Multi-Sensor Fault Diagnosis of Underwater Thruster Propeller Based on Deep Learning

With the rapid development of unmanned surfaces and underwater vehicles, fault diagnoses for underwater thrusters are important to prevent sudden damage, which can cause huge losses. The propeller causes the most common type of thruster damage. Thus, it is important to monitor the propeller’s health reliably. This study proposes a fault diagnosis method for underwater thruster propellers. A deep convolutional neural network was proposed to monitor propeller conditions. A Hall element and hydrophone were used to obtain the current signal from the thruster and the sound signal in water, respectively. These raw data were fast Fourier transformed from the time domain to the frequency domain and used as the input to the neural network. The output of the neural network indicated the propeller’s health conditions. This study demonstrated the results of a single signal and the fusion of multiple signals in a neural network. The results showed that the multi-signal input had a higher accuracy than the one-signal input. With multi-signal inputs, training two types of signals with a separated neural network and then merging them at the end yielded the best results (99.88%), as compared to training two types of signals with a single neural network.

Bootstrap Aggregation Technique for Evaluating the Significance of Manufacturing Process Parameters in the Glass Industry

The article presents the application of the bootstrap aggregation technique to create a set of artificial neural networks (multilayer perceptron). The task of the set of neural networks is to predict the number of defective products on the basis of values of manufacturing process parameters, and to determine how the manufacturing process parameters affect the prediction result. For this purpose, four methods of determining the significance of the manufacturing process parameters have been proposed. These methods are based on the analysis of connection weights between neurons and the examination of prediction error generated by neural networks. The proposed methods take into account the fact that not a single neural network is used, but the set of networks. The article presents the research methodology as well as the results obtained for real data that come from a glassworks company and concern a production process of glass packaging. As a result of the research, it was found that it is justified to use a set of neural networks to predict the number of defective products in the glass industry, and besides, the significance of the manufacturing process parameters in the glassworks company was established using the developed set of neural networks.

Tandem Deep Learning Side-Channel Attack on FPGA Implementation of AES

Side-channel attacks have become a realistic threat to implementations of cryptographic algorithms, especially with the help of deep-learning techniques. The majority of recently demonstrated deep-learning side-channel attacks use a single neural network classifier to extract the secret from implementations of cryptographic algorithms. The potential benefits of combining multiple classifiers using the ensemble learning method have not been fully explored in the side-channel attack’s context. In this paper, we propose a tandem approach for the attack in which multiple models are trained on different attack points but are used in parallel to recover the key. Such an approach allows us to considerably reduce (33.5% on average) the number of traces required to recover the key from an FPGA implementation of AES by power analysis. We also show that not all combinations of classifiers improve the attack efficiency.

An individual claims reserving model for reported claims

We present a claims reserving technique that uses claim-specific feature and past payment information in order to estimate claims reserves for individual reported claims. We design one single neural network allowing us to estimate expected future cash flows for every individual reported claim. We introduce a consistent way of using dropout layers in order to fit the neural network to the incomplete time series of past individual claims payments. A proof of concept is provided by applying this model to synthetic as well as real insurance data sets for which the true outstanding payments for reported claims are known.

A Stopping Criterion for the Training Process of the Specific Signal Generator

Mathematical description of a complex signal is very important in engineering but nearly impossible in many occasions. The emergence of the Generative Adversarial Network (GAN) shows the possibility to train a single neural network to be a Specific Signal Generator (SSG), which is only controlled by a random vector with several elements. However, there is no explicit criterion for the GAN training process to stop, and in real applications the training always stops after a certain big iteration. In this paper, a serious issue was discussed during the process to use GAN as a SSG. And, an explicit criterion for the GAN as a SSG to stop the training process were proposed. Several experiments were carried out to illustrate the issues mentioned above and the effectiveness of the stopping criterion proposed in this paper.

Real-Time Object Detection Model

Object Detection is a study in the field of computer vision. An object detection model recognizes objects of the real world present either in a captured image or in real-time video where the object can belong to any class of objects namely humans, animals, objects, etc. This project is an implementation of an algorithm based on object detection called You Only Look Once (YOLO v3). The architecture of yolo model is extremely fast compared to all previous methods. Yolov3 model executes a single neural network to the given image and then divides the image into predetermined bounding boxes. These boxes are weighted by the predicted probabilities. After non max-suppression it gives the result of recognized objects together with bounding boxes. Yolo trains and directly executes object detection on full images.