regarded as a problem; consequently, various methods of estimating highly reliable grades from scores given by multiple students have been proposed. Under most of the existing methods, a nonadaptive allocation pattern, which performs allocation in advance, is assumed. In this study, we analyze the effect of student-submission allocation on score estimation in peer assessment under a nonadaptive allocation setting. We examine three types of nonadaptive allocation methods, random allocation, circular allocation and group allocation, which are considered the commonly used approaches among the existing nonadaptive peer assessment methods. Through simulation experiments, we show that circular allocation and group allocation tend to yield lower accuracy than random allocation. Then, we utilize this result to improve the existing adaptive allocation method, which performs allocation and assessment in parallel and tends to make similar allocation result to circular allocation. We propose the method to replace part of the allocation with random allocation, and show that the method is effective through experiments.

Previous studies on anomaly detection in videos have trained detectors in which reconstruction and prediction tasks are performed on normal data so that frames on which their task performance is low will be detected as anomalies during testing. This paper proposes a new approach that involves sorting video clips, by using a generative network structure. Our approach learns spatial contexts from appearances and temporal contexts from the order relationship of the frames. Experiments were conducted on four datasets, and we categorized the anomalous sequences by appearance and motion. Evaluations were conducted not only on each total dataset but also on each of the categories. Our method improved detection performance on both anomalies with different appearance and different motion from normality. Moreover, combining our approach with a prediction method produced improvements in precision at a high recall.

With the high development of computation requirements in Internet of Things, resource-limited edge servers usually require to cooperate to perform the tasks. Most related studies usually assume a static cooperation approach which might not suit the dynamic environment of edge computing. In this paper, we consider a dynamic cooperation approach by guiding edge servers to form coalitions dynamically. It raises two issues: 1) how to guide them to optimally form coalitions and 2) how to cope with the dynamic feature where server statuses dynamically change as the tasks are performed. The coalitional Markov decision process (CMDP) model proposed in our previous work can handle these issues well. However, its basic solution, coalitional Q-learning, cannot handle the large scale problem when the task number is large in edge computing. Our response is to propose a novel algorithm called deep coalitional Q-learning (DCQL) to solve it. To sum up, we first formulate the dynamic cooperation problem of edge servers as a CMDP: each edge server is regarded as an agent and the dynamic process is modeled as a MDP where the agents observe the current state to formulate several coalitions. Each coalition takes an action to impact the environment which correspondingly transfers to the next state to repeat the above process. Then, we propose DCQL which includes a deep neural network and so can well cope with large scale problem. DCQL can guide the edge servers to form coalitions dynamically with the target of optimizing some goal. Furthermore, we run experiments to verify our proposed algorithm's effectiveness in different settings.

Distributed edge cloud computing is an important computation infrastructure for Internet of Things (IoT) and its task offloading problem has attracted much attention recently. Most existing work on task offloading in distributed edge cloud computing usually assumes that each self-interested user owns one edge server and chooses whether to execute its tasks locally or to offload the tasks to cloud servers. The goal of each edge server is to maximize its own interest like low delay cost, which corresponds to a non-cooperative setting. However, with the strong development of smart IoT communities such as smart hospital and smart factory, all edge and cloud servers can belong to one organization like a technology company. This corresponds to a cooperative setting where the goal of the organization is to maximize the team interest in the overall edge cloud computing system. In this paper, we consider a new problem called cooperative task offloading where all edge servers try to cooperate to make the entire edge cloud computing system achieve good performance such as low delay cost and low energy cost. However, this problem is hard to solve due to two issues: 1) each edge server status dynamically changes and task arrival is uncertain; 2) each edge server can observe only its own status, which makes it hard to optimize team interest as global information is unavailable. For solving these issues, we formulate the problem as a decentralized partially observable Markov decision process (Dec-POMDP) which can well handle the dynamic features under partial observations. Then, we apply a multi-agent reinforcement learning algorithm called value decomposition network (VDN) and propose a VDN-based task offloading algorithm (VDN-TO) to solve the problem. Specifically, the motivation is that we use a team value function to evaluate the team interest, which is then divided into individual value functions for each edge server. Then, each edge server updates its individual value function in the direction that can maximize the team interest. Finally, we choose a part of a real dataset to evaluate our algorithm and the results show the effectiveness of our algorithm in a comparison with some other existing methods.

