An Adaptive Noise Reduction Method Based on Improved Dislocation Superposition Method for Abnormal Noise Fault Component of Automotive Engine

The acoustic signal generated by mechanical motion contains the information of its motion state, but when the signal-to-noise ratio (SNR) is low, the accuracy of real-time monitoring mechanical motion state by the acoustic signal is low. This study proposes an adaptive noise reduction method based on the dislocation superposition method (DSM), which can realize the adaptive noise reduction and the extraction of fault a component from the automobile engine abnormal noise signal of low SNR. Firstly, the wavelet coefficients of engine abnormal noise signal are obtained by continuous wavelet transform (CWT), and the fault feature points of the abnormal noise signal in each period are extracted by setting hard threshold function, window function, and feature points extraction algorithm. Then, the signal segments containing fault components are obtained by using the position of feature points to extend the length of the fault component forward and backward, respectively, and Pearson’s correlation is calculated by traversal to determine the starting superposition point of each signal segment containing fault components. Finally, the signal segments of the odd group and even group are selected for superposition calculation. When the superposition stop condition is not satisfied, the number of superpositions increased until the stop condition is satisfied, and the superposition signal can be used as a fault component. The experimental results show that, compared with the improved DSM, this method has a good effect on the noise reduction and extraction of fault components of automobile engine cylinder knocking fault, and the effectiveness of this method is verified. This method is used to reduce the noise and extract the fault components of automobile engine cylinder missing fault and knock fault, and good results are obtained.

Classifying the tracing difficulty of 3D neuron image blocks based on deep learning

Quickly and accurately tracing neuronal morphologies in large-scale volumetric microscopy data is a very challenging task. Most automatic algorithms for tracing multi-neuron in a whole brain are designed under the Ultra-Tracer framework, which begins the tracing of a neuron from its soma and traces all signals via a block-by-block strategy. Some neuron image blocks are easy to trace and their automatic reconstructions are very accurate, and some others are difficult to trace and their automatic reconstructions are inaccurate or incomplete. The former are called low Tracing Difficulty Blocks (low-TDBs), while the latter are called high Tracing Difficulty Blocks (high-TDBs). We design a model named 3D-SSM to classify the tracing difficulty of 3D neuron image blocks, which is based on 3D Residual neural Network (3D-ResNet), Fully Connected Neural Network (FCNN) and Long Short-Term Memory network (LSTM). 3D-SSM contains three modules: Structure Feature Extraction (SFE), Sequence Information Extraction (SIE) and Model Fusion (MF). SFE utilizes a 3D-ResNet and a FCNN to extract two kinds of features in 3D image blocks and their corresponding automatic reconstruction blocks. SIE uses two LSTMs to learn sequence information hidden in 3D image blocks. MF adopts a concatenation operation and a FCNN to combine outputs from SIE. Neuron images of a whole mouse brain are segmented into many overlapped 3D image blocks along the gold standard reconstruction of each marked neuron. With the help of gold standard reconstructions and APP2 automatic reconstructions, 12732 training samples and 5342 test samples are constructed based on those 3D image blocks. The 3D-SSM achieves classification accuracy rate 87.04% on the training set and 84.07% on the test set. The trained 3D-SSM is tested on samples from another whole mouse brain and its accuracy rate is 83.21%. 3D-SSM can be used as a stop condition of an automatic tracing algorithm in the Ultra-Tracer framework. With its help, neuronal signals in low-TDBs can be traced by the automatic algorithm and in high-TDBs may be reconstructed by annotators.

Analysis of Footprint of Uncertainty of the IT2 Mamdani Controllers Using Different Type-Reducers

Appropriate Footprint of Uncertainties (FOUs) are beneficial to the performance of Interval Type-2 (IT2) fuzzy controller, revealing the effect of FOUs is a key problem. In our published work, as the FOUs increase, the IT2 Mamdani and TS fuzzy controllers, using KM or EKM type-reducer (TR), approach the constant and piecewise linear controllers, respectively, while they finally become constant and piecewise linear controllers. To verify the validation of the above results, when a different TR is used, in this study, the effects of other popular TRs (i.e., Nie-Tan, Wu-Mendel, Iterative Algorithm with Stop Condition) on output of IT2 Mamdani fuzzy controller, are explored. We proven that, (1) as the FOUs increase, irrespectively of the TRs used, the IT2 Mamdani fuzzy controllers approach constant controllers, (2) when all the FOUs are equal to 1 (i.e., at their maximum ), the fuzzy controllers using Nie-Tan and Iterative Algorithm with Stop Condition TR become constant controllers. The FOUs of the controllers using Wu-Mendel TR can be infinitely approaching 1 and cannot be equal to 1 (otherwise, the denominator of the TR output expression are equal to 0), hence when FOUs are infinitely approaching 1, the controller will approach the constant controller infinitely. These results imply regardless of which popular TR is used, the IT2 Mamdani fuzzy controller, when using larger FOUs, the fluctuation of the input variables have a limited impact on the output, the ability to deal with system uncertainties will deteriorate. Laboratory control experiments are provided to demonstrate these findings.

Operational Risk Reverse Stress Testing: Optimal Solutions

Selecting a suitable method to solve a black-box optimization problem that uses noisy data was considered. A targeted stop condition for the function to be optimized, implemented as a stochastic algorithm, makes established Bayesian methods inadmissible. A simple modification was proposed and shown to improve optimization the efficiency considerably. The optimization effectiveness was measured in terms of the mean and standard deviation of the number of function evaluations required to achieve the target. Comparisons with alternative methods showed that the modified Bayesian method and binary search were both performant, but in different ways. In a sequence of identical runs, the former had a lower expected value for the number of runs needed to find an optimal value. The latter had a lower standard deviation for the same sequence of runs. Additionally, we suggested a way to find an approximate solution to the same problem using symbolic computation. Faster results could be obtained at the expense of some impaired accuracy and increased memory requirements.

