Residual Loss Research Articles

An Enhanced Autoencoder-based Anomaly Detection Model for time series Data from Wearable Medical Devices.

In today's era of rapidly advancing technologies, such as sensors and the Internet of Things (IoT), and the increasing focus on promoting healthy lifestyles, smart wearable devices play a crucial role in real-time detection and diagnosis of physical health conditions. Through analyzing the multi-featured time series data captured by these devices with multiple sensors, we can uncover hidden diseases and provide timely treatment. Therefore, it is imperative to study an anomaly detection model with robust feature learning and anomaly diagnosis capabilities. To address this need, this paper proposes an enhanced autoencoder-based anomaly detection model for time series data obtained from wearable medical devices. Initially, the model utilizes a convolutional neural network to learn the correlations between multiple features. Subsequently, a long and short-term memory network is employed to capture the sequence correlations, and an multi-head attention mechanism is used to mitigate the performance degradation caused by increasing the sequence length. The residual loss is also used to effectively mitigate the vanishing gradient problem. Finally, the model is evaluated using two widely recognized public datasets: the HeartDisease dataset, which contains information on patients with heart conditions, and the MIMIC dataset, a comprehensive database of de-identified health data related to critical care. The experimental results demonstrate that our model can achieve an accuracy of 95.37% and 95.56% on the two datasets, respectively. Compared to the best performing baseline methods, our model improves 8.6% and 12.3% on the two datasets, respectively. Overall, our model enables efficient analysis of sequential data, effectively captures long-term dependencies, and significantly improves the success rate of early health diagnosis for individuals.

Wideband loss separation model of nanocrystalline materials considering microscopic magnetization mechanism

The core loss directly determines the performance of high-frequency transformers. This paper presents an enhanced method for calculating core losses by integrating the LLG equation and Maxwell's diffusion equation. By the method, the losses caused by the rotation of the magnetic moment can be obtained. The loss attributed to domain wall displacement is derived from the total loss by subtracting losses due to magnetic moment rotation, further categorized into the hysteresis loss and the residual loss. The observation of the dynamic variation of nanocrystalline magnetic domains with applied excitation verifies the feasibility of the improved method. Finally, the causes of residual loss generation are also analyzed the model accuracy is verified by comparing the measured and modelled values and the traditional loss separation model.

A Rhodnius prolixus catalytically inactive Calpain protease patterns the insect embryonic dorsal-ventral axis

The calcium dependent Calpain proteases are modulatory enzymes with important roles in cell cycle control, development and immunity. In the fly model Drosophila melanogaster Calpain A cleaves Cactus/IkappaB and consequently modifies Toll signals during embryonic dorsal-ventral (DV) patterning. Here we explore the role of Calpains in the hemiptera Rhodnius prolixus, an intermediate germband insect where the Bone Morphogenetic Protein (BMP) instead of the Toll pathway plays a major role in DV patterning. Phylogenetic analysis of Calpains in species ranging from Isoptera to Diptera indicates an increase of Calpain sequences in the R. prolixus genome and other hemimetabolous species. One locus encoding each of the CalpC, CalpD and Calp7 families, and seven Calpain A/B loci are present in the R. prolixus genome. Several predicted R. prolixus Calpains display a unique architecture, such as loss of Calcium-binding EF-hand domains and loss of catalytic residues in the active site CysPc domain, yielding catalytically dead Calpains A/B. Knockdown for one of these inactive Calpains results in embryonic DV patterning defects, with expansion of ventral and lateral gene expression domains and consequent failure of germ band elongation. In conclusion, our results reveal that Calpains may exert a conserved function in insect DV patterning, despite the changing role of the Toll and BMP pathways in defining gene expression territories along the insect DV axis.

Identification of Increased Grain Length 1 (IGL1), a novel gene encoded by a major QTL for modulating grain length in rice.

