Analysis of medical images super-resolution via a wavelet pyramid recursive neural network constrained by wavelet energy entropy.

Recently, multi-resolution pyramid-based techniques have emerged as the prevailing research approach for image super-resolution. However, these methods typically rely on a single mode of information transmission between levels. In our approach, a wavelet pyramid recursive neural network (WPRNN) based on wavelet energy entropy (WEE) constraint is proposed. This network transmits previous-level wavelet coefficients and additional shallow coefficient features to capture local details. Besides, the parameter of low- and high-frequency wavelet coefficients within each pyramid level and across pyramid levels is shared. A multi-resolution wavelet pyramid fusion (WPF) module is devised to facilitate information transfer across network pyramid levels. Additionally, a wavelet energy entropy loss is proposed to constrain the reconstruction of wavelet coefficients from the perspective of signal energy distribution. Finally, our method achieves the competitive reconstruction performance with the minimal parameters through an extensive series of experiments conducted on publicly available datasets, which demonstrates its practical utility.

Security-guaranteed filter design for discrete-time Markovian jump delayed systems subject to deception attacks and sensor saturation.

This work is devoted the problem of a security-guaranteed filter design for a class of discrete-time Markov jump systems that are vulnerable to stochastic deception attacks and have random sensor saturation. Deception attacks, in particular, are taken into account in the filter when the attacker attempts to modify the broadcast signal in communication networks by inserting some misleading information data into the assessment output. The Bernoulli distribution is satisfied by two sets of introduced stochastic variables. It shows the likelihood that the broadcaster's data transmissions will be the focus of deception attacks and sensor saturation. The Lyapunov functional technique is established, and criteria are derived to ensure that the system is mean-square stable. Furthermore, explicit expression of the filter gains is obtained by solving a set of linear matrix inequalities. Lastly, two simulation examples including a synthetic genetic regulatory network are provided to further demonstrate the validity and efficiency of the suggested theoretical results.

Stabilization of Periodic Piecewise Time-Varying Systems With Time-Varying Delay Under Multiple Cyber Attacks: An Augmented Lyapunov Functional Approach.

This article investigates the asymptotic stabilization of periodic piecewise time-varying systems with time-varying delay under various cyber attacks, particularly deception and DoS attacks. The addressed system is reformed into a number of time-varying subsystems based on the time interval for each period. Following that, a state-feedback controller with periodic time-varying gain parameters is developed to solve the stabilization problem. The control design depicts the possibility of the aforementioned cyber attacks with two mutually exclusive stochastic Bernoulli distributed parameters. Then, an augmented Lyapunov-Krasovskii functional with periodically varying matrices is used to determine the conditions for designing the proposed controller that ensures the mean-square asymptotic stability of the addressed system. The results of numerical examples support the conclusion that the proposed method is effective and superior, regardless of the cyber attacks involved.

Nonfragile power and frequency control in islanded microgrids under abnormal asynchronous stochastic cyber attacks.

In this paper, we focus on the real power sharing and frequency regulation of distributed generators in islanded microgrids with abnormal asynchronous stochastic cyber attacks, which is of great significance to the information security and stable operation of microgrids. Firstly, considering the possible cyber attacks in the communication network, a distributed non-fragile controller with coupled memory delay is proposed according to the nonperiodic sampled-data control. Then, the construction of delay-dependent two-sided looped-functional makes the Lyapunov-Krasovskii functional contain more delay and sampling information and relaxes constraints on free matrices. In addition, based on the enhanced integral inequality technique and the linear convex combination method, a sampling-based consensus protocol is presented to solve the issues of real power sharing and frequency regulation in the islanded microgrids. Finally, to verify the effectiveness and feasibility of the designed control strategy, a modified IEEE 34-bus test system is used for the experiment.

Observer-based state estimation for discrete-time semi-Markovian jump neural networks with round-robin protocol against cyber attacks.

This paper investigates an observer-based state estimation issue for discrete-time semi-Markovian jump neural networks with Round-Robin protocol and cyber attacks. In order to avoid the network congestion and save the communication resources, the Round-Robin protocol is used to schedule the data transmissions over the networks. Specifically, the cyber attacks are modeled as a set of random variables satisfying the Bernoulli distribution. On the basis of the Lyapunov functional and the discrete Wirtinger-based inequality technique, some sufficient conditions are established to guarantee the dissipativity performance and mean square exponential stability of the argument system. In order to compute the estimator gain parameters, a linear matrix inequality approach is utilized. Finally, two illustrative examples are provided to demonstrate the effectiveness of the proposed state estimation algorithm.

Fixed-Time Stability of Nonlinear Impulsive Systems and Its Application to Inertial Neural Networks.

