Joint feature enhancement mapping and reservoir computing for improving fault diagnosis performance

Complicated industrial robot structure and harsh working conditions may cause signal features collected in the condition monitoring process to be seriously disturbed. In this paper, a joint feature enhancement mapping and reservoir computing (FEM-RC) method is presented to handle the industrial robot fault diagnosis problem. Firstly, a feature enhancement mapping (FEM) method is proposed to achieve intraclass distance minimization and interclass distance equalization to obtain an enhanced feature matrix. Then, the first reservoir computing (RC) network is adopted to map the original feature matrix to the feature enhancement matrix, and the second RC network is for fault type classification. The results of the experiment carried out on a six-axial industrial robot demonstrate that compared with other peer models, the present FEM-RC has better fault diagnosis performance and robustness.

Real-time vehicle detection and computer intelligent recognition through improved YOLOv4

Vehicle detection is one of the key techniques of intelligent transportation system with high requirements for real-time and accuracy. To better balance the requirements, a vehicle detection algorithm based on the You Only Look Once (YOLO) v4 is proposed in this paper. On the one hand, the improved depthwise separable convolution is adopted to ensure the real-time performance. On the other hand, a novel feature fusion network is designed to gather more original feature information of different depth network layer. Experimental results show that the proposed algorithm can reduce the detection time by half while ensuring the accuracy, compared with the pristine YOLOv4.

Feature Selection for Regression Based on Gamma Test Nested Monte Carlo Tree Search

This paper investigates the Nested Monte Carlo Tree Search (NMCTS) for feature selection on regression tasks. NMCTS starts out with an empty subset and uses search results of lower nesting level simulation. Level 0 is based on random moves until the path reaches the leaf node. In order to accomplish feature selection on the regression task, the Gamma test is introduced to play the role of the reward function at the end of the simulation. The concept Vratio of the Gamma test is also combined with the original UCT-tuned1 and the design of stopping conditions in the selection and simulation phases. The proposed GNMCTS method was tested on seven numeric datasets and compared with six other feature selection methods. It shows better performance than the vanilla MCTS framework and maintains the relevant information in the original feature space. The experimental results demonstrate that GNMCTS is a robust and effective tool for feature selection. It can accomplish the task well in a reasonable computation budget.

A Coarse-to-Fine Contour Optimization Network for Extracting Building Instances from High-Resolution Remote Sensing Imagery

Building instances extraction is an essential task for surveying and mapping. Challenges still exist in extracting building instances from high-resolution remote sensing imagery mainly because of complex structures, variety of scales, and interconnected buildings. This study proposes a coarse-to-fine contour optimization network to improve the performance of building instance extraction. Specifically, the network contains two special sub-networks: attention-based feature pyramid sub-network (AFPN) and coarse-to-fine contour sub-network. The former sub-network introduces channel attention into each layer of the original feature pyramid network (FPN) to improve the identification of small buildings, and the latter is designed to accurately extract building contours via two cascaded contour optimization learning. Furthermore, the whole network is jointly optimized by multiple losses, that is, a contour loss, a classification loss, a box regression loss and a general mask loss. Experimental results on three challenging building extraction datasets demonstrated that the proposed method outperformed the state-of-the-art methods’ accuracy and quality of building contours.

Собака як спостерігач і свідок історії (за романом Вікторії Амеліної Дім для Дома)

The authors of the article analyze one of the contemporary Ukrainian novels – Home for Dom (Дім для Дома, 2015) of Viktoria Amelina. Original feature of the plot is that protagonist and narrator of this work is dog Dom (Dominic). Writer, using an animalistic hero, has achieved not only a success between readers but she also has founded a new version for emotional rethinking of the past. The matter is that Viktoria Amelina tried to reveal the peculiarities of individual, family, city and national memory. Dog’s perception of the past in the novel is the author’s effort to replace accents from total estimates to relative and subtle ones. Different “faces” of memory is a value which writer shows in the examples of one Lviv family history. She combines all difficult and contradictory processes of the twentieth century – wars, genocides, repression, deportation, enslavement of man and people.

Competitive precision genome editing (CGE) assay for functional analysis at single base resolution

The two major limitations of applying CRISPR/Cas9-technology for analysis of the effect of genotype on phenotype are the difficulty of cutting DNA exactly at the intended site, and the decreased cell proliferation and other phenotypic effects caused by the DNA cuts themselves. Here we report a novel competitive genome editing assay that allows analysis of the functional consequence of precise mutations. The method is based on precision genome editing, where a target sequence close to a feature of interest is cut, and the DNA is then repaired using a template that either reconstitutes the original feature, or introduces an altered sequence. Introducing sequence labels to both types of repair templates generates a large number of replicate cultures, increasing statistical power. In addition, the labels identify edited cells, allowing direct comparison between cells that carry wild-type and mutant features. Here, we apply the assay for multiplexed analysis of the role of E-box sequences on MYC binding and cellular fitness.