A novel quantitative trait locus qIGL1, which performed a positive function in regulating grain length in rice, was cloned by the map-based cloning approach; further studies revealed that it corresponded to LOC_Os03g30530, and the IGL1 appeared to contribute to lengthening and widening of the cells on the surface of grain hulls. Grain length is a prominent determinant for grain weight and appearance quality of rice. In this study, we conducted quantitative trait locus mapping to determine a genomic interval responsible for a long-grain phenotype observed in a japonica cultivar HD385. This led to the identification of a novel QTL for grain length on chromosome 3, named qIGL1 (for Increased Grain Length 1); the HD385 (Handao 385)-derived allele showed enhancement effects on grain length, and such an allele as well as NIP (Nipponbare)-derived allele was designated qigl1 HD385 and qIGL1NIP, respectively. Genetic analysis revealed that the qigl1HD385 allele displayed semidominant effects on grain length. Fine mapping further narrowed down the qIGL1 to an ~ 70.8-kb region containing 9 open reading frames (ORFs). A comprehensive analysis indicated that LOC_Os03g30530, which corresponded to ORF6 and carried base substitutions and deletions in HD385 relative to NIP, thereby causing changes or losses of amino-acid residues, was the true gene for qIGL1. Comparison of grain traits between a pair of near-isogenic lines (NILs), termed NIL-igl1HD385 and NIL-IGL1NIP, discovered that introduction of the igl1HD385 into the NIP background significantly resulted in the elevations of grain length and 1000-grain weight. Closer inspection of grain surfaces revealed that the cell length and width in the longitudinal direction were significantly longer and greater, respectively, in NIL-igl1HD385 line compared with in NIL-IGL1NIP line. Hence, our studies identified a new semidominant natural allele contributing to the increase of grain length and further shed light on the regulatory mechanisms of grain length.

Age-induced excellence with green solvents: the impact of residual solvent and post-treatments in screen-printed carbon perovskite solar cells and modules

Screen-printed perovskite devices made using green γ-valerolactone precursors are found to improve significantly following fabrication due to slow residual solvent loss over time. This is also observed in large-scale 220 cm2 modules.

High Intensity Beam Operation of J-PARC RCS With Minimum Beam Loss

The 3-GeV RCS (Rapid Cycling Synchrotron) of J-PARC (Japan Proton Accelerator Research Complex) at present operates at a high intensity beam near to 1 MW beam power. The beam loss and the corresponding residual radiation are key issues for beam intensity ramp up. Based on detail numerical simulations and systematic beam studies the beam loss has been well mitigated to a minimum level. The residual beam loss at 1 MW beam power is mostly due to unavoidable foil scattering of the circulating beam during injection. We have identified almost all major beam loss sources and optimized to minimize the beam loss for achieving a stable operation at 800 kW beam power since April 2022.

Ion emission from 1-10 MDa salt clusters: individual charge state resolution with charge detection mass spectrometry.

Salt cluster ions produced by electrospray ionization are used for mass calibration and fundamental investigations into cluster stability and charge separation processes. However, previous studies have been limited to relatively small clusters owing to the heterogeneity associated with large, multiply-charged clusters that leads to unresolved signals in conventional m/z spectra. Here, charge detection mass spectrometry is used to measure both the mass and charge distributions of positively charged clusters of KCl, CaCl2, and LaCl3 with masses between ∼1 and 10 MDa by dynamically measuring the energy per charge, m/z, charge, and mass of simultaneously trapped individual ions throughout a 1 s trapping time. The extent of remaining hydration on the clusters, determined from the change in the frequency of ion motion with time as a result of residual water loss, follows the order KCl < CaCl2 < LaCl3, and is significantly lower than that of a pure water nanodrop, consistent with tighter water binding to the more highly charged cations in these clusters. The number of ion emission events from these clusters also follows this same trend, indicating that water at the cluster surface facilitates charge loss. A new frequency-based method to determine the magnitude of the charge loss resulting from individual ion emission events clearly resolves losses of +1 and +2 ions. Achieving this individual charge state resolution for ion emission events is an important advance in obtaining information about the late stages of bare gaseous ions formation. Future experiments on more hydrated clusters are expected to lead to a better understanding of ion formation in electrospray ionization.