An ERP study on proactive and reactive response inhibition in individuals with schizotypy

Schizotypy, a subclinical group at risk for schizophrenia, has been found to show impairments in response inhibition. However, it remains unclear whether this impairment is accompanied by outright stopping (reactive inhibition) or preparation for stopping (proactive inhibition). We recruited 20 schizotypy and 24 non-schizotypy individuals to perform a modified stop-signal task with electroencephalographic (EEG) data recorded. This task consists of three conditions based on the probability of stop signal: 0% (no stop trials, only go trials), 17% (17% stop trials), and 33% (33% stop trials), the conditions were indicated by the colour of go stimuli. For proactive inhibition (go trials), individuals with schizotypy exhibited significantly lesser increase in go response time (RT) as the stop signal probability increasing compared to non-schizotypy individuals. Individuals with schizotypy also exhibited significantly increased N1 amplitude on all levels of stop signal probability and increased P3 amplitude in the 17% stop condition compared with non-schizotypy individuals. For reactive inhibition (stop trials), individuals with schizotypy exhibited significantly longer stop signal reaction time (SSRT) in both 17% and 33% stop conditions and smaller N2 amplitude on stop trials in the 17% stop condition than non-schizotypy individuals. These findings suggest that individuals with schizotypy were impaired in both proactive and reactive response inhibition at behavioural and neural levels.

A Semi-supervised Learning Approach Based on Adaptive Weighted Fusion for Automatic Image Annotation

To learn a well-performed image annotation model, a large number of labeled samples are usually required. Although the unlabeled samples are readily available and abundant, it is a difficult task for humans to annotate large numbers of images manually. In this article, we propose a novel semi-supervised approach based on adaptive weighted fusion for automatic image annotation that can simultaneously utilize the labeled data and unlabeled data to improve the annotation performance. At first, two different classifiers, constructed based on support vector machine and covolutional neural network, respectively, are trained by different features extracted from the labeled data. Therefore, these two classifiers are independently represented as different feature views. Then, the corresponding features of unlabeled images are extracted and input into these two classifiers, and the semantic annotation of images can be obtained respectively. At the same time, the confidence of corresponding image annotation can be measured by an adaptive weighted fusion strategy. After that, the images and its semantic annotations with high confidence are submitted to the classifiers for retraining until a certain stop condition is reached. As a result, we can obtain a strong classifier that can make full use of unlabeled data. Finally, we conduct experiments on four datasets, namely, Corel 5K, IAPR TC12, ESP Game, and NUS-WIDE. In addition, we measure the performance of our approach with standard criteria, including precision, recall, F-measure, N+, and mAP. The experimental results show that our approach has superior performance and outperforms many state-of-the-art approaches.

Conditional Expression Fault Classification and Repair Method of Aviation Control Software

Aviation control software has become the core control decision-making unit of the aviation system. The Boolean conditional expressions are the main parts of the branch and loop control logic of aviation control software. This paper studies the fault classification and repair method of conditional expression of aviation control software. 1) a two-level Boolean conditional expression fault classification method is proposed; 2) based on the design of mutation operator, repair solution and stop condition of mutation repair, an automatic repair method of conditional expression is proposed; 3) a repair assistant tool for conditional expression is designed and implemented, and 155 fault expressions are repaired. The results show that 55.5% of fault expressions can achieve accurate repair, 44.5% of fault expressions can generate multiple correct versions after a repair, and some versions have operation redundancy, so it is necessary to select the optimal version manually.

Adaptive Algorithm on Block-Compressive Sensing and Noisy Data Estimation

In this paper, an altered adaptive algorithm on block-compressive sensing (BCS) is developed by using saliency and error analysis. A phenomenon has been observed that the performance of BCS can be improved by means of rational block and uneven sampling ratio as well as adopting error analysis in the process of reconstruction. The weighted mean information entropy is adopted as the basis for partitioning of BCS which results in a flexible block group. Furthermore, the synthetic feature (SF) based on local saliency and variance is introduced to step-less adaptive sampling that works well in distinguishing and sampling between smooth blocks and detail blocks. The error analysis method is used to estimate the optimal number of iterations in sparse reconstruction. Based on the above points, an altered adaptive block-compressive sensing algorithm with flexible partitioning and error analysis is proposed in the article. On the one hand, it provides a feasible solution for the partitioning and sampling of an image, on the other hand, it also changes the iteration stop condition of reconstruction, and then improves the quality of the reconstructed image. The experimental results verify the effectiveness of the proposed algorithm and illustrate a good improvement in the indexes of the Peak Signal to Noise Ratio (PSNR), Structural Similarity (SSIM), Gradient Magnitude Similarity Deviation (GMSD), and Block Effect Index (BEI).

Early Stopping Criteria for Levenberg-Marquardt Based Neural Network Training Optimization

In this research we train a direct distributed neural network using Levenberg-Marquardt algorithm. In order to prevent overtraining, we proposed correctly recognized image percentage based on early stop condition and conduct the experiments with different stop thresholds for image classification problem. Experiment results show that the best early stop condition is 93% and other increase in stop threshold can lead to decrease in the quality of the neural network. The correct choice of early stop condition can prevent overtraining which led to the training of a neural network with considerable number of hidden neurons.