Bloody Disgusting! Abjection, Excess & Absurdity: The Carnivalesque Cohesion Between Horror & Comedy in Film & Television

Horror and comedy. Screaming and laughing. Two genres and the visceral responses which they provoke, broadly considered to be polarised, apparently juxtaposed. This thesis argues that horror and comedy can be significantly more cohesive in their thematic traits, visual presentation and narrative events, than might initially be considered. Expanding a relatively underexplored academic field and building on the work of Paul (1994), the doctorate explores gross-out cinema and television in both theory and praxis. Part One opens with scholarly exploration of core theories of genre, horror and comedy. Semiotic and historical analysis and close reading of key texts in the horror, comedy, and hybrid horror comedy genre identifies and considers shared representation across the genres. Analysed texts include The Evil Dead series (1981-1992), Grimsby (2016), Nighty Night (2004-2005) and Braindead (1992). The core shared themes and representations across the genres are posited as abjection, excess and absurdity. Each of these elements is then explored in context of the tension of horror and humour co-present in the grotesque (Thomson, 1972). The paradoxical pleasure in reception (often in the disgust response) is found to align to the transgressions of the carnivalesque, and moreover, the carnivalesque grotesque (Danow, 1995, Bakhtin, 1974 et al.). These findings are then uniquely applied in praxis in Part Two in the original feature length film script Knitters! in which the women of the Potter’s Bluff Townswomen’s Guild must face an indestructible supernatural foe in an isolated Lake District resort. In the Lake District no-one can hear you scream! The Part Three exegesis reflects rigorously on the application of thesis findings in praxis, alongside detailed gnosis of the practical construction of a feature length script including close consideration of plotting, narrative pacing and characterisation.

Intelligent Prediction Mathematical Model of Industrial Financial Fraud Based on Data Mining

The essence of enterprise financial modeling is to use mathematical models to classify and sort out all kinds of enterprise information according to the main line of value creation and on this basis to complete the analysis, prediction, and value evaluation of enterprise financial situation. A reasonable financial model is also an effective means to reduce financial fraud. In this paper, a financial fraud identification model is constructed based on empirical data. In the process of model construction, the primary feature set is selected according to the financial fraud motivation theory, and then, the original feature set is obtained by Mann–Whitney test on the primary feature set, and the final fraud identification feature set is selected from the original feature set by using Relief and Boruta algorithms. Finally, based on the final fraud identification feature set, the data algorithms such as decision tree, logistic regression, support vector machine, and random forest are used to identify financial fraud. The experimental results show that the combination of financial fraud identification features constructed by the Relief algorithm and random forest model has the best recognition effect. The evaluation indexes of the G mean value and the F value were 75.86% and 78.33%, respectively.

Centered Multi-Task Generative Adversarial Network for Small Object Detection

Despite the breakthroughs in accuracy and efficiency of object detection using deep neural networks, the performance of small object detection is far from satisfactory. Gaze estimation has developed significantly due to the development of visual sensors. Combining object detection with gaze estimation can significantly improve the performance of small object detection. This paper presents a centered multi-task generative adversarial network (CMTGAN), which combines small object detection and gaze estimation. To achieve this, we propose a generative adversarial network (GAN) capable of image super-resolution and two-stage small object detection. We exploit a generator in CMTGAN for image super-resolution and a discriminator for object detection. We introduce an artificial texture loss into the generator to retain the original feature of small objects. We also use a centered mask in the generator to make the network focus on the central part of images where small objects are more likely to appear in our method. We propose a discriminator with detection loss for two-stage small object detection, which can be adapted to other GANs for object detection. Compared with existing interpolation methods, the super-resolution images generated by CMTGAN are more explicit and contain more information. Experiments show that our method exhibits a better detection performance than mainstream methods.

Detecting Web Attacks Based on Clustering Algorithm and Multi-branch CNN

Abstract—This paper proposes and develops a web attack detection model that combines a clustering algorithm and a multi-branch convolutional neural network (CNN). The original feature set was clustered into clusters of similar features. Each cluster of similar features was generalized in a convolutional structure of a branch of the CNN. The component feature vectors are assembled into a synthetic feature vector and included in a fully connected layer for classification. Using K-fold cross-validation, the accuracy of the proposed method 98.8%, F1-score is 98.9% and the improvement rate of accuracy is 1.479%.Tóm tắt—Bài báo đề xuất và phát triển mô hình phát hiện tấn công Web dựa trên kết hợp thuật toán phân cụm và mạng nơ-ron tích chập (CNN) đa nhánh. Tập đặc trưng ban đầu được phân cụm thành các nhóm đặc trưng tương ứng. Mỗi nhóm đặc trưng được khái quát hoá trong một nhánh của mạng CNN đa nhánh để tạo thành một vector đặc trưng thành phần. Các vector đặc trưng thành phần được ghép lại thành một vector đặc trưng tổng hợp và đưa vào lớp liên kết đầy đủ để phân lớp. Sử dụng phương pháp kiểm thử chéo trên mô hình đề xuất, độ chính xác đạt 98,8%, F1-score đạt 98,8% và tỉ lệ cải tiến độ chính xác là 1,479%.