This article is concerned with the fixed-time stability (FTS) problem of nonlinear impulsive systems (NISs). By means of the impulsive control mechanism and Lyapunov functions theory, several sufficient conditions are established to ensure the FTS of general NISs. Meanwhile, some novel impulse-dependent settling-time estimation schemes are developed, which fully considers the influence of stabilizing impulses and destabilizing impulses on the convergence rate of the system states. The proposed schemes establish a quantitative relationship between the upper bound of the settling time and impulse effects. It shows that stabilizing impulses can accelerate the convergence rate of the system states and leads to the upper bound of the settling time being smaller. Conversely, destabilizing impulses can reduce it and make the upper bound of the settling time larger. Then, the theoretical results are applied to delayed inertial neural networks (DINNs), where two kinds of controllers are designed to realize fixed-time synchronization of the considered systems in the impulse sense. Finally, some numerical examples are provided to illustrate the validity of the proposed theoretical results.

How to handle noisy labels for robust learning from uncertainty.

Most deep neural networks (DNNs) are trained with large amounts of noisy labels when they are applied. As DNNs have the high capacity to fit any noisy labels, it is known to be difficult to train DNNs robustly with noisy labels. These noisy labels cause the performance degradation of DNNs due to the memorization effect by over-fitting. Earlier state-of-the-art methods used small loss tricks to efficiently resolve the robust training problem with noisy labels. In this paper, relationship between the uncertainties and the clean labels is analyzed. We present novel training method to use not only small loss trick but also labels that are likely to be clean labels selected from uncertainty called "Uncertain Aware Co-Training (UACT)". Our robust learning techniques (UACT) avoid over-fitting the DNNs by extremely noisy labels. By making better use of the uncertainty acquired from the network itself, we achieve good generalization performance. We compare the proposed method to the current state-of-the-art algorithms for noisy versions of MNIST, CIFAR-10, CIFAR-100, T-ImageNet and News to demonstrate its excellence.

Application of Macro-Instrumented Indentation Test for Superficial Residual Stress and Mechanical Properties Measurement for HY Steel Welded T-Joints.

HY-80 and HY-100 steels, widely used in constructing large ocean vessels and submarine hulls, contain mixed microstructures of tempered bainite and martensite and provide high tensile strength and toughness. Weld integrity in HY steels has been studied to verify and optimize welding conditions. In this study, the T-joint weld coupons, HY80 and HY100, were fabricated from HY-80 and HY-100 steel plates with a thickness of 30 mm as base metals by submerged-arc welding. Flux-cored arc welding was performed on an additional welding coupon consisting of HY-100 to evaluate the effect of repair welds (HY100RP). Microstructures in the heat-affected zones (HAZ) were thoroughly analyzed by optical observation. Instrumented indentation testing, taking advantage of local characterization, was applied to assess the yield strength and the residual stress of the HAZ and base regions. The maximum hardness over 400 HV was found in the HAZ due to the high volume fraction of untempered martensite microstructure. The yield strength of the weld coupons was evaluated by indentation testing, and the results showed good agreement with the uniaxial tensile test (within 10% range). The three coupons showed similar indentation residual stress profiles on the top and bottom surfaces. The stress distribution of the HY100 coupon was comparable to the results from X-ray diffraction. HY100RP demonstrated increased tensile residual stress compared to the as-welded coupon due to the effect of the repair weld (323 and 103 MPa on the top and bottom surfaces). This study verifies the wide applicability of indentation testing in evaluating yield strength and residual stress.

Closeness-Centrality-Based Synchronization Criteria for Complex Dynamical Networks With Interval Time-Varying Coupling Delays.

This paper investigates synchronization in complex dynamical networks (CDNs) with interval time-varying delays. The CDNs are representative of systems composed of a large number of interconnected dynamical units, and for the purpose of the mathematical analysis, the leading work is to model them as graphs whose nodes represent the dynamical units. At this time, we take note of the importance of each node in networks. One way, in this paper, is that the closeness-centrality mentioned in the field of social science is grafted onto the CDNs. By constructing a suitable Lyapunov-Krasovskii functional, and utilizing some mathematical techniques, the sufficient and closeness-centrality-based conditions for synchronization stability of the networks are established in terms of linear matrix inequalities. Ultimately, the use of the closeness-centrality can be weighted with regard to not only the interconnection relation among the nodes, which was utilized in the existing works but also more information about nodes. Here, the centrality will be added as the concerned information. Moreover, to avoid the computational burden causing the nonconvex term including the square of the time-varying delay, how to deal with it is applied by estimating it to the convex term including time-varying delay. Finally, two illustrative examples are given to show the advantage of the closeness-centrality in point of the robustness on time-delay.

Potential of phosphoric acid-catalyzed pretreatment and subsequent enzymatic hydrolysis for biosugar production from Gracilaria verrucosa.

This study combined phosphoric acid-catalyzed pretreatment and enzymatic hydrolysis to produce biosugars from Gracilaria verrucosa as a potential renewable resource for bioenergy applications. We optimized phosphoric acid-catalyzed pretreatment conditions to 1:10 solid-to-liquid ratio, 1.5 % phosphoric acid, 140 °C, and 60 min reaction time, producing a 32.52 ± 0.06 % total reducing sugar (TRS) yield. By subsequent enzymatic hydrolysis, a 68.61 ± 0.90 % TRS yield was achieved. These results demonstrate the potential of phosphoric acid to produce biosugars for biofuel and biochemical production applications.