Preparation, characterization, and in vitro/vivo evaluation of a multifunctional electrode coating for cochlear implants

Cochlear implantation (CI) is the primary intervention for patients with sensorineural hearing loss to restore their hearing. However, approximately 90 % of CI recipients experience unexpected fibrosis around the inserted electrode arrays due to acute and chronic inflammation. This fibrosis leads to progressive residual hearing loss. Addressing this complication is crucial for enhancing CI outcomes, yet an effective treatment has not yet been found. In this study, we developed a multifunctional dexamethasone (DXM)-loaded polytrimethylene carbonate (PTMC) electrode coating to mitigate inflammatory reactions and fibrosis after CI. This thin and flexible coating could preserve the mechanical performance of the electrode and reduce the implantation resistance for CI. The in vitro release studies demonstrated the DXM-PTMC coating's efficient drug loading and sustained release capability over 90 days. DXM-PTMC also showed long-term stability, high biocompatibility, and effective anti-inflammatory effects in vitro and in vivo. Compared with the uncoated group, DXM-PTMC coating significantly inhibited the expression of inflammatory factors, such as NO, TNF-α, IL-1β, and IL-6. DXM-PTMC coating suppressed fibrosis in rat implantation models for 3 weeks by reducing both acute and chronic inflammation. Our findings suggest that DXM-PTMC coating is a novel strategy to improve the outcomes of CI.

DSRNet: Depth Super-Resolution Network guided by blurry depth and clear intensity edges

Although high resolution (HR) depth images are required in many applications such as virtual reality and autonomous navigation, their resolution and quality generated by consumer depth cameras fall short of the requirements. Existing depth upsampling methods focus on extracting multiscale features of HR color image to guide low resolution (LR) depth upsampling, thus causing blurry and inaccurate edges in depth. In this paper, we propose a depth super-resolution (SR) network guided by blurry depth and clear intensity edges, called DSRNet. DSRNet differentiates effective edges from a number of HR edges with the guidance of blurry depth and clear intensity edges. First, we perform global residual estimation based on an encoder–decoder architecture to extract edge structure from HR color image for depth SR. Then, we distinguish effective edges from HR edges in the decoder side with the guidance of LR depth upsampling. To maintain edges for depth SR, we use intensity edge guidance that extracts clear intensity edges from HR image. Finally, we use residual loss to generate accurate high frequency (HF) residual and reconstruct HR depth maps. Experimental results show that DSRNet successfully reconstructs depth edges in SR results as well as outperforms the state-of-the-art methods in terms of visual quality and quantitative measurements.11The proposed model with some test image pairs are available in https://github.com/lanhui-123/DSRNet.

Microstructure and properties of FeSiCr/Al2O3/epoxy resin soft magnetic composites with high flowability by dry milling and coupling modification

FeSiCr/Al2O3 modified powders were prepared by dry milling and coupling for different durations, and FeSiCr/Al2O3 epoxy resin composites were prepared by banburying and low-pressure molding. The melt index, bending strength, magnetic permeability, and magnetic loss of the FeSiCr/Al2O3 epoxy resin composites were measured. The microstructures of the different FeSiCr/Al2O3 modified powders and FeSiCr/Al2O3 epoxy resin composites were observed using scanning electron microscopy (SEM), and their effects on the flowability, magnetic, and mechanical properties of the composites are discussed. The results showed that dry milling and coupling modification greatly improved the flowability of the FeSiCr/Al2O3/epoxy resin composites. Moreover, the bending strength and effective magnetic permeability increased, and the core loss decreased. Based on the loss data, a new loss separation method that is different from the conventional low-frequency method is proposed. Residual, hysteresis, and eddy current loss were in descending order in the frequency range of 100–2000 kHz.