In this letter, we analyze the error rate performance of M-ary coherent free-space optical (FSO) communications under strong atmospheric turbulence. Specifically, we derive the exact error rates for M-ary phase shift keying (MPSK) and M-ary quadrature amplitude modulation (MQAM) based on moment-generating function (MGF) with negative exponential distributed turbulence, where maximum ratio combining (MRC) receiver is adopted to mitigate the turbulence effects. Additionally, by evaluating the asymptotic error rate in high signal-to-noise ratio (SNR) regime, it is possible to effectively investigate and predict the error rate performance for various system configurations. The accuracy and the effectiveness of our theoretical analyses are verified via numerical results.

In this paper, event-triggered control over a sensor network is studied as one of the control methods of cyber-physical systems. Event-triggered control is a method that communications occur only when the measured value is widely changed. In the proposed method, by solving an LMI (Linear Matrix Inequality) feasibility problem, an event-triggered output feedback controller such that the closed-loop system is asymptotically stable is derived. First, the problem formulation is given. Next, the control problem is reduced to an LMI feasibility problem. Finally, the proposed method is demonstrated by a numerical example.

Energy efficiency is one of the critical issues for Wireless Sensor Networks (WSN). IEEE 802.15.4 beacon-enabled MAC protocol achieves low energy consumption by having periodical inactive portions, where nodes run in low power. However, IEEE 802.15.4 beacon-enabled protocol cannot respond to dynamic changes in the number of sensor nodes and data rates in WSN because its duty cycle is fixed and immutable. In this paper, we propose a dynamic superframe duration adaptation scheme based on the Markov chain-based analysis methods for IEEE 802.15.4 beacon-enabled protocol. The proposed methods are flexible enough to accommodate changes in the number of sensor nodes and differences in data rates in WSNs while maintaining low latency and low energy consumption despite slight degradation in packet delivery ratio.

AlphaSeq is a new paradigm to design sequencess with desired properties based on deep reinforcement learning (DRL). In this work, we propose a new metric function and a new reward function, to design an improved version of AlphaSeq. We show analytically and also through numerical simulations that the proposed algorithm can discover sequence sets with preferable properties faster than that of the previous algorithm.

Real-time and scalable multicast services are of paramount importance to Industrial Internet of Things (IIoT) applications. To realize these services, the multicast algorithm should, on the one hand, ensure the maximum delay of a multicast session not exceeding its upper delay bound. On the other hand, the algorithm should minimize session costs. As an emerging networking paradigm, Software-defined Networking (SDN) can provide a global view of the network to multicast algorithms, thereby bringing new opportunities for realizing the desired multicast services in IIoT environments. Unfortunately, existing SDN-based multicast (SDM) algorithms cannot meet the real-time and scalable requirements simultaneously. Therefore, in this paper, we focus on SDM algorithm design for IIoT environments. To be specific, the paper first converts the multicast tree construction problem for SDM in IIoT environments into a delay-bounded least-cost shared tree problem and proves that it is an NP-complete problem. Then, the paper puts forward a shared tree (ST) algorithm called SDM4IIoT to compute suboptimal solutions to the problem. The algorithm consists of five steps: 1) construct a delay-optimal shared tree; 2) divide the tree into a set of subpaths and a subtree; 3) optimize the cost of each subpath by relaxing the delay constraint; 4) optimize the subtree cost in the same manner; 5) recombine them into a shared tree. Simulation results show that the algorithm can provide real-time support that other ST algorithms cannot. In addition, it can achieve good scalability. Its cost is only 20.56% higher than the cost-optimal ST algorithm. Furthermore, its computation time is also acceptable. The algorithm can help to realize real-time and scalable multicast services for IIoT applications.

This paper presents a physical layer wireless device identification method that uses a convolutional neural network (CNN) operating on a quadrant IQ transition image. This work introduces classification and detection tasks in one process. The proposed method can identify IoT wireless devices by exploiting their RF fingerprints, a technology to identify wireless devices by using unique variations in analog signals. We propose a quadrant IQ image technique to reduce the size of CNN while maintaining accuracy. The CNN utilizes the IQ transition image, which image processing cut out into four-part. An over-the-air experiment is performed on six Zigbee wireless devices to confirm the proposed identification method's validity. The measurement results demonstrate that the proposed method can achieve 99% accuracy with the light-weight CNN model with 36,500 weight parameters in serial use and 146,000 in parallel use. Furthermore, the proposed threshold algorithm can verify the authenticity using one classifier and achieved 80% accuracy for further secured wireless communication. This work also introduces the identification of expanded signals with SNR between 10 to 30dB. As a result, at SNR values above 20dB, the proposals achieve classification and detection accuracies of 87% and 80%, respectively.