PDE 5 inhibition with udenafil improves left ventricular systolic/diastolic functions and exercise capacity in patients with chronic heart failure with reduced ejection fraction; A 12-week, randomized, double-blind, placebo-controlled trial.

BACKGROUND: Impaired nitric oxide-mediated pulmonary vascular tone is commonly found in heart failure with reduced ejection fraction (HFrEF), and is associated with derangement of left ventricular (LV) hemodynamics and decreased exercise capacity, which may be reversed by PDE5 inhibitor. This study investigated the effects of a new, long-acting PDE5 inhibitor on LV hemodynamics and exercise capacity in HFrEF. METHODS: Patients with chronic HFrEF on optimal medical therapy for >30 days before enrollment were randomly assigned to placebo or udenafil at a dose of 50mg 2x/day for the first 4 weeks followed by 100mg 2x/day for the next 8 weeks. All patients underwent cardiopulmonary exercise echocardiography before and after the 12-week treatment. RESULTS: Improvement of subjective functional capacity was more frequently reported in the udenafil group (P = 0.002). Also, a higher increase in peak VO2 (Δpeak VO2, 21.6% (6.9 ~ 106.4%) vs 1.9% (-15.7 ~ 21.0%) in the placebo group, P = 0.04) and a larger decrease in ventilatory efficiency were observed in the udenafil group (Δ-6.4 ± 9.7 vs Δ1.9 ± 12.1 in the placebo group, P = 0.03). Regarding LV systolic function, the extent of increment in LV ejection fraction was significantly greater in the udenafil group (6.6 ± 6.4% vs 2.3 ± 4.8% in the placebo group, P = 0.02). In the udenafil group, an echocardiographic surrogate of LV filling pressure was more prominently decreased (P = 0.006) along with a significant reverse remodeling of left atrial volume index (57 ± 25mL at baseline to 44 ± 23 at 12th week, P = 0.04) and a progressive fall in B-type natriuretic peptide level (589 ± 679pg/mL at baseline to 220 ± 225pg/mL at 12th week, P < 0.001), indicating LV diastolic function improvement. Udenafil was well tolerated without excess of adverse events compared to placebo. CONCLUSIONS: Udenafil improves LV systolic/diastolic functions and exercise capacity in conjunction with established conventional pharmacotherapy, without significant adverse events in HFrEF.

Extended dissipative analysis for neural networks with time-varying delays.

In this brief, an extended dissipativity analysis was conducted for a neural network with time-varying delays. The concept of the extended dissipativity can be used to solve for the H∞, L2-L∞, passive, and dissipative performance by adjusting the weighting matrices in a new performance index. In addition, the activation function dividing method is modified by introducing a tuning parameter. Examples are provided to show the effectiveness and less conservatism of the proposed method.

Assessment of left ventricular mass in hypertrophic cardiomyopathy by real-time three-dimensional echocardiography using single-beat capture image.

BACKGROUND: Left ventricular (LV) mass is an important prognostic indicator in hypertrophic cardiomyopathy. Although LV mass can be easily calculated using conventional echocardiography, it is based on geometric assumptions and has inherent limitations in asymmetric left ventricles. Real-time three-dimensional echocardiographic (RT3DE) imaging with single-beat capture provides an opportunity for the accurate estimation of LV mass. The aim of this study was to validate this new technique for LV mass measurement in patients with hypertrophic cardiomyopathy. METHODS: Sixty-nine patients with adequate two-dimensional (2D) and three-dimensional echocardiographic image quality underwent cardiac magnetic resonance (CMR) imaging and echocardiography on the same day. Real-time three-dimensional echocardiographic images were acquired using an Acuson SC2000 system, and CMR-determined LV mass was considered the reference standard. Left ventricular mass was derived using the formula of the American Society of Echocardiography (M-mode mass), the 2D-based truncated ellipsoid method (2D mass), and the RT3DE technique (RT3DE mass). RESULTS: The mean time for RT3DE analysis was 5.85 ± 1.81 min. Intraclass correlation analysis showed a close relationship between RT3DE and CMR LV mass (r = 0.86, P < .0001). However, LV mass by the M-mode or 2D technique showed a smaller intraclass correlation coefficient compared with CMR-determined mass (r = 0.48, P = .01, and r = 0.71, P < .001, respectively). Bland-Altman analysis showed reasonable limits of agreement between LV mass by RT3DE imaging and by CMR, with a smaller positive bias (19.5 g [9.1%]) compared with that by the M-mode and 2D methods (-35.1 g [-20.2%] and 30.6 g [17.6%], respectively). CONCLUSIONS: RT3DE measurement of LV mass using the single-beat capture technique is practical and more accurate than 2D or M-mode LV mass in patients with hypertrophic cardiomyopathy.

Stability for neural networks with time-varying delays via some new approaches.

This paper considers the problem of delay-dependent stability criteria for neural networks with time-varying delays. First, by constructing a newly augmented Lyapunov-Krasovskii functional, a less conservative stability criterion is established in terms of linear matrix inequalities. Second, by proposing novel activation function conditions which have not been proposed so far, further improved stability criteria are proposed. Finally, three numerical examples used in the literature are given to show the improvements over the existing criteria and the effectiveness of the proposed idea.