Enhancing physics informed neural networks for solving Navier–Stokes equations

SummaryFluid mechanics is a critical field in both engineering and science. Understanding the behavior of fluids requires solving the Navier–Stokes equation (NSE). However, the NSE is a complex partial differential equation that can be challenging to solve, and classical numerical methods can be computationally expensive. In this paper, we propose enhancing physics‐informed neural networks (PINNs) by modifying the residual loss functions and incorporating new computational deep learning techniques. We present two enhanced models for solving the NSE. The first model involves developing the classical PINN for solving the NSE, based on a stream function approach to the velocity components. We have added the pressure training loss function to this model and integrated the new computational training techniques. Furthermore, we propose a second, more flexible model that directly approximates the solution of the NSE without making any assumptions. This model significantly reduces the training duration while maintaining high accuracy. Moreover, we have successfully applied this model to solve the three‐dimensional NSE. The results demonstrate the effectiveness of our approaches, offering several advantages, including high trainability, flexibility, and efficiency.

Effect of geotextiles with different masses per unit area on water loss and cracking under bottom water loss soil conditions

Soil water loss is an important component of the water balance in irrigated agriculture. This study investigated the effects of geotextiles on water loss during soil drying and cracking. The results indicate that the residual water content of soil samples increased by 98.5%, 145.5%, and 164.7% with geotextile masses per unit area of 100, 300, and 400 g/m2, compared that of soil without geotextiles. There are two water loss stages of soil, the "rapid loss" and the "residual loss,” under the condition of bottom water loss, which is different from the evaporation stage of normal soil without bottom water loss. When a geotextile is added to the soil, the stages of bottom water loss will become "rapid loss, deceleration loss, and residual loss." When the mass per unit area of 400 g/m2 geotextile was used, the crack ratio, probability entropy, and fractal dimensions decreased by 15.19%, 6.47%, and 5.81%, respectively. The geotextile mass per unit area increased the specific surface area of the soil, and water retention was improved. When the mass per unit area of the geotextile increased, the interface friction between the soil and geotextile increased, and the cracking of the soil was effectively inhibited.

Effects of magnetic domain morphology on the magnetic spectrum and high‐frequency core losses of MnZn ferrites

AbstractMnZn ferrites that exhibit low core losses at high frequencies are suitable for the integration and miniaturization of power conversion devices. To obtain low core losses at high frequencies, it is important to understand the underlying mechanism. In this work, MnZn ferrites with various grain sizes were produced by changing the sintering temperature. The relationship between domain morphology, dynamic magnetization behavior, and core losses at high frequencies was carefully investigated. As the average grain size approached the critical size of single‐domain state, the contribution of domain wall displacement to complex permeability decreased, while domain wall resonant frequency improved significantly, resulting in the suppression of residual loss at 3 MHz. As the average grain size increased, the domain morphology changed from a single‐domain state to a multidomain state, which increased the initial permeability and decreased hysteresis loss. However, the number of domain walls increased, causing a significant decrease in the domain wall resonant frequency, and resulting in a sharp increase in residual loss. These results provide valuable insights into reducing high‐frequency core losses in MnZn ferrites by domain engineering.

Finger vein denoising algorithm based on gradient-oriented residual structure and LBP texture loss

Finger vein denoising algorithm based on gradient-oriented residual structure and LBP texture loss

Ultrafast synthesis of amorphous molybdenum sulfide by magnetic induction heating for hydrogen evolution reaction