Direction of arrival (DOA) estimation of wireless signals is demanded in many applications. In addition to classical methods such as MUSIC and ESPRIT, non-linear algorithms such as compressed sensing have become common subjects of study recently. Deep learning or machine learning is also known as a non-linear algorithm and has been applied in various fields. Generally, DOA estimation using deep learning is classified as on-grid estimation. A major problem of on-grid estimation is that the accuracy may be degraded when the DOA is near the boundary. To reduce such estimation errors, we propose a method of combining two DNNs whose grids are offset by one half of the grid size. Simulation results show that our proposal outperforms MUSIC which is a typical off-grid estimation method. Furthermore, it is shown that the DNN specially trained for a close DOA case achieves very high accuracy for that case compared with MUSIC.

Although deep neural networks (DNNs) have achieved high performance across a variety of applications, they can often be deceived by adversarial examples that are generated by adding small perturbations to the original images. Adversaries may generate adversarial examples using the property of transferability, in which adversarial examples that deceive one model can also deceive other models because adversaries do not obtain any information on the DNNs deployed in real scenarios. Recent studies show that adversarial examples with feature space perturbations are more transferable than others. Adversarial training is an effective method to defend against adversarial attacks. However, it results in a decrease in the classification accuracy for natural images, and it is not sufficiently robust against transferable adversarial examples because it does not consider adversarial examples with feature space perturbations. We propose a novel adversarial training method to train DNNs to be robust against transferable adversarial examples and maximize their classification accuracy for natural images. The proposed method trains DNNs to correctly classify natural images and adversarial examples and also minimize the feature differences between them. The robustness of the proposed method was similar to those of the previous adversarial training methods for MNIST dataset and was up to average 6.13% and 9.24% more robust against transfer adversarial examples for CIFAR-10 and CIFAR-100 datasets, respectively. In addition, the proposed method yielded an average classification accuracy that was approximately 0.53%, 6.82%, and 10.60% greater than some state-of-the-art adversarial training methods for all datasets, respectively. The proposed method is robust against a variety of transferable adversarial examples, which enables its implementation in security applications that may benefit from high-performance classification but are at high risk of attack.

Speech separation is the task of extracting target speech while suppressing background interference components. In applications like video telephones, visual information about the target speaker is available, which can be leveraged for multi-speaker speech separation. Most previous multi-speaker separation methods are mainly based on convolutional or recurrent neural networks. Recently, Transformer-based Seq2Seq models have achieved state-of-the-art performance in various tasks, such as neural machine translation (NMT), automatic speech recognition (ASR), etc. Transformer has showed an advantage in modeling audio-visual temporal context by multi-head attention blocks through explicitly assigning attention weights. Besides, Transformer doesn't have any recurrent sub-networks, thus supporting parallelization of sequence computation. In this paper, we propose a novel speaker-independent audio-visual speech separation method based on Transformer, which can be flexibly applied to unknown number and identity of speakers. The model receives both audio-visual streams, including noisy spectrogram and speaker lip embeddings, and predicts a complex time-frequency mask for the corresponding target speaker. The model is made up by three main components: audio encoder, visual encoder and Transformer-based mask generator. Two different structures of encoders are investigated and compared, including ResNet-based and Transformer-based. The performance of the proposed method is evaluated in terms of source separation and speech quality metrics. The experimental results on the benchmark GRID dataset show the effectiveness of the method on speaker-independent separation task in multi-talker environments. The model generalizes well to unseen identities of speakers and noise types. Though only trained on 2-speaker mixtures, the model achieves reasonable performance when tested on 2-speaker and 3-speaker mixtures. Besides, the model still shows an advantage compared with previous audio-visual speech separation works.

This paper presents a deep network based on morphological filters for Gaussian denoising. The morphological filters can be applied with only addition, max, and min functions and require few computational resources. Therefore, the proposed network is suitable for implementation using a small microprocessor. Each layer of the proposed network consists of a top-hat transform, which extracts small peaks and valleys of noise components from the input image. Noise components are iteratively reduced in each layer by subtracting the noise components from the input image. In this paper, the extensions of opening and closing are introduced as linear combinations of the morphological filters for the top-hat transform of this deep network. Multiplications are only required for the linear combination of the morphological filters in the proposed network. Because almost all parameters of the network are structuring elements of the morphological filters, the feature maps and parameters can be represented in short bit-length integer form, which is suitable for implementation with single instructions, multiple data (SIMD) instructions. Denoising examples show that the proposed network obtains denoising results comparable to those of BM3D [1] without linear convolutions and with approximately one tenth the number of parameters of a full-scale deep convolutional neural network [2]. Moreover, the computational time of the proposed method using SIMD instructions of a microprocessor is also presented.

