A robust neural network model for fault detection in the presence of mislabelled data

Several data-driven methodologies for process monitoring and detection of faults or abnormalities have been developed for the safety of processing systems. The effectiveness of data-based models, however, is impacted by the volume and quality of training data. This work presents a robust neural network model for addressing the mislabelled and low-quality data in detecting faults and process abnormalities. The approach is based on harnessing data quality features along with supervisory labels in the network training. The data quality has been computed using the Mahalanobis distances and trusted centres of each class of data such as normal and faulty data. The method has been examined for detecting abnormalities in two case studies; a continuous stirred tank heater problem for detecting leaks and the Tennessee Eastman chemical process for detecting step and sticking faults. The performance of the proposed robust artificial neural networks (ANN) model is evaluated in terms of accuracy, fault detection rate, false alarm rate, and classification index at varying extents of mislabelling, namely, 1%, 5%, and 10% mislabelled data. The proposed model demonstrates higher detection performance, especially at increased labels of mislabelled data where the performance of the conventional ANN is severely impacted. The proposed methodology can be advantageous in handling mislabelled and low-quality data issues which are crucial in the data-driven modelling of processing systems.     

Characteristics of Bi-metallic Interfaces Formed During Direct Energy Deposition Additive Manufacturing Processing

Metallurgical and Materials Transactions B - Various additive manufacturing processes are being evaluated to reduce the time and cost for fabrication of low volume, complex, and multifunctional...     

Teriflunomide Preserves Neuronal Activity and Protects Mitochondria in Brain Slices Exposed to Oxidative Stress

Teriflunomide (TFN) limits relapses in relapsing–remitting multiple sclerosis (RRMS) by reducing lymphocytic proliferation through the inhibition of the mitochondrial enzyme dihydroorotate dehydrogenase (DHODH) and the subsequent modulation of de novo pyrimidine synthesis. Alterations of mitochondrial function as a consequence of oxidative stress have been reported during neuroinflammation. Previously, we showed that TFN prevents alterations of mitochondrial motility caused by oxidative stress in peripheral axons. Here, we aimed to validate TFN effects on mitochondria and neuronal activity in hippocampal brain slices, in which cellular distribution and synaptic circuits are largely preserved. TFN effects on metabolism and neuronal activity were investigated by assessing oxygen partial pressure and local field potential in acute slices. Additionally, we imaged mitochondria in brain slices from the transgenic Thy1-CFP/COX8A)S2Lich/J (mitoCFP) mice using two-photon microscopy. Although TFN could not prevent oxidative stress-related depletion of ATP, it preserved oxygen consumption and neuronal activity in CNS tissue during oxidative stress. Furthermore, TFN prevented mitochondrial shortening and fragmentation of puncta-shaped and network mitochondria during oxidative stress. Regarding motility, TFN accentuated the decrease in mitochondrial displacement and increase in speed observed during oxidative stress. Importantly, these effects were not associated with neuronal viability and did not lead to axonal damage. In conclusion, during conditions of oxidative stress, TFN preserves the functionality of neurons and prevents morphological and motility alterations of mitochondria.     

Evidence for Oxidative Pathways in the Pathogenesis of PD: Are Antioxidants Candidate Drugs to Ameliorate Disease Progression?

Parkinson’s disease (PD) is a progressive neurodegenerative disorder that arises due to a complex and variable interplay between elements including age, genetic, and environmental risk factors that manifest as the loss of dopaminergic neurons. Contemporary treatments for PD do not prevent or reverse the extent of neurodegeneration that is characteristic of this disorder and accordingly, there is a strong need to develop new approaches which address the underlying disease process and provide benefit to patients with this debilitating disorder. Mitochondrial dysfunction, oxidative damage, and inflammation have been implicated as pathophysiological mechanisms underlying the selective loss of dopaminergic neurons seen in PD. However, results of studies aiming to inhibit these pathways have shown variable success, and outcomes from large-scale clinical trials are not available or report varying success for the interventions studied. Overall, the available data suggest that further development and testing of novel therapies are required to identify new potential therapies for combating PD. Herein, this review reports on the most recent development of antioxidant and anti-inflammatory approaches that have shown positive benefit in cell and animal models of disease with a focus on supplementation with natural product therapies and selected synthetic drugs.     

Dried Blood Spots as Matrix for Evaluation of Valproate Levels and the Immediate and Delayed Metabolomic Changes Induced by Single Valproate Dose Treatment

The immediate and delayed metabolic changes in rats treated with valproate (VPA), a drug used for the treatment of epilepsy, were profiled. An established approach using dried blood spots (DBS) as sample matrices for gas chromatography/mass spectrometry-based metabolomics profiling was modified using double solvents in the extraction of analytes. With the modified method, some of the previously undetectable metabolites were recovered and subtle differences in the metabolic changes upon exposure to a single dose of VPA between males and female rats were identified. In male rats, changes in 2-hydroxybutyric acid, pipecolic acid, tetratriacontane and stearic acid were found between the control and treatment groups at various time points from 2.5 h up to 24 h. In contrast, such differences were not observed in female rats, which could be caused by the vast inter-individual variations in metabolite levels within the female group. Based on the measured DBS drug concentrations, clearance and apparent volume of distribution of VPA were estimated and the values were found to be comparable to those estimated previously from full blood drug concentrations. The current study indicated that DBS is a powerful tool to monitor drug levels and metabolic changes in response to drug treatment.     

Beitr?ge zur Untersuchung und Beurteilung des Citronensaftes

Ohne ZusammenfassungMitteilung aus dem Chemischen Untersuchungsamte der Stadt Dresden.     

Waves in a Circular Plasma Column with General Impedance Boundary Conditions

Electromagnetic waves in a plasma column subjected to generalized impedance boundary conditions are studied. A general dispersion equation is obtained which is solved numerically on a high speed digital computer, the boundary impedances being either read in on cards as fixed impedances or calculated from a subroutine written for a specific problem. The effects of resistive and reactive components in the boundary impedances are shown. True zeros of the determinantal equation are found to occur for real ? only when pure reactance is presented to the plasma surface.