Molybdenum sulfides have emerged as viable alternatives to noble-metal catalysts for green hydrogen production via the hydrogen evolution reaction (HER). Herein, magnetic induction heating (MIH) is exploited for the rapid preparation of carbon-supported MoSx nanocomposites. The sample prepared at 200 A for 10 s shows an amorphous Mo3S7Cly-like structure and a low overpotential (η10) of −184 mV at 10 mA cm−2 in acidic media, whereas samples prepared at higher induction currents display a largely crystalline MoS2 structure and drastically lower HER performance. This is due to the formation of dimeric Mo6S14 moieties in amorphous MoSx that is facilitated by the loss of Cl residues during electrochemical reaction and enhanced H adsorption at both the S and Mo sites, in comparison to crystalline MoS2, as shown in First-principles calculations. These results show that MIH may be used as a powerful tool in the preparation of nonequilibrium structures as high-performance electrocatalysts.

Residual Bond Strength of Highly Corroded Reinforced Concrete

Corrosion of reinforcement due to chloride attacks in the marine environment is a natural phenomenon. Corrosion of reinforcement produces corrosion products of high volume, which deteriorates the structural integrity of reinforced concrete structures due to loss of bond, cracking, and spalling of the concrete. Existing literature has documented tests investigating the bond behavior of uncorroded and corroded specimens, but there is a dearth of data pertaining to a more advanced stage (higher mass loss) of corrosion. In the current study, an accelerated corrosion test was conducted on cylindrical (lollypop) specimens, which involved utilizing an impressed current laboratory technique to induce three distinct levels of corrosion (10%, 20%, and 40% mass loss). Moreover, to assess bond strength, the pull-out tests were performed on both corroded and uncorroded specimens. The present study deals with the residual bond strength at three different corrosion levels, as a function of different parameters such as clear cover (CC), water-cement (W/C) ratio, and two different reinforcement diameters. Experimental data reveals that the mass loss achieved is lesser than the target mass loss for all specimens. It is observed that at higher corrosion levels, where the mass loss exceeds 10% or cracks appear on the surface of the reinforced structure, both an increase in mass loss and a decrease in residual bond strength is consistently observed. These effects remain consistent regardless of whether the parameters such as bar diameter, W/C ratio, and CC are increased or decreased. The statistical analysis was performed on the experimental data to develop predictive models for estimating the residual bond strength and mass loss. For higher mass loss of 30% to 35%, the corresponding bond strength for all of the specimens falls within the range of 4 Mpa to 6 MPa.

Biodegradation of LDPE (low density polyethylene) using bacterial strain

The applications of polyethylene are enormous and they are popular because of their relatively low cost, lack of difficulty in manufacture, adaptability and durability. In the past decade, the increase in the utilization of plastics and the accumulation of polyethylene in the environment all around the globe has drawn the attention of environmentalists and scientists. The plastic materials are commonly derived from petrochemicals extracted from coal and natural gas. They are produced in a wide range in synthetic, semi synthetic organic polymers. The degradation of LDPE film was determined by residual dry weight loss of the LDPE films. The potent strain showed 42.5% degradation ability in 90 d. Degradation of LDPE results in the breakdown of the polymer backbone chain producing CO2 which was checked by gravimetric analysis. Enzyme kinetic studies showed the production of three ligninolytic extracellular enzymes- lignin, laccase and manganese peroxidase by bacterial strain AJ01 in LDPE degradation. Degradation of LDPE strips followed first order kinetics model with a rate constant (R2) of 0.9783 d-1. Estimation of protein from post treated LDPE using bacterial isolate showed a total concentration of 0.440mg/0.1mL and 0.703mg/0.2mL. LDPE degradation was confirmed by using analytical techniques such as SEM which showed changes in the surface topology of LDPE strips and FTIR analysis showed changes in complex chemical structure of LDPE post 90 d of degradation using microorganisms thereby reducing carbonyl index (CI) value.