A method to generate color palettes from images is proposed. Here, deep neural networks (DNN) are utilized in order to consider human perception. Two aspects of human perception are considered; one is attention to image, and the other is human preference for colors. This method first extracts N regions with dominant color categories from the image considering human attention. Here, N is the number of colors in a color palette. Then, the representative color is obtained from each region considering the human preference for color. Two deep neural-net systems are adopted here, one is for estimating the image area which attracts human attention, and the other is for estimating human preferable colors from image regions to obtain representative colors. The former is trained with target images obtained by an eye tracker, and the latter is trained with dataset of color selection by human. Objective and subjective evaluation is performed to show high performance of the proposed system compared with conventional methods.

The narrow bandwidth limitation of 300-3400Hz on the public switching telephone network results in speech quality deterioration. In this paper, we propose an artificial bandwidth extension approach that reconstructs the missing lower bandwidth of 50-300Hz using sinusoidal synthesis based on the first formant location. Sinusoidal synthesis generates sinusoidal waves with a harmonic structure. The proposed method detects the fundamental frequency using an autocorrelation method based on YIN algorithm, where a threshold processing avoids the false fundamental frequency detection on unvoiced sounds. The amplitude of the sinusoidal waves is calculated in the time domain from the weighted energy of 300-600Hz. In this case, since the first formant location corresponds to the first peak of the spectral envelope, we reconstruct the harmonic structure to avoid attenuating and overemphasizing by increasing the weight when the first formant location is lower, and vice versa. Consequently, the subjective and objective evaluations show that the proposed method reduces the speech quality difference between the original speech signal and the bandwidth extended speech signal.

Estimating the spatial parameters of the signals by using the effective data of a single snapshot is essential in the field of reconnaissance and confrontation. Major drawback of existing algorithms is that its constructed covariance matrix has a great degree of rank loss. The performance of existing algorithms gets degraded with low signal-to-noise ratio. In this paper, a three-parallel linear array based algorithm is proposed to achieve two-dimensional direction of arrival estimates in a single snapshot scenario. The key points of the proposed algorithm are: 1) construct three pseudo matrices with full rank and no rank loss by using the single snapshot data from the received signal model; 2) by using the rotation relation between pseudo matrices, the matched 2D-DOA is obtained with an efficient parameter matching method. Main objective of this work is on improving the angle estimation accuracy and reducing the loss of degree of freedom in single snapshot 2D-DOA estimation.

In this paper, an automatic retransmission request (ARQ) scheme for IEEE 802.11ac is presented, which can solve the problem of severe packet loss and greatly improve the performance in error-prone environments. The proposed solution only requires to be deployed on the sender and is compatible with the 802.11 protocol. The algorithm utilizes the basic strategy of sliding retransmission and then adds the method of copying frames. The media access control (MAC) protocol data unit (MPDU) lost in the transmission and the newly added data frame brought by the sliding window change are replicated. The scheme retransmits the duplicated aggregated packet and further improves the throughput by increasing the probability of successful transmission of sub-frames. Besides, we also establish a mathematical model to analyze the performance of the proposed scheme. We introduce the concept of average aggregated sub-frames and express the sliding retransmission strategy as the aggregated transmission of average aggregated sub-frames, thereby simplifying the model and effectively analyzing the theoretical throughput of the proposed algorithm. The simulation results of Network simulator 3 (NS-3) simulation results demonstrate that the performance of the proposed algorithm is better than the traditional sliding retransmission ARQ algorithm in error-prone channels with a higher physical layer rate.

In this paper, the scattering far-field from a circular electric conducting cylinder has been analyzed by physical optics (PO) approximation for both H and E polarizations. The evaluation of radiation integrations due to the PO current is conducted numerically and analytically. While non-uniform and uniform asymptotic solutions have been derived by the saddle point method, a separate approximation has been made for forward scattering direction. Comparisons among our approximation, direct numerical integration and exact solution results yield a good agreement for electrically large cylinders.

The electromagnetic environment is increasingly complex and changeable, and radar needs to meet the execution requirements of various tasks. Modern radars should improve their intelligence level and have the ability to learn independently in dynamic countermeasures. It can make the radar countermeasure strategy change from the traditional fixed anti-interference strategy to dynamically and independently implementing an efficient anti-interference strategy. Aiming at the performance optimization of target tracking in the scene where multiple signals coexist, we propose a countermeasure method of cognitive radar based on a deep Q-learning network. In this paper, we analyze the tracking performance of this method and the Markov Decision Process under the triangular frequency sweeping interference, respectively. The simulation results show that reinforcement learning has substantial autonomy and adaptability for solving such problems.