Enhancing the efficiency of an axial impulse turbine with a diffuser

In oscillating water column wave energy converters, the aerodynamic losses caused by flow misalignment between the rotor outlet and the outlet guide vane of self-rectifying impulse turbines are a significant design problem. The symmetrical positioning of guiding vanes on both sides of the rotor is the reason for it. In comparison to the equivalent Wells turbine, the efficiency of impulse turbines is lower because of these losses in the outlet guide vane. This paper presents a strategy to develop animproved design of an axial impulse turbine to reduce the losses in the outlet guide vane along with the residual kinetic energy loss by integrating a diffuser between the rotor and guide vanes. The proposed model was investigated numerically using RANS simulations. Multiple diffuser geometries were tested against the reference turbine for comparison of the performance characteristics. The results support the hypothesis behind the proposed design. The performance was compared extensively with the reference case in terms of the dimensionless loss coefficients for a better insight into the contribution of all the turbine sectors. The current work shows a possible design path for the performance improvement of the turbine. Results also support the fact that the guide vanes need to beredesigned for obtaining maximizing efficiency with the proposed design. The losses in the outlet guide vane were reduced by approximately 50% with an overall efficiency rise of around 2%.

Engineering Attributes of Ternary Geopolymer Mortars Containing High Volumes of Palm Oil Fuel Ash: Impact of Elevated Temperature Exposure

Geopolymer mortars made from various waste products can appreciably reduce carbon dioxide emissions and landfill-related issues, making them viable substitutes for ordinary Portland cement, a workhorse in the concrete industry. Thus, a series of ternary geopolymer mortars were made and characterized to determine the effects of exposure to elevated temperatures (from room temperature up to 900 °C) on their engineered (residual compressive strength, weight loss, and slant shear bond strength) and microstructural properties. These mortars, which contain fly ash, ground blast furnace slag, and a high volume of palm oil fuel ash, were designed to activate via the incorporation of an alkali activator solution at a low concentration (molarity of 4). The elevated temperature-mediated deterioration of the ternary geopolymer mortar was quantified using Fourier transform infrared spectroscopy, X-ray diffraction, scanning electron microscopy, and thermogravimetric analysis. The results revealed an improvement in the ternary geopolymer mortars’ resistance against elevated temperatures when the palm oil fuel ash level in the mortar matrix was raised from 50 to 70% and when slag was replaced by fly ash. It was asserted that the proposed ternary geopolymer mortars may contribute to the advancement of green concretes demanded by the construction sectors.

Machine-learning-assisted prediction and optimized kinetic modelling of residual chlorine decay for enhanced water quality management

The quality of water changes from source to tap, presenting challenges in maintaining consistent water quality across the system. Predicting water quality in distribution systems, including disinfectant residual loss and by-product formation, has been the subject of research since the early 1990s. Although numerous models have been proposed to predict residual chlorine decay, disputes exist among researchers and experts over the superiority of certain models. Accordingly, this study modified the existing process-based bulk decay models by replacing the initial Total Residual Chlorine (TRC) concentration parameter with TRC demand, leading to an improvement in the models' performance. The modification resulted in a 38.03%, 28.02%, 23.11%, and 33.29% average improvement in Mean Squared Error (MSE) values for the First Order Model (FOM), Parallel First Order Model (PFOM), Second Order Model (SOM), and Parallel Second Order Model (PSOM), respectively. The study also introduced an online predictive method based on a Machine Learning (ML) algorithm that predicts the first-order TRC bulk decay rate by using water quality parameters as inputs. A Gaussian Process Regression (GPR) model was used to predict the kinetic parameters in FOM, which accurately predicted the test sets for most of the cases. In addition, a new methodology was proposed in this study for predicting TRC in water distribution systems that incorporates the variability of source natural organic matter, operational actions, and water demands. This method seeks to develop high-fidelity and robust water quality predictions that provide operational decision support for optimized distribution system management. In conclusion, this study emphasizes the importance of understanding water quality changes from source to tap and the challenges of maintaining consistent water quality across the system. The study suggests modifying existing models and introducing a novel methodology for predicting residual chlorine in water distribution systems that can improve water quality management and, ultimately, better public health